Disentangling Discursive Spaces of Knowledge Refused by Science: An Analysis of the Epistemic Structures in the Narratives Repertoires on Health During the Covid-19 Pandemic

Abstract

This chapter provides an understanding of the social configurations with which Refused Knowledge Communities (RKCs) attribute credibility to knowledge about healthcare and wellbeing. This study focuses on how RKCs enrol knowledge claims and heterogeneous actors to build, maintain and legitimise forms of knowledge refused by science. The analysis relies on empirical materials related to the online discourses shared in the Alkaline Water (AW) and Five Biological Laws (5BLs) RKCs from January 2020 to December 2021—a time span characterised by the emergence of the Covid-19 pandemic and the management of the related health crisis—by identifying in each RKC distinct claims of refused knowledge and the actors that sustain these claims. Through a combination of qualitative analysis and network-analytic techniques, we examine the epistemic structures of the AW and 5BLs RKCs and formalise the connections between claims and actors within each RKC by a two-mode network in which claims are connected to actors. By means of community detection, we provide a visual analysis of the configuration of claim–actor connections, while using betweenness centrality scores to denote ‘flexible’ objects that link diverse sub-groups of nodes—that is, claims or actors that act as ‘boundary objects’ within these complex social worlds.

6.1 Introduction

Epistemic positioning matters in defining the social worlds that build knowledge claims. As Chap. 1 argued, the research project of which this study forms part labels its object of study refused knowledge (RK), taking into account the positioning of science, which situates RK claims outside the boundaries of knowledge corpora that it considers legitimate. Consistent with the symmetric perspective of STS, this chapter aims to understand how refused knowledge communities (RKCs) position science with respect to their knowledge claims, to comprehend if these social worlds refuse the science that denies them validity or adopt strategies designed to enrol science—i.e. scientific knowledge’s claims, technoscientific devices, scientists’ institutions, scientists and scientific papers. To this end, we will enquire into the meaning-making processes regarding health within social worlds made up of people who, within RKCs, work with shared objects to legitimise their knowledge claims. The hypothesis of this work is that knowledge can be analysed as a discursive assemblage made up of both knowledge claims and heterogeneous actors enrolled to legitimise this knowledge. We have therefore enquired into the association processes via which RKCs enrol claims and actors within their discursive universes and, thus, the ways in which they build meanings and attribute credibility to knowledge about health. As we shall see, a particular kind of enrolment process concerns how science is recruited by RKCs in legitimising the knowledge they build.

The methodology chosen is based on quantitative and qualitative procedures combined in a nested research design. More specifically, ours is a narrative approach (Czarniawska, 2004) integrated with the methodological framework of social network analysis (SNA; Wasserman & Faust, 1994; Scott, 2000). This set of techniques allowed us to visualise and analyse the relational properties of knowledge assemblages shared by RKCs, thus uncovering the structures of discursive configurations that build, maintain and legitimise these forms of knowledge. Finally, analysing the narrative repertoires shared in different discursive configurations permitted us to identify the primary narrative structures within the RKCs’ discursive universes.

The analysis focuses on the online discourses shared in the Alkaline Water (AW) and Five Biological Laws (5BLs) RKCs from January 2020 to December 2021 during a time span characterised by the emergence of the Covid-19 pandemic and the management of the related health crisis. Health issues gained prominence during the pandemic not only for RKCs but also in society as a whole. Our interest in these social worlds was motivated by the fact that RKCs developed a corpus of knowledge on health and wellbeing which is refused by scientific institutions, but without refusing institutional science. This peculiarity makes such RKCs of interest in the study of the ways in which they incorporate science into their discourses.

This chapter is organised as follows. The next section focuses on concepts borrowed from the theory of social worlds and employed them in the network analysis performed in our study. Section 6.3 describes the methodology and empirical material used in this study. This section clarifies the use of SNA in the context of the theoretical framework of social worlds, considering the use of SNA in Science and Technology Studies (STS), which has been the subject of much debate. Section 6.4 focuses on describing the analysis and its main results. Finally, the last section focuses on discussion and conclusions.

6.2 Analysing Spaces of Epistemic Enrolment Within RK Social Worlds

To understand how, within their discourse universes, RKCs enrol diverse claims and actors to legitimise their knowledge, we will borrow a number of key concepts from the social worlds’ perspective. According to Clarke and Star (2008), the social world framework focuses on meaning-making processes within groups of actors ‘doing things together’ (Becker, 1986) and working with shared objects. Here, the focus is not on the ‘doing’ but on the linguistic utterances as part of the discursive construction of such objects, which can also be pieces of knowledge and play a central role in our analysis. It is around these objects that knowledge claims are built and conveyed in these social worlds. Our main reference is thus to the definition of social worlds as universes of discourse (Strauss, 1978), namely shared discursive spaces that are profoundly relational in nature, which prompted us to adopt a narrative approach (Czarniawska, 2004) to enquire into the most significant forms of narratives used by RKCs to legitimise and thus stabilise the knowledge they perceive as being refused by science and mainstream world views.

To identify the configurations on which RK relies, we opted for an approach derived from the sociology of associations (Callon, 1984; Latour, 2005) designed to trace the connections between knowledge claims and heterogeneous actors enrolled within these discursive worlds to support those claims. Analysing knowledge as an assemblage of claims and actors underlines the profoundly relational nature of knowledge itself and understands the latter’s sharing within a community as one of the main factors attributing it the status of knowledge, regardless of the truthfulness or falsity of its contents (Bloor, 1976).

Claims of knowledge are defined here as the cognitive elements considered true within a social world and constituting segments of its corpus of shared knowledge. The epistemic enrolment space is the set of discursive structures that guide, focus and delimit RK credibility attribution processes by assembling and re-assembling epistemic, social and political structures. In the case studies considered here, epistemic enrolment space analysis focuses on the discursive texture built by RKCs’ entrepreneurs (Clarke & Star, 2008), namely individuals, or groups of individuals, who are deeply committed to, and active in, promoting RK in online spaces within the social worlds observed.

In accordance with the social worlds’ perspective, we will also examine the role played in these discursive universes by implicated actors, i.e. ‘actors silenced or only discursively present—constructed by others for their own purposes’ (Clarke & Star, 2008, p. 119). As discursively constructed primarily by RKC entrepreneurs to sustain RK, implicated actors are neither actively involved in negotiating self-representation in social worlds nor considered for what they say, write and argue; yet they can play a determinant role in enrolment processes into forms of knowledge. Finally, we will consider how both human and non-human actors are mobilised in the making up of epistemic configurations and thus we use the term actors to refer to both human and non-human actors.

Hence, within the epistemic enrolment space we will investigate knowledge claims concerning health as it is maintained by RKCs, along with the networks of enrolment and counter-enrolment (Callon & Law, 1982) built to affirm this knowledge. The elements assembled in such networks were identified via web-ethnography during our research to enable us to explore RKC narratives as proxies to the re-assembling of the social (Latour, 2005), i.e. as a way to grasp how the various narratives bring heterogeneous elements together and into meaningful wholes (Czarniawska, 2004). The narratives constructed by RKC entrepreneurs contribute to sustaining wider narrative structures through which meanings and their relation to social worlds can be built and shared. Diverse sets of actors are enrolled into these narrative structures, to support and entangle the discourse universes deemed significant by RKCs. Moreover, in these discursive structures, our interest was identifying the objects (claims and actors) coexisting in the diverse structures making up the epistemic enrolment space and building narratives centred on a range of repertoires. Our analysis will focus on boundary objects (Star & Griesemer, 1989), nodes in the narrative structure network where various social worlds meet in arenas of mutual concern. Our interest in these objects was based on the key role they play within the translation processes (re)constructing meanings to meet the specific needs or demands of the various social worlds involved (Star, 1989).

6.3 Methodology and Data

Our analysis used a mixed methods perspective by combining the narrative approach (Czarniawska, 2004), designed to identify the discursive structures of the social worlds, and SNA.¹ This methodological strategy was chosen with a view to examining the relational structures at play in the enrolment of the various types of actors supporting RK claims within RKCs’ online discursive spaces and the narrative structures that inform the epistemic enrolment space of these social worlds. The stages in our analysis are shown in Fig. 6.1. As the entire data collection and analytical process dealt with qualitative data and prioritised the interpretation and analysis of texts and network graphs over formalisation, we consider our work to be concerned with qualitative networks (Bellotti, 2014; Hollstein, 2011).

Fig. 6.1

Analytical process in each case study

In the first step in this research, from January 2020 to December 2021, our research group conducted web-ethnography on AW and 5BLs RKCs’ online spaces (blogs, Facebook pages and profiles, YouTube channels and the like) (Chap. 1). Using content analysis tools (Lieblich et al., 1998), we analysed the diaries resulting from this web-ethnography and, through an iterative coding and recoding process, we identified: (a) the health-related claims constituting the core of the corpus of refused knowledge;² (b) the enrolled actors interacting in the discursive universes (including implicated actors); (c) the linkages between claims and enrolled actors.

The three entity classes were translated into the elements constituting the two-mode networks discussed below. We also considered mimicry practices to be one of the strategies pursued by RKCs to give epistemic legitimacy to an RK corpus and depict it as an attribute of claim–actor linkages (see Chap. 1). Of the four enrolment strategies identified by the project, we chose to focus only on mimicry because of its significance in RKCs’ attempts to enrol in science. Performing mimicry strategies—from the simplest reference (either textual or visual) to technoscientific devices, e.g. a microscope or an oxidation-reduction potential (ORP) metre to more sophisticated biological elements and processes, e.g. extracellular pH, tumour micro-environment or T cell apoptosis—means borrowing science’s constitutive ‘marks’ of scientific authority, including scientific institutions, scientists and scientific papers. The third step formalised the connections between claims and actors through two-mode networks, which proved to be useful in operationalising linkages between heterogeneous actors (Mützel, 2009)³ and emerged as appropriate to the kind of relational structure we intended to analyse. In fact, in two-mode networks, ties are allowed only between nodes belonging to two distinct node sets, as is the case with claims (Set 1) connected to actors (Set 2). These two-mode networks permitted us to visually explore the way claims are connected to actors, and thus the way actors aggregate around narratives expressed by claims. The analytical strategy we pursued therefore implied that actor–actor connections are mediated by the claims they jointly support whilst, conversely, claims are connected to each other insofar as they are sustained by the same enrolled actors—which is precisely one of the main features of two-mode (affiliation) networks, i.e. their duality (Breiger, 1974).

Adopting Actor Network Theory’s perspective (Latour, 2005) led us to consider the associations between these elements as social in that such networks constitute a representation of social worlds as assemblages of heterogeneous actors and claims at work in discursive enrolment. As far as actor heterogeneity is concerned, it is well known that two-mode affiliation networks can help scholars produce ‘heterogeneous maps’, but do not allow distinctions between the nodes or ties comprised in a single ‘mode’, which is considered a limitation (Cambrosio et al., 2004; Mützel, 2009, p. 874, 878; Venturini et al., 2019, p. 515). There is no doubt that, in our analysis, heterogeneous associations coalesce into a single set of linkages—whilst paying attention to the diverse strategies pursued to sustain claims through actors’ enrolment (see below)—but the differences between types of actors are retained as a node attribute (see Sect. 6.4).

Our network graphs were drawn up using Gephi 0.9.7 software which we also used to obtain network statistics. We made use of a force-directed layout named ForceAtlas2 (Jacomy et al., 2014) to spatialise networks and thus exploit the potential of visual network analysis (Venturini et al., 2021). As a basic centrality measure, we considered betweenness centrality for two-mode networks (Faust, 1997; Brandes, 2001; Everett & Borgatti, 2005). Notably, betweenness centrality expresses the potential for a node to act as broker or intermediary in a network (Scott, 2000, p. 86), which means a claim or actor connects different areas of the graph. In addition, we performed a community detection analysis using the Louvain modularity algorithm implemented in Gephi (Blondel et al., 2008). Each cluster (or modularity class) emerging from this analysis combined densely connected claims and actors.⁴ As far as the mimicry strategy was concerned, we highlighted this in the network graph by tie colour (see below).

Lastly, we qualitatively analysed the narratives assembled in each clustering structure by assigning a given repertoire to each of them, where this repertoire was the result of a further content analysis of the ensemble of claims and actors making up the clusters.⁵ This also enabled us to detect several sub-structures within these networks, namely different sub-worlds each of which can be made of distinct clusters or even a combination of different clusters. This last step drew on the qualitative side of our analysis to interpret the betweenness centrality scores: when a high betweenness score expresses a ‘flexibility’ of objects in connecting diverse sub-groups of nodes that relate to it for different purposes, these objects (claims or actors) can be regarded as boundary objects that ‘inhabit several intersecting social worlds’ (Star & Griesemer, 1989, p. 393).

Finally, we analysed the network obtained from a union between the two social world networks examined. This analysis allowed us to identify the crucial statements and actors present in both RKCs, thus understanding which Covid-19 pandemic period actors and statements helped legitimise the construction and dissemination of forms of RK and which discursive structures were activated within these social worlds to provide epistemic credibility to these forms of knowledge.

6.4 Analysis

The content analysis of the entrepreneur actors’ discursive universes performed in the first step of the study provided two sets of claims (192 for the AW RKC; 365 for the 5BLs RKC) and actors (1939 for AW RKC, 1940 for 5BLs RKC) which, as a whole, constitute the ‘dual’ health-related knowledge cores of each of these universes. By enrolling these actors and setting forth these claims, entrepreneurs handle the knowledge cores assembled to build and legitimise RKCs’ claims. We identified various categories of enrolled actors, such as (1) organs, tissues and cells; (2) diseases; (3) polluting pathogens; (4) scientific disciplines; (5) distinguished international scientific scholars; (6) authors of scientific papers; (7) public figures active in the debate on Covid-19; (8) media and social networks; (9) scientific journals; (10) scientific institutions; (11) people who participate in chats (e.g. with comments) on the online spaces run by the entrepreneurs and (12) other concerned actors (e.g. children, the elderly and shopkeepers).

Following the steps outlined in the previous section, we translated these knowledge cores regarding health into networked form by focusing on the links between each claim and the various actors, which were in turn connected to other claims. The assemblage as a whole thus resulted in a complex configuration of network nodes constituting one possible representation of these RKCs and provided a map of their shared knowledge.

Indeed, a first examination of the networks analysed revealed a degree of complexity that hindered their readability due to excessive relational data ‘noise’. In other words, using these networks as maps required moving upward from a poorly informative terrain in which claims and actors may be associated with a minimum of one or two nodes (actors or claims, respectively) to a richer analytical framework in which associations involve at least three units for each claim or actor. We therefore focused analysis on a sub-network of each RKC extracted through a degree-based procedure called k-cores⁶ (Seidman, 1983) and then chose to limit our analysis to a subgraph with k = 3, that is, a 3k-core (Scott, 2000, p. 110; see Table 6.1).

Table 6.1 Number of nodes in the networks built for the two RKCs and their unions

For both RKCs, the community detection algorithm generated a clusterisation of claims and actors. This was the first main finding in our analysis, i.e. that the discursive spaces depicted via SNA were organised around various narrative repertoires that could be seen in the clusters resulting from modularity analysis. The structures observed rendered the heterogeneity of assemblages and highlighted the differential associations revolving around knowledge-specific cores represented by the repertoires characterising the clusters. We thus analysed the structural configuration of claims and actors emerging from modularity analysis and identified the main repertoire within each cluster. Indeed, the clustering of the discursive universes showed not only that RKCs use a range of repertoires but also that such repertoires adopt enrolled actor types that are specific to them. For instance, Cluster 5 in the AW RKC (see below) was concerned with cancer and chronic disease prevention through an alkaline diet and its actors included two scientific institutions, namely the American Association for Cancer Research and the American Society for Biochemistry and Molecular Biology, along with two papers (published in the journal Cancer Research), tumour cells and metabolic processes.

The node colours in the graphs shown below for the two RKCs and their combination denote modularity class. Edge colours denote the presence (black) or absence (grey) of a mimicry strategy. A caption in each figure shows modularity class number and colour, plus the percentage (in parentheses) of nodes included in each class.

6.4.1 The Structure of Discursive Universes Legitimising RK Within the Alkaline Water RKC

Figure 6.2 illustrates the clustering structure of the claim–actor network within the AW RKC (modularity = 0.645). The content analysis of the claims showed a high degree of homogeneity of repertoires in each cluster, consequently different partitions can be classified as belonging to the narrative repertoires shown in Table 6.2.

Fig. 6.2

Two-mode network of claims and enrolled actors: the case of AW RKC (black lines = mimicry strategy)

Table 6.2 Clusters and related narrative repertoires within the AW RKC

By visually analysing the network, we detected some central clusters (i.e. those with modularity classes 8-9-6-4-0-1-2-3) and some peripheral ones (modularity classes 5-10-7). As far as the related narratives are concerned, Clusters 8-9-6—those mostly scattered across the core of the network—provided the primary repertoires used by RKC members to sustain alkaline water’s and food’s ability to purify the body and defend it against the effects of toxic and carcinogenic pathogens, including the power of the alkaline lifestyle to enhance the immune system. Note also that clusters 9-6 (which are identically labelled) refer to the same repertoire, although they are distinct in the modularity analysis because of their different network connection patterns. As far as actors are concerned, biomolecular actors prevail in Cluster 9, while Cluster 6 includes actors better fitting the highly energetic lifestyle idea. Both Clusters 9 and 6—along with Cluster 3—encompassed claims asserting that water and alkaline nutrition improve physical and mental performance, stimulate fertility, are beneficial during pregnancy and counteract inflammation caused by acidosis of tissues responsible for serious diseases and tumours. In the central Clusters 4 and 0, the promotion of water and alkaline food as a healthy lifestyle focused on different repertoires: acid–base balance as a characterising element for a healthy body and as an anti-ageing factor (Cluster 4); alkaline water and diet as a defence against viruses, including the SARS-CoV-2 coronavirus (Cluster 0). The latter cluster includes claims contesting government anti-Covid-19 health policies. Cluster 1 is strongly connected to Cluster 0 and presents alkaline treatment as a way of enhancing personal awareness. Cluster 2 is less pervasive in the graph but still significant in its focus on claims arguing against the scientific approach to cancer, chronic diseases and Covid-19 treatment.

Another area is made up of peripheral Clusters 5-10-7. The Cluster 5 repertoire focused on alkaline water and food’s ability to prevent cancer and chronic disease. Although claims relating to causes of tumours and their treatment were distributed across all clusters, this one featured claim–actor relationships seeming to favour the enrolment of scientific actors in arguments for alkaline water as a way of preventing and treating tumours, such as scientific journal articles, their authors or tumour physiology subjects. In addition, science also plays an important part in the Cluster 10 repertoire although, in this case, it is not enrolled to legitimise the AW RKC corpus of knowledge but rather falls into this repertoire for its perceived inability to understand and treat diseases, given its failure to consider the mind–body relationship. Cluster 7 includes claims arguing for the use of alkaline water to improve physical performance and strengthen the immune system, mainly on the basis of actors in the biology and physiology spheres, as well as scientific institutions or physicians.

Tables 6.3 and 6.4 show the claims and actors with the highest betweenness centrality values. We noted that although both SARS-CoV-2 and Covid-19 are present in alkaline RKC narratives and feature in Cluster 0-2 repertoires, they do not play a key bridging role in them–either as a component in the claims or as actors. The narratives revolve mainly around associations between acidic and alkaline body conditions and their consequences. In this sense, claim N113 (‘Acidosis causes chronic diseases’) bridged Cluster 5 and the rest of the network by focusing on chronic diseases as a consequence of acidity, while claim N137 (‘Alkaline water has an anti-ageing effect’) lay mainly in a central position, arguing for the anti-ageing effects of alkaline water, i.e. a less extreme assertion that helps explain this location in the network. The following are the three most central actors: (1) publishers, because of their ability to provide RK with a readership; (2) ionisers, for their chief role as ‘flexible’ technological devices—as they serve various needs and have a range of possible uses and purposes (drinking, cleaning, saving money, avoiding plastic, etc.); (3) inflammation, as a widespread condition impacting health and wellbeing with various degrees of severity. Finally, the ties in the network denote the widespread use of mimicry practices (64.6% of the ties) to legitimise RKC entrepreneurs’ knowledge claims. Here, mimicry also goes along with the enrolment of scientific and institutional sphere actors. Exceptions to the prevalence of mimicry are provided in Cluster 10, visibly peripheral and related to criticisms of institutional medicine, and in other more central areas of the network, mostly involving Clusters 9-6-0, where the strategy is partly mixed.

Table 6.3 Betweenness centrality of claims in the AW RKC network

Table 6.4 Betweenness centrality of actors in the AW RKC

Finally, we noted that the analysis highlights not only the significant role in holding different narrative repertoires together played by ‘tumour’ enrolment but also by knowledge claims relating to the use of alkaline water to prevent tumours and ‘acidosis’ (combated by the alkaline diet) as a cause of tumours.

6.4.2 The Structure of Discursive Universes Legitimising RK Within the Five Biological Laws RKC

The 5BLs RKC network would seem to be more complex than AW (Fig. 6.3). Basically, both the network and its 3k-core partition are larger than the other cases (see Table 6.1). This is due to (1) the higher level of interaction observed in the relevant online spaces, also evidenced by a large number of online users coded as actors, and (2) the wider spectrum of the knowledge contents coded as claims. As in the former case, the networked 5BLs RKC was divided up into different areas identified using the modularity algorithm. Its community structure (modularity = 0.762) was rather fragmented with the algorithm yielding 17 modularity classes. Similar to AW, these clusters form both a core area (bottom centre of the graph) and a number of peripheral areas, plus several clusters occupying less dense areas and with sparser distribution than the others.

Starting from partitions with the largest proportion of nodes, Cluster 7, with its ‘Causes of disease: fear and psycho-social conflicts’ repertoire, is located bottom left in the graph and divided up into two parts, one of which is highly cohesive and peripheral while the other is sparser and closer to the core of the graph. The two parts of Cluster 7 are mainly held together by one of the claims with the highest betweenness in this graph, namely CLB186 (‘Fear of death causes pneumonia’), which is representative of the repertoire of this cluster and also one of the claims relating health narratives to the Covid-19 crisis. Interestingly, the connection between this claim and the right half of the cluster is based on strategies other than mimicry: the separation, then, concerns the way this claim’s narrative is oriented towards enrolled actors, which form two seemingly unrelated sub-groups⁷ joined up by a ‘boundary claim’—which owes its role to its connection to actors from other clusters (Table 6.5).

Cluster 1’s repertoire relates to Covid-19 counter-narratives and occupies a central position in the graph, with some ramifications towards other nearby areas. This cluster comprises the two central claims CLB84 (‘Covid-19 is no more harmful than seasonal flu’) and CLB80 (‘Mortality rates from Covid-19 are very low’, see Table 6.6), both of which are related to denying the dangers of Covid-19 (Table 6.7). These two claims are responsible, to a considerable extent, for Cluster 1’s central position because of their connections with nodes from other clusters—notably often related to people who comment on content online (one of which has the highest betweenness of all the actors). Cluster 4’s location (bottom of the graph) in the network is also a subtle one. The narrative repertoire of this cluster is about denying the dangers of viruses in general. It is split up into two sub-partitions, plus other sparse nodes. The two-halves of the cluster are kept connected by the highly central CLB134 claim, which states that ‘The virus is not dangerous’. In sum, the positions of the first three clusters reflect the way in which the Covid-19-related repertoire tends to spread across the RKC, albeit in different forms.

A further set of repertoires belongs to clusters intersecting with the above. This is the case of Cluster 6 (‘Links between institutional medicine, economic interests and policies’), Cluster 13 (‘Epistemic relativism on Covid-19 and health’) and Cluster 14 (‘Unreliability of experts and institutions’), which lie at the core of the graph, though with some ramifications. This is a set of repertoires that more directly question the validity and legitimacy of science and medicine. Furthermore, Cluster 13 is split into two-halves joined by the central CLB478 claim (‘Social distancing brings about health-related and social damage’—see right-hand side of the graph), and the upper-right branch of this cluster extends to another area where Cluster 16 (‘Opacity of health institutions’) and topics more closely related to 5BLs and the latter’s interpretation of conflict and trauma (Clusters 8 and 5) are located.

Hence, Covid-19 and health institution narratives are scattered across the RKC and intersect one another, especially those related to clusters with more ties towards the core of the network than towards its edges. In this respect, Clusters 5-8-9-0 are located along the periphery. These denote narrative repertoires specific to the 5BLs and are thus more extreme than other narratives, such as more Covid-19 and related counter-narrative focused ones. As far as the presence of mimicry as a strategy is concerned, the graph shows not only how prevalent this is (67% of all ties) but also how it flows through several branches of the network, following traces of Covid-19 pandemic counter-narratives and criticisms of medicine and science. In this respect, the way Cluster 15 (‘Pandemic as a social experiment vs self-determination in health matters’) is positioned merits consideration. This partition’s subsets are detached from one another, denoting a presence within the core of the graph, along with more ‘relaxed’ narrative repertoires, and also towards the periphery (upper-right-hand side)—hence with more extreme subjects (such as totalitarianisms or ‘quantum field theory’). The extreme sub-partitions also differ in strategy; the nodes in the upper branch of Cluster 15 are bound together by linkages unrelated to mimicry, while the opposite is true of the lower branch. Finally, this strategy also characterises the peripheral cluster ties related to 5BLs specific repertoires.

Fig. 6.3

Two-mode network of claims and enrolled actors: the case of 5BLs

Table 6.5 Clustering of narrative repertoires within the 5BLs RKC

Table 6.6 Betweenness centrality of claims for the 5BLs RKC network

Table 6.7 Betweenness centrality of enrolled actors for the 5BLs RKC network

6.4.3 Combining the Network Structures of the Two RKCs

The network combining the two RKCs (Fig. 6.4) is modular in structure (modularity = 0.762) with a complex appearance in that it at least partly parallels the two distinct RKC groupings but also, at the same time, reveals some merging between the AW and 5BLs RKC repertoires. The clustering results shown in Table 6.8 indicate that several clusters share the same repertoire and relate to claims from one of the two social worlds or to their combinations. The repertoires of the two RKCs tend to preserve their specificity. The 5BLs RKC focuses on criticisms of the way governments and health institutions have managed the pandemic (modularity Classes 13–14), general criticism of institutional medicine along with the denial of the Covid-19 pandemic (modularity Classes 1-7-8-12-15, even though Clusters 7 and 8 include a minimum of alkaline RKC claims) and 5BLs’ interpretation of the psychological and social causes of the pandemic and the damaging effects of social distancing and protection devices (links to a counter-narrative on vaccines, Clusters 0-3-5). Clusters 2 and 4 are related to the AW RKC repertoires; namely, the acidity-alkalinity dichotomy and the beneficial effects of food and water for health and wellbeing (the latter with a minimum of claims from the 5BLs RKC). Cluster 6 is the only truly mixed one in terms of the origins of its claims.

Fig. 6.4

Two-mode network of claims and enrolled actors obtained by joining the AW and 5BLs RKCs

Table 6.8 Clustering of narrative repertoires within the network obtained by joining the AW and 5BLs RKCs

Table 6.9 Betweenness centrality of claims for the joint AW and 5BLs RKC network

Table 6.10 Betweenness centrality of claims for the joint alkaline and 5BLs SW network

Despite this apparent segmentation, the two RKCs seem to interact in some way, particularly if we look at the claims and enrolled actors with the highest betweenness centrality scores (Tables 6.9 and 6.10). The graph in Fig. 6.4 shows the labels of the nodes with the highest scores for betweenness only. As far as the claims are concerned, the first five sorted by betweenness score lie on the upper (5BLs) and lower (AW) sides of the graph and concern, respectively, Covid-19-related narratives and those regarding acidic/alkaline polarities in relation to health; this also means that the network is virtually divided up into these two RKCs.⁸

As far as the five enrolled actors (children, physicians, viruses, the elderly and symptoms) with the highest betweenness score are concerned, it is noteworthy that these are the actors that truly connect the two-halves of the graph and, more generally, the two different RKCs in them. Topologically, they are also exactly central in the graph, and when their connections are observed in detail, they link claims from both social worlds. More importantly, these actors can be considered boundary objects for their potential to translate interests from the different RKCs.

6.5 Discussion and Conclusions: Disassembling and Re-assembling Science

This chapter presented an enquiry into the epistemic structures constituting the AW and 5BLs RKCs, in an attempt to discover (1) the core RK claims concerning health within the discursive universes in which they are constructed and disseminated, (2) enrolment by advocates of RK and (3) how entrepreneurial actors position their claims vis-à-vis science. In this respect, the choice to pursue a combination of qualitative analysis and network analysis techniques is in line with the chapter’s aim to study RKCs as social worlds and map their organisation as assemblages among claims and heterogeneous actors. The aim of the chapter’s modularity analysis was to cast light on the structure of these social worlds and make them more understandable. This analysis provides the reader with an overview of the associations that emerged as significant in their support for RK, but with an eye to how entrepreneurial actors enrol other heterogeneous actors.

Thus, the clustering of the online discursive universes examined showed not only that entrepreneurial actors use differentiated repertoires but also that these repertoires rely on specific types of enrolled actors. The configuration of the assemblages involving claims and actors relates to RKCs’ shared purposes and practices, however the latter are internally differentiated. Thus, one or more clusters may comprise elements whose association emanates from a given ‘commitment’ regarding health (e.g. alkaline water as a cure, symptoms as psychic-bodily reactions, etc.). This commitment is highlighted as the analysis displays the different repertoires constituting the knowledge core represented by the clusters obtained through community detection. In this respect, this analysis revealed that knowledge about health as it is practised within social worlds may be constituted through commitment to, and participation in, one or more RKCs, leading to broader arenas made up of multiple worlds organised ecologically around issues of mutual concern and commitment to action (Clarke & Star, 2008; Shibutani, 1955; Strauss, 1959). What is of chief interest in the configuration of the RKCs examined here is that discursive enrolment occurs with a deployment of different actors functioning as allies irrespective of their origin, usage or function.

Thus, the analysis showed the hybrid nature of the RKC epistemic enrolment space. Reticular representation of the discursive universes of entrepreneurial actors who play a leading role in RK legitimisation processes allowed us to analyse the way such RK is supported by networks constructed by assembling actors and claims from different RKCs. In the transposition from one social world to another, the roles and interests of actors and claims change. These latter, as elements in assemblages, are arranged and aligned to respond to the cognitive needs defined in the various clusters identified in the analysis, but without being forced to conform entirely to the different local settings in which they are enrolled.

In this regard, observing the science transposition processes, which are enrolled in various forms in an attempt to support the legitimation of RK, is particularly significant. In the discursive universes designed to increase RK credibility, ongoing processes involving moving closer to and further away from science were observable. Within the RKCs epistemic enrolment space, the enrolment of science emerges as the result of a continuous (re)negotiation of science contestation processes, on one hand, and science purification processes, on the other. These two trends emerged from our analysis as follows.

The narrative repertoires marshalled to contest science revolve around the medicine betrayed theme. Having abandoned a holistic vision of treatment that conceives of wellbeing as an expression of an integrated mind–body organism and the medical profession’s ethical principles as set out in the Hippocratic Oath, modern medicine is unable to understand the causes of diseases. Institutional medicine focuses on treatment of disease rather than healthcare. In science-critique narratives, a key role is played by interpretations of cancer treatment and the practices pursued to limit the dissemination of Covid-19. Both the scientific community and communication structures are enrolled in these narratives as organisations manipulated by lobbies which include denying scientific claims, thus providing alternative claims to explain health problems. The methods and tools of institutional medicine (including diagnostic ones) are rejected, as they focus on disease rather than health and on a conceptualisation of the body as a set of distinct organs, including the mind, rather than the unitary organism propounded by holistic models. Epidemiological data is denied, i.e. deemed untrustworthy on the grounds of institutional medicine’s distorted perspectives or corruption in the scientific community and information systems. Finally, official views of the causes of disease are considered to be wrong. In the case of the AW RKC, diseases are attributable to a state of acidosis in the body’s tissues, while in that of 5BLs, they are to be explained by conflicts bound up with past psychic traumas leaving biomolecular scars in human tissues.

Thus, for these RKCs health is a matter of rebuilding a state of lost equilibrium. In the AW RKC, the equilibrium referred to is the acid–base equilibrium. Displacement from this balance causes cellular ageing, inflammation, malaise, chronic and/or degenerative diseases and tumours. Alkaline water is thus considered capable of restoring this balance, and a wellbeing and prevention practice as well as a treatment for diseases and tumours. The body of knowledge advocated by the 5BLs, on the other hand, refers to a body–mind balance. Diseases are, in fact, interpreted as imbalances generated by psychic conflicts deriving from prenatal and natal traumas. Care practices within this narrative infrastructure are presented as paths of awareness requiring subjects’ active agency. And it is essentially through this process of acquiring—strongly practical and experiential—scientific knowledge that science is purified. Science itself is enrolled to legitimise forms of RK, e.g. the biomolecular claims attesting to the benefits of alkaline water on health or the scientific evidence marshalled to testify to the veracity of Hamer’s psychobiological framework model.

These disassembling and re-assembling science processes are driven by a constant reworking of claims and actors within the RKC epistemic enrolment space. A key role in these processes is played by the boundary objects identified in the analysis of the unions between the two RKCs. The merging of the repertoires belonging to the two RKCs highlights claims and enrolled actors acting as boundary objects in the narratives examined. What counts in this respect is the role played by these boundary objects in the processes of translation between different repertoires: indeed, these objects allow us to move from, say, a biomolecular repertoire to a political repertoire—as happens, for e.g. with viruses, a recombinant agent in these repertoires. Narrative structures are also sustained by these translations.

Within this heterogeneous epistemic enrolment space, the tension deriving from these science disassembling and re-assembling processes is balanced by boundary objects responding to the need for network coherence. Paradoxically, our analysis showed that a key role in holding different social worlds together is played by actors such as children, physicians, viruses, the elderly and symptoms, together contributing to reinforcing a narrative on health entirely played out within the contested narratives of these RKCs.

In this sense, a key role is played by Covid-19, which acts as an arena within these social worlds and allows further elements designed to augment RK credibility to be added. The SARS-CoV-2 virus, infection, Covid-19 symptoms, social distancing and the health-related and social damage it causes, pandemic fake news allegations and the epidemiological death and infection figures spread by governments and the media are all enrolled to bring together elements from different social worlds (among these, the social worlds of science) to further legitimise the forms of knowledge advocated. In this sense, an analysis of RKCs focusing on the assemblages at work within these discursive spaces can increase our understanding of the extent to which RK is the result of bricolage processes and a reworking of conceptions and practices which acquire meaning in relation to one another, even when the pieces of knowledge thus deployed and articulated come from science itself and are reframed and recombined, as needed, to make sense of these assemblages.