2 Towards a Multi-Layered Account of Collective Memory

There can be no such thing as an isolated utterance. It always presupposes utterances that precede and follow it. No one utterance can be either the first or the last. Each is only a link in the chain, and none can be studied outside this chain.

M.M. Bakhtin (198, p. 136)6

Memory is a vast and broad concept that has received much attention across the humanities, social sciences, and natural sciences. It is perhaps because of this wide breadth that it is difficult to pin down precisely what we mean by memory, and why the concept has resisted domestication by any singular discipline. In ordinary speech, words like ‘memory’, ‘remember,’ and ‘forget’ have relatively straightforward meanings. One has a ‘memory’ of something if they can successfully recall a personal experience, such as an early childhood experience of their birthday party, or if they can summon information that they have ‘memorised’, such as a phone number, an address, or the multiplication table. ‘Forgetting’ is usually seen as a failure of memory, referring to an inability to recall experiences and information, or an attempt to erase the memory of particular events: one can ‘forget’ to lock the front door as they leave the house in a hurry, or one can try to ‘forget’ an unpleasant encounter with a colleague by taking an extended vacation. A systemic inability to recall information is usually considered a medical condition (such as amnesia and dementia).

In addition to these cognitive, neurological, and psychological properties, memory also plays a critical function for social life, such as family histories and generational identities, as well as identities of culture, ethnicity/race, religion, and nation. What we ‘remember’ is not purely personal; we remember, share, and re-narrate experiences and anecdotes as we are told them by our parents, grandparents, friends, partners and colleagues. Through various media, we consume and share stories of important historical events in the societies in which we live. Scholars have coined terms such as ‘national memory’ to describe a body of knowledge that a ‘nation’ collectively shares with reference to ‘monumental’ events, such as wars, political events, elections, demonstrations, and celebrations. Starting in the early 1990s, many parts of the world, especially Europe, North America, and East Asia, experienced a ‘memory boom’ – an upsurge of academic and popular discourses surrounding these ‘national memories’ that roughly coincided with the fiftieth anniversary since the end of the Second World War. Spanning across the humanities and social sciences in disciplines including history, literary criticism, sociology, anthropology, archaeology, and cultural studies, terms such as ‘collective memory’, ‘cultural memory’, and ‘social memory’ have captured these socially shared and reproduced aspects of memory and memorial practices (Erll 2011; Olick 1999; Olick, Vinitzky-Seroussi, and Levy 2011; Misztal 2003).

In the past several decades, the field of memory studies has decisively shifted from classical notions of collective memory as an enduring, objective structure that constrains individuals belonging to a society, to instead treat memories as contingent and contentious objects of political struggle within and between societies. As a result, the field of memory studies has produced a dizzying array of concepts and theories to highlight different types of commemoration, from official ceremonies and national holidays to media narratives, art, and street protests, which, through their performance, enact and realise partial (in both senses of the word) interpretations of the past; this book refers to these as ‘commemorative speech acts’. The act of commemoration is a type of ‘performative utterance’, or a ‘declarative’ statement, in which the speaker creates, rather than describes, social reality through speech acts (Austin 1962; Searle 1995).

Building on these approaches that stress the contingency of commemorative processes, this chapter seeks to develop a sociologically driven account of collective memory that places the uneven differentials of power and resources as key factors that influence both the form (what kind of commemorative medium and method is used) and the content (what kinds of messages are disseminated through commemorative speech acts). As such, this chapter develops a theoretical approach informed by cultural trauma theory, as well as concepts by the literary theorist Mikhail Bakhtin. Specifically, it argues that Bakhtin’s conceptual vocabulary of ‘dialogue’, ‘polyphony’, and ‘heteroglossia’ is eminently suited to a sociological analysis of contending cultural trauma narratives.

The goal of this chapter is neither to advocate for the creation of yet new categorisations of memory nor to proclaim the superiority of Bakhtinian theory above existing ones. Instead, the chapter proposes an approach that puts the agency of social actors at the heart of struggles over memory narratives within a socially stratified and historically dependent arena of discourse. It argues that this model is useful not only for understanding the long-term impact of the Aum Affair, but for understanding how the Japanese government actors, mass media outlets, civic institutions, and individuals responded to the perceived social crisis in disparate ways.

There are many ways to define ‘collective memory’, but at its core, the term stresses the fact that memory is not solely an individual’s capacity to remember, but a type of common property shared by members of a group and passed down generations. The French sociologist Maurice Halbwachs was arguably the first to systematically theorise collective memory in this way. Halbwachs was heavily influenced by Durkheim’s methodological holism, which argues that a society shares a singular ‘collective consciousness’ (Durkheim 2008). Following Durkheim’s central argument – that the collective consciousness is more than the mere sum of its constituent parts but an objective social fact – Halbwachs held that society, as an external structural force, was integral to an individual’s capacity to interpret the world: ‘it is in society that people normally acquire their memories. It is also in society that they recall, recognize, and localize their memories’ (Halbwachs 1992: 38). Nevertheless, Halbwachs was also careful to point out that collective memory was not coterminous with Durkheim’s ‘collective consciousness’. Whereas Durkheim highlighted the powerful normative and moral drive of a single collective consciousness shared by all members of a society, for Halbwachs, collective memory existed at the level of concrete social groups, such as occupational groups, religions, and family units that reproduced narratives about themselves across multiple generations.

Much of memory studies debates has emerged as a response to Halbwachs’s influential formulation of collective memory as a supra-individual structure and an objective ‘social fact’. On the one hand, Halbwachs’s formulation is appealing because some aspects of social life are inextricably linked to experiences to which no single individual has a claim. Modern nation-states routinely create, reinforce, and reproduce national identities through commemorative events, museums, statues, street and place names, and national holidays (Billig 1995; Connerton 1989; Nora 1989; Winter 2010; 2014; Zerubavel 2003). Such an approach usefully highlights that collective memory, just like other social phenomena that precede and outlast any single individual such as language and religion, are objective social facts, rather than an aggregate of individual reflections of the past (Olick 1999).

On the other hand, the Halbwachsian approach has an implicit bias towards historical continuity and cultural conservatism: individuals can only receive and reproduce memories of their predecessors, but with little flexibility to change or challenge them. To overcome this stability bias, numerous scholars have highlighted how social actors actively shape how the past should be remembered with reference to concepts such as ‘moral entrepreneurs’ (Wagner-Pacifici and Schwartz 1991), ‘agents of memory’ (Vinitzky-Seroussi 2002), or ‘memory choreographers’ (Conway 2010a).

Others have stressed that ‘national memories’ or ‘collective memories’ are never unitary, but divided across structural inequalities and often imposed from above. For instance, the Popular Memory Group at the Birmingham Centre for Contemporary Cultural Studies (CCCS) highlighted the importance of vernacular articulations of memory beyond historical sources as told by and for elites. Reflecting the Birmingham School’s wider interests in ‘mass’ culture, the Group called for an expansion of sources of ‘memory’ to include popular culture such as commercial works of biographies and autobiographies, fiction, television, radio, and film, as well as practices such as oral history and community publishing (CCCS Popular Memory Group 1982; Seaton 2007). In a similar vein, Schwartz and Schuman (2005) have noted that there are patent discrepancies between how elites portray historical events and figures, and how the masses remember them through individual recollections, suggesting that official institutions of memory such as commemorative ceremonies, national holidays, museums, monuments, and history textbooks are not necessarily reflective of civic, public, and popular memories more generally. It is not surprising that memory scholars have become increasingly alert to the need to study ‘bottom-up’ forms of commemoration and remembrance in addition to ‘top-down’ models which have tended to dominate influential studies of the past (Winter 2010: 317).

Indeed, a recognition that collective memory is reproduced along uneven distributions of economic, political, military, cultural, and social resources is critical for understanding why some forms of commemoration are possible but not others (Conway 2010a). For instance, public remembrances of historical atrocities may be banned by the state, as is the case for the remembrance of the 1989 Tiananmen Square massacre in China. Conversely, mass protests and collective civic action can overturn (quite literally) pre-existing symbols of the past through dramatic actions. Following the police murder of George Floyd in Minneapolis, Minnesota, in 2020, the Black Lives Matter protests led to the toppling of the statue of slave trader Edward Colston in Bristol, UK, and the removal of statues of Confederate leaders across the United States.

While the conflict between elite (‘top-down’) memories and mass (‘bottom-up’) becomes most visible at flashpoints such as these mass demonstrations against public ‘sites of memory’, it is important to note that elite and mass memories are not always contradictory forces. As this book argues, how ‘elite’ and ‘mass’ memories concur or clash with one another is an empirical, rather than a theoretical question. Ordinary people’s willingness to challenge dominant representations of the past, whether at an individual or collective level, is dependent on specific contexts and variables including the size and power of counter-hegemonic groups and movements, the salience and the relevance of historical events and social issues at a specific given time, as well as ‘political opportunity structures’ such as the level of state repression, the level of pluralism among politicians, and ties between elite and ordinary actors.

Collective memory is not always conflictual, and indeed, as Durkheim originally argued with reference to religious rituals as a site for celebrating social unity, commemorative actions can often result in a strong sense of moral cohesion. This is especially true in the aftermath of terrorist attacks, when there is often a strong congruence between state, media, and civic reactions to the violence. These actors often employ frames of good and evil to celebrate their own values, while calling on external threats to be exterminated (Simko 2012; Smelser 2004a; Wagner-Pacifici 1986). In the wake of Islamic extremist terrorist attacks across Europe and America, buttressed by enduring public support for the ‘war on terror’, states have successfully strengthened their hold on individuals’ behaviour, through greater investigative powers, enhanced security checks, greater surveillance in public spaces and online, and increased budget spending for security authorities.

This book does not seek to fully explain how and when ordinary people’s perceptions of the past concur or diverge with those of elite actors. Instead, it proceeds with a basic presumption that collective memory is not always consensual, nor is it always agonic: cultural ‘hegemony’ – the agreement between elite and mass perceptions of social reality – is a contingent, rather than a necessary phenomenon that is to be explained through the empirical study of specific case studies. As this chapter proposes below, a sociological model of collective memory must be attuned to questions of social differentiation (in terms of the division of labour), as well as social stratification (in terms of the uneven distribution of social resources that make commemoration possible). Moreover, the model must be attuned to the direction of commemorative speech acts and discourses, in terms of whether such discourses agree or disagree with representations of social reality that elite actors seek to impose.

There are two distinct bodies of research on memory and social discourse that are conducive to developing such a theoretical outlook: cultural trauma theory, on the one hand, and Bakhtin’s literary theory on the other.

Emerging concurrently, and in conversation with, wider debates on the relationship between collective memory and collective identity, cultural trauma has emerged as an influential and prolific research programme for studying the long-term, intergenerational trajectories of intensely negative events such as wars, genocides, terrorism, and other acts of violence. Cultural trauma is defined as an instance in which a collectivity experiences an ‘acute discomfort entering into the core of the collectivity’s sense of its own identity’ (Alexander 2004b: 10). Cultural trauma developed as an explicit critique of ‘naturalistic’ theories of trauma which hold that entire collective groups such as ethnic groups or nations are automatically traumatised by horrific events. Instead, cultural trauma arises as a result of concerted efforts to attribute moral meanings to past events through narratives: ‘cultural traumas are for the most part historically made, not born’ (Smelser 2004b: 37).

The cultural sociologist Jeffrey Alexander identifies a four-step ‘trauma process’ by which stakeholders socially construct cultural trauma (Alexander 2004b: 12–15). Firstly, the nature of the pain and hurt must be identified; secondly, the victims must be named; thirdly, the relationship of the victims to wider collectivities must be established, and fourthly, responsibility and blame are attributed to perpetrators. Cultural trauma emphasises the fundamental role of symbolic representations, and the presence of supra-individual, cultural schemata (Alexander 2004a: 200): the events need to be ‘coded’, ‘weighted’, and ‘narrated’ according to binary oppositions such as victims and perpetrators, good and evil, purity and impurity, sacred and profane (ibid.: 202–4). Various ‘carrier groups’ of trauma such as public intellectuals, the state, the mass media, and survivors of violence play instrumental roles in relating the moral significance of an event to indirectly implicated audiences with no first-hand experience of it. When carrier groups are successful in the ‘symbolic extension’ from those directly affected by the event to wider audiences, cultural traumas can become important markers of cultural identity. As carrier groups and institutions successfully routinise these trauma narratives through commemorative processes, cultural traumas come to occupy an integral position in collective memory and collective identity. Cultural trauma has produced numerous and wide-ranging works on the importance of horrendous events in shaping collective identities, from slavery in African American identity (Eyerman 2001) and Holocaust memory (Alexander 2004a), to Japanese war memory (Hashimoto 2015; Saito 2006) and natural disasters (Eyerman 2015; Geilhorn and Iwata-Weickgennant 2017).

As a theory that emphasises the constructivist and contingent nature of commemoration, cultural trauma has been particularly successful in demonstrating how historical events result in suppressed or delayed articulations of trauma. For instance, Bartmanski and Eyerman (2011) show that the mass murder of Polish soldiers and civilians in the Katyn Massacre by Soviet forces during the Second World War failed to emerge as a national trauma in Poland until after the fall of the socialist regime in 1989, due to political repression and the weak political status accorded to victims. In drawing attention to cases of delayed or failed commemoration, cultural trauma provides a useful alternative approach to existing studies which have tended to privilege successful cases of commemoration (Conway 2010b: 448).

Despite its utility, cultural trauma suffers from two shortcomings which arise from the insistence that narratives, rather than broader processes of commemoration, should be the primary object of sociological analysis. The first problem relates to the analytical separation of ‘values’ from ‘interests’. For cultural trauma theorists, the struggle over symbolic ‘values’ takes place on a different plane to that of the struggle over material ‘interests’. As such, cultural trauma theorists insist that symbols and narratives need to be analysed on their own terms, rather than as reflections or instruments of economic and political interests. While the analytical separation of (symbolic) ‘values’ from (material) ‘interests’ is certainly a useful one, it is less useful when applied to investigating what certain forms of narratives and commemorative speech acts accomplish socially and politically. As discussed above, commemoration can never be entirely separated from questions about power and stratification: every commemorative speech act must eventually answer the mundane but inevitable question: ‘who pays for what?’ Jay Winter (2014) emphasises that commemoration entails a distribution of resources, such as construction materials, money, specialist knowledge, labour, and time. Who buys the land on which a memorial is built? Who pays for the architects, builders, landscapers, cleaners, and (if required) security guards? Who will clean, maintain, and repair the site of memory? Who organises and attends regular commemorative ceremonies? Winter notes through a discussion of First World War memorials that,

[h] owever sacred the task of commemoration, it still touched all the chords of local loyalties, petty intrigues, favouritism, apathy, and indifference. It also was about contracts, payments, and profit…The business of commemoration was always that: a business, shaped by the character of the community that undertook it.

(Winter 2014: 90)

If one acknowledges that all forms of commemoration are products of social resources and present interests of various stakeholders in one way or another, it follows that one must study these flows of material interests, political agendas, and commercial incentives in conjunction with, rather than separate from, the competition over symbolic representations.

The second problem arises from the first, in that the caesura between the ‘material’ and ‘symbolic’ worlds forecloses an analysis of the potential ‘meanings’ of the various commemorative media at the physical, material, and spatial level. To date, many cultural trauma theorists have been primarily interested in the ‘meanings’ and ‘narratives’ expressed through different media. However, this has meant that a consideration of the unique properties of different modes of commemoration which act as vessels of these narratives is often lacking in these analyses. To return to Marshall McLuhan’s (1964) adage that the ‘medium is the message’, it is not simply the content of commemoration that requires analysis, but also the material medium that is itself meaningful for the reproduction of a particular memory narrative. As such, it is imperative to consider the material possibilities and ‘agency’ of different mnemonic tools and practices by considering how these vary in cost, duration, and affect.

For example, commemorative ceremonies on anniversary dates can invoke powerful emotions through vivid language and iconography. However, the duration of such heightened emotions through ceremonies is short-lived. As such, routinisation and professionalisation of commemorative ceremonies is necessary for them to have a more enduring effect on collective identity (Olick 2016: 57). By contrast, physical commemorative objects such as memorials and monuments are often costly in terms of time, money, and space. The spatiality and design of the site of memory may itself be meaningful; Wagner-Pacifici and Schwartz (1991) have highlighted how certain memorials such as the Vietnam Veterans Memorial act as sites of ‘dissensual’ memory politics, in which the physical design of the memorial site enables individual visitors from different political orientations to derive different meanings from them.

In short, materiality matters. Every commemorative speech act, from holding a minute of silence to constructing a monument in a public area, imposes a particular vision of the ‘social organisation of time’ (Zerubavel 2003). Questions such as whether the commemorative medium is a human being or a material object, or what is the colour, size, design, and material of the object are pertinent for exploring the ‘meanings’ of sites of memory which are not reducible to the content of ‘narratives’ as such. Cultural trauma theory’s focus on abstract ‘(meta)narratives’ and ‘cultural codes’ unfortunately neglects how materiality itself is a meaningful object of analysis in the circulation and reproduction of social memory.

If an analysis of commemorative ‘narratives’ in isolation is insufficient, how can one be more sensitive to these interaction between ‘values’ and ‘interests’ as well as to the meaningfulness of different commemorative media? Mikhail Bakhtin’s three concepts of ‘dialogue,’ ‘polyphony,’ and ‘heteroglossia’ provide insights for capturing the conflictual and dynamic nature of commemorative speech acts. To be sure, this is not the first attempt to develop Bakhtinian concepts in memory studies. Notably, Jeffrey Olick (2007; 2016) has articulated a sophisticated theoretical model inspired by Bakhtin, Bourdieu, and Elias, arguing that commemoration is ‘always fundamentally dialogical’ (Olick 2007: 82). Olick’s (2016) study of official German commemorations of the Second World War illustrates how actors address political issues surrounding the question of commemoration in a path-dependent manner, as political leaders respond to preceding events, official statements, and evolving political priorities. Highlighting a different aspect of Bakhtin’s work, the anthropologist Andrea Smith has developed Bakhtin’s concept of ‘heteroglossia’ – the co-presence of official and lay discourse – to explore how differentials in class, status, and ethnicity shape how contradictory official and lay discourses are co-present in how people narrate their autobiographies (Smith 2004). Contrasting French official discourses of ‘race blindness’ with French settlers’ actual experiences of racism, Smith highlights how different social contexts shaped whether interviewees would refer to official or lay discourse: ‘official voice was most widely used in large gatherings of relative strangers, whereas the contrasting voice was especially apparent in more intimate settings with only a few present, and never in mixed French – non-French company’ (Smith 2004: 261). Smith argues how inequalities in race, ethnicity, class, and status can shape internally contradictory – but by no means incoherent – memories of the past.

While these studies have been helpful for highlighting different aspects of commemoration and collective memory, this chapter proposes that the concepts of ‘dialogue’, ‘polyphony’, and ‘heteroglossia’ cannot be applied to collective memory in isolation, but that they must be considered in tandem. The concepts of ‘dialogue’ and ‘polyphony’ are perhaps the most well-known of Bakhtin’s theoretical formulations. Drawing on dialogue as an everyday practice, Bakhtin drew attention to how some literary genres emphasise the interaction of multiple voices expressing differing opinions.

For Bakhtin, whereas the epic poem represents the protagonist’s point of view through an authoritative narrative voice, the novel – as exemplified by the works of Dostoevsky – represents a new literary genre in which multiple characters express contrasting, but equally legitimate, viewpoints (Bakhtin 1981; 1984a). Dialogue differs decisively from the concept of Hegelian dialectics or Habermas’ concept of communicative action. Unlike dialectics and communicative action, in which two opposing perspectives are resolved to a superior argument, dialogue does not presuppose an end point (Sennett 2012). Instead, dialogue treats difference as a given; it continues indefinitely as a response to what has been said before, in anticipation of what will be said in the future. As a result, dialogue exposes a state in which different consciousnesses mutually illuminate one another, leading to a greater understanding between participants that does not result in agreement or a Hegelian synthesis (Bakhtin 1984a: 97). Although dialogue does not necessarily result in a resolution, it does anticipate social change. Since dialogue does not have a logical conclusion, any attempts to fix meanings are always open to challenge, rebuttal, and reinterpretation. Bakhtin stresses that no single participant gets the ‘last word’ (Bakhtin 1986: 170).

In Bakhtinian discourse, since multiple viewpoints are not fully resolved, dialogue always results in polyphony, a situation which he defines as a ‘world of a multitude of objectively existing and interacting psychologies’ (Bakhtin 1984a: 37). Although intimately connected, there is a subtle difference between dialogue and polyphony that makes them conceptually distinct: dialogue highlights the sequential nature of an interaction between different points of view; in contrast, polyphony highlights the co-presence of multiple opinions at a given time. Thus, dialogue can be understood as the process which gives rise to polyphony as a product.

In contrast to Bakhtin’s better-known concepts of dialogue and polyphony, the concept of ‘heteroglossia’ has arguably received less attention in memory studies. Heteroglossia refers to the many ‘accents’, or ‘speech genres’, that exist according to existing social hierarchies. Within a linguistic system, there are many speech genres, that range from official state discourse, professional jargon and urban speech to regional dialects, polemic, and everyday language (Bakhtin 1981: 263). Like the Marxist philosopher Antonio Gramsci, who was similarly sensitive to the relationship between class and language, Bakhtin emphasised that inequalities were constitutive of social discourse. Different forms of speech genres are distributed according to differences in social characteristics, such as between official and non-official discourse, professional and lay language, and city and country.

This perspective is more pronounced in works by V.N. Voloshinov – one of Bakhtin’s collaborators – who maintained that ‘[c] ommunication and the forms of communication may not be divorced from the material basis’ (Voloshinov 1986: 21).¹ For Voloshinov/Bakhtin, as different social groups, each with their own ‘accent’ (speaking with multiple ‘accents’) speak to one another, language itself becomes ‘an arena of the class struggle’ (ibid.: 23). Bakhtin highlighted that each stratum of language has a certain ‘direction’. On the one hand, official, ‘sacred’, ‘high’ discourse is characterised by a ‘centripetal force’: an attempt to bring ideological unity and to represent social reality in the interests of the ruling class. As with Gramsci’s concept of hegemony, in which the existing social order is accepted by the masses as natural and inevitable, Voloshinov/Bakhtin argued that ‘[t]he ruling class strives to impart a supraclass, eternal character to the ideological sign, to extinguish or drive inward the struggle between social value judgments which occurs in it, to make the sign uniaccentual’ (Voloshinov 1986: 23). For example, elite discourses such as legal documents, professional documents and academic papers seek to make generalisable and all-encompassing statements about entire populations across time. In commemorative speech acts, one can often find centripetal narratives during wars and following events such as natural disasters, terrorist attacks, and pandemics, when state leaders and figures of political and cultural authority call for unity amidst a ‘national emergency’ and potentially overpower dissenting voices through coercive state institutions.

On the other hand, multiple speech genres or ‘accents’ that exist across society as ‘low’, ‘mass’, ‘profane’ discourse resist attempts by elite strata through a ‘centrifugal force’, leading to a dispersion of meaning through resistance, satire, and parody. Unlike Gramsci, for whom the masses’ perception of the world was obscured by hegemony and ‘common sense’, Voloshinov/Bakhtin held that the opposition of speech genres across social divides provides linguistic signs with ‘vitality and dynamism and the capacity for further development’ (Voloshinov 1986: 23).

Although heteroglossia highlights the interaction of linguistic discourses between multiple strata as a relatively normal condition in society, Bakhtin recognised that the social differentials can occasionally be subverted at a societal level. In what can be considered a case of ‘centrifugal’ discourse that destabilises centres of social power, in Rabelais and His World (Bakhtin 1984b), Bakhtin draws attention to the ‘carnivalesque’ as a temporary state of the suspension of social hierarchies and norms, a type of ‘anti-structure’ (Turner 1982). The carnival, much like Durkheim’s religious ritual which results in a moment of heightened emotions which he called ‘collective effervescence’ (Durkheim 2008), is an occasion in which the social differences that separate individuals temporarily break down, resulting in a moment of unity. During the carnival, social norms and expectations are subverted. ‘Lowly’ bodily functions of eating, excretion, and sex are on full display. Satire, parody, and blasphemy become permissible. The sacred becomes profaned, and the hidden becomes apparent. By upending existing cultural norms and social hierarchies, and as opinions are exchanged between people as equals, carnivals dissolve existing social structures, albeit temporarily.

Bakhtin’s attention to the innate capacity of lay actors to engage in parody, satire, and other forms of resistance means that heteroglossia provides a more dynamic model than Gramsci’s, in which ordinary speech is reduced to a reflection of elite interests and laypeople are assumed to be ‘cultural dopes’, unaware of their true class interests until prompted by ‘organic intellectuals’. Heteroglossia stresses that, while lay actors are still subject to centripetal speech genres, they also have the innate capacity to challenge, parody, and attack these representations through centrifugal speech. Moreover, applied to the study of commemoration, an explicit consideration of heteroglossic aspects of commemoration overcomes a common tendency to equate commemoration to ‘elite’ and ‘official’ discourses by taking into consideration non-elite and subaltern actors in the creation of speech acts.

Having considered these theoretical perspectives, this chapter concludes with three theoretical propositions that undergird the following chapters.

Collective memory is not a rigid and stable social structure, nor is commemoration an unreflexive reproduction of knowledge about what has happened in the past. Nor can collective memory be reduced to an aggregate of individual psychological reflections about the past. As Halbwachs highlighted, cognition – or rather, recognition – of the past is always social, filtered through existing structural constraints, including language, social norms, kinship and friendship groups, and, as this book highlights, the mass media. Moreover, the style and content of specific commemorative media are dependent on cultural and historical factors that determine the discursive rules of what can and cannot be said.

Rather than treating collective memory as an unchanging social structure, collective memory is best understood as a shorthand to describe the collection of competing claims, symbolic representations, and social practices pertaining to shared understandings of the past. Some of these claims, when successfully shared across the population, have a structuring effect on individual perceptions of the past: they may even become taken for granted explanations about the origins and history of a society or a people. Thus, while this book has ‘Japanese collective memory’ as its focus, it does not mean that there is a permanent and objective social structure that all Japanese people are conditioned to accept. Rather, ‘collective memory’ refers to constellations of social practices that create and reproduce certain interpretations about what it means to be ‘Japanese’ across multiple generations. Collective memory is neither a purely ‘objective’ nor an impartial reflection of the past; as such, it is analytically distinct from ‘history’.

Commemorative speech acts are concerted social efforts by individuals, groups, and institutions that create and reproduce partial interpretations of the past through a variety of mnemonics devices and practices across disparate social groups and across generations. Because there are always many ways to interpret the past, there is never a single ‘collective memory’; even if certain narratives and symbolism become hegemonic, one can still expect to find multiple and mutually exclusive narratives and symbolism relating to the past.

Given that different social groups inevitably have different understandings of the past. especially regarding the factual, normative, and affective importance of historical events, commemoration is inherently dialogical – in the sense that each intervention responds to what precedes it – and polyphonic – in the sense that there are multiple perspectives coexisting in the same discursive space: this means that commemoration is always provisional and open to challenge (Olick 2016).

Commemorative speech acts, whether of positive or negative events, rely on symbolism to represent particular interpretations of the past with reference to fundamental values and norms. Symbolic expressions can range from physical sites of memory, such as monuments and museums, to mediated forms of communication, such as mass media, artwork, books, academic work and commemorative processes that rely on embodied practices such as oral history, commemorative ceremonies, demonstrations, music, and dance.

As cultural trauma theorists have argued, these representations of the past very often rely on a foregrounding of binary cultural codes consisting of positive and negative oppositions: good and evil, victims and perpetrators, sacred and profane, pure and impure. As such, sites of memory are replete with moral meanings and inseparable from appraisals of the past, making commemoration an intensely political struggle over who occupies the side of the ‘good’, ‘sacred’, and ‘pure’. As commemoration relies on the reification and celebration of the ‘in-group’ as being on the side of the ‘good’, it can also intensify representations of certain groups as ‘evil’, ‘profane’, and ‘impure’, thereby justifying exclusionary or discriminatory behaviours and policies against those who are identified as ‘perpetrators’ and the cultural ‘Other’. As Alexander and Breese (2011: xxxiii) note, cultural trauma construction does not necessarily result in civil repair; instead, it can lead to enduring social division, as cultural codes of good and evil, pure and impure, and victims and perpetrators become entrenched in civil discourse between opposing groups.

At the same time, cultural practices are meaningful beyond their expressions of abstract binary values through narratives. The materiality of commemorative media, such as the design, colour, and material of a memorial object, are themselves affective qualities even if they may not relate directly or explicitly to ‘narratives’ or ‘values’. Similarly, the very presence of a memorial object demonstrates a willingness to remember, while its absence may suggest either an inability to remember or a desire to forget and exclude a certain event from public memory.

Many types of commemoration, from ceremonies and rituals to the construction of monuments and museums, are resource intensive. Commemoration requires a range of social resources such as time, capital, labour, physical space, access to media, political power, and social ties. These social resources are unevenly distributed across society, meaning that those with more social resources are likely to be able to engage in commemorative speech acts that are high-impact and far-reaching. For example, political elites hold the power to enact legislation, hold military parades, determine educational curricula, and set official holidays. Rich and powerful corporate actors such as mass media conglomerates can disseminate narratives and imagery that reach national or even global audiences. Economic elites may engage indirectly in commemorative speech acts through patronage and financing the creation of monuments, museums, and archives, or by funding research in a field of study related to distinct peoples or past events. These elite-controlled commemorative speech acts usually constitute the ‘dominant’ representations, or the ‘master narratives’ of collective memory that have structural effects on individuals’ perceptions of the past.

As Bakhtin (among others) recognised, elites tend to use language in a ‘centripetal’ manner that neutralises and justifies the current social order. However, this is not always the case for all elites. Those in relative positions of privilege, but without direct access to political power – artists, novelists, academics, and journalists, to name a few – can challenge these dominant representations through counter-narration, by ‘speaking truth to power’. These elites in intermediate positions can destabilise and challenge established meanings through ‘centrifugal’ discourse. Hence, social discourse cannot simply be reduced to an opposition between ‘elite’ and ‘lay’ discourse: discourse is multi-layered into numerous substrata, each of which can be oriented in ‘centripetal’ (towards the political and linguistic centre) or ‘centrifugal’ (towards the political and linguistic periphery) directions.

While social differentials of power have structuring effects on lay actors’ perceptions of the past, they do not determine the outcomes of commemorative practices. It is indeed often the case that ordinary populations will be influenced by elite representations by the state and media, and history is replete with instances in which national populations are prone to propaganda about the mythical origins and the superiority of nations and ethnic groups. Yet, this is not to suggest that ‘ordinary’ people cannot challenge or question established narratives and frames. Even when lacking access to social resources, subordinate strata can still rely on different commemorative media with relatively lower barriers for entry, such as social media, public demonstrations, self-publishing, and commemorative ceremonies. The power of subaltern groups to engage in counter-memory means that ‘elites’ never truly have a monopoly over the meanings of the past. Whilst those at the centres of power enjoy a structural advantage in shaping collective memory within a society, given the centrifugal aspects of language among peripheral groups, ‘there can be no consensus’ of collective memory at any given time (Schwartz 2012: 529). Ordinary people are not under the influence of ‘hegemony’ or ‘common sense’ by default. Rather, elite ‘master narratives’ are always vulnerable to resistance, parody, and challenge. Every utterance, whether it comes from the elite or subaltern, anticipates a response; dialogue never ends.

The following chapters seek to demonstrate how each of these three principles is crucial for understanding the complex relationships between various actors involved in the narration and commemoration of the Aum Affair. The Aum Affair was a landmark moment for Japanese society, raising fundamental questions about the state of current society. However, perhaps more importantly, it was an event that exposed the fault-lines of who constituted the Japanese ‘us’ and the non-Japanese ‘them’. These divisive categorisations, operating at the levels of civil society, religion, and ethnicity, have endured long after the immediate impact of Aum’s violence has faded.