Collective memory: An hourglass between the collective and the individual

Abstract Collective memories are memories shared by a group that influence their social identity. The goal of this paper is to focus on two major limitations in current studies on collective memory and show how the hourglass metaphor can overcome those limitations. The first limitation concerns the partial nature of studies devoted to the analysis of collective memory. Studies tend to focus either on the choice of the past (how memory agents mobilise the past) or the weight of the past (how the past affects the individual or the group). The second limitation relates to the temporal dimension of research conducted so far. Most studies only assess memory over a single generation, yet it can have long-term effects. In this paper, we suggest considering memory work as an hourglass, with the collective and the individual at opposite ends and the sand of memories passing from one to the other, filtered through family values and representations. The hourglass metaphor thus provides a helpful tool to explain the formation of collective memories over time and the interactions between the macro, meso, and micro levels. We approach the study of collective memory from an interdisciplinary perspective, mainly involving psychology, political science, and history. We conclude by suggesting three challenges that future studies of memory will need to address: (1) the need to combine multiple approaches; (2) the need to consider the role of generations; and (3) the need to bridge discussion across disciplines.

history, anthropology, philosophy, and communication studies. These fields investigate 'memory in the wild' and most experts in these fields stress that memory cannot be an exact and perfect reflection of the past: it is only its trace or evocation (Kensinger 2009;Noiriel 2004). Yet the links between memory in the head and in the wild remain unclear (Barnier and Hoskins 2018). The study of collective memory resides at the intersection of these two worlds, where individuals and the social context in which they live collide to form an ensemble of representations and memories shared by a group. Throughout this paper, we argue that both perspectives need to be considered jointly in order to fully grasp the scale of how collective memories are formed, transformed, and transmitted, as well as how they impact the individual and the collective.
Memories are not literally preserved, but they are reconstructed according to present conditions and the social context. It is from this premise that Halbwachs (1997Halbwachs ( [1950) developed the concept of collective memory. Contesting the notionwhich he considered an impossibilityof isolated individual memory, the French sociologist emphasised above all the influence of the social on the content of individual memories. In Les cadres sociaux de la mémoire (1994 [1925]), he argued that, over the course of an era, it is a groups' shared beliefs and its collective experiences which shape the meaning of individual memories and not the other way around. When the notion of collective memory resurfaced at the end of the 1970s, it was in the realm of history not sociology. According to Nora, collective memory is 'what remains of the past in the groups' experiences, or rather what these groups do with the past ' (1978, 401). Since then, the notion has gradually spread throughout the humanities and social sciences (Baussant 2002;Gluck 2007;Hirst and Manier 2008;Kim 2016;Kurze and Lamont 2019;Olick et al 2011;Winter 2001).
Despite the wealth of existing literature, two major limitations remain in collective memory studies. The first concerns the relatively partial nature of many studies. They generally focus on only one of the two dimensions that constitute the subject under examination (Lavabre 1994). The first of these is the choice of the pastwhich refers to how memory agents strategically mobilise the past (Assmann and Conrad 2010;Foong Khong 1992;Langenbacher and Shain 2010;Todorov 1995) and the second is the weight of the pastwhich investigates the traces or imprints left by the past on individuals and groups (Bell 2010;Davoine and Gaudillière 2006;Rosenblum 2009). The hypothesis underlying this paper is that it is crucial to understand the articulation between the macro level, most often reduced to the strategic dimension (the choice of the past), and the micro level, which has been often analysed through the prism of 'trauma', in the broadest sense (the weight of the past) (Margry and Sanchez-Carretero 2011;Resende and Budryte 2014). We argue that another important dimension also lies within the meso level, the smaller groups in which individuals interact, exchange stories, representations, and narratives. Within the scope of this paper, we examine the meso level through the role of families and how the family values and representations act as a memory and information filter. 1 We are not the first researchers to call for more studies to examine the interaction between the different levels of memory (e.g., Heinrich and Weyland 2016; Van de Putte 2021). Among those who have made similar calls, communication studies researchers Keightley, Pickering and Bisht (2019) have suggested an 'interscalar approach' to the study of memory, which takes experience as a starting point before analysing the different communicative processes between the individual and collective dimensions. For them, remembering at different scales is produced by 'mnemonic imagination', a mechanism that combines memory and imagination. This suggestion is closely related to current arguments in psychology, which assert that memory uses the same reconstructive process as imagining the future or considering counterfactual versions of past events (Cordonnier et al 2018;Schacter et al 2007).
This is but one of many good examples that demonstrate how different research disciplines working in the field of memory would benefit from collaboration. Consequently, in order to examine and comprehend the intertwining mechanisms operating at the collective, interpersonal and individual levels, multi-disciplinary work is essential. Only by combining expertise and methodologies is it possible to identify phenomena related to the instrumentalisation of memory, its relational, cognitive, and emotional aspects, as well as its possible distortions over time.
The second limitation relates to the temporal dimension of collective memory research. Understanding the mechanisms of memory transmission and evolution necessarily implies a broadening of the timescale (Joutard 1977). Rather than limiting itself to the study of one generation of actors, future research should attempt to identify tensions, gaps, and even contradictions from one generation to the next, as well as investigate the active role that different generations take in the sharing of memories and knowledge. Although there have already been studies that have examined intergenerational transmission within families or in schools (Billaud et al 2015;Buono 2004;Fivush et al 2008;Leonhard 2002;Stone et al 2014), very few, so far, have tackled the articulation of memory at macro, meso, and micro levels over three generations (one of the rare examples being Welzer 2005).
In order to counter these limitations, we need more interconnected, ambitious approaches that strive to combine the methodological tools and analytical lenses of multiple disciplines to interpret the multifaceted concept that is collective memory. Within this article, we propose the hourglass metaphor to facilitate discussions across disciplines that wish to investigate the formation, transmission, and evolution of collective memories over time, as well as the interactions between the macro, meso, and micro levels.

Memory work as an hourglass
The malleability of memory is one of the main objects of consensus across all fields of memory studies. Memory functions in 'a shadow land of nuance and subtlety' (Storrie 2014), depending on the needs and requirements of the present. Indeed, the notion of collective memory is based on the interactions between public narratives of the past (whether presented by official representatives or historians) and individual memories (experienced and/or transmitted by the population) that operate according to their own logic but affect each other. Their interactions can be approached in a top-down manner, examining how the weight of the collective bears on the individual, or from a bottom-up perspective, by assuming that extant collective memory must take its root in the collection of individual memories (Hirst and Manier 2008). In this sense, we suggest that memory research can be seen as an hourglass, with the collective and the individual at opposite ends and the sand of memories passing from one to the other, depending on the way it has been turned (see Figure 1). In the next section, we will start by briefly describing the metaphor, before delving into each approach in more detail.
Beyond the apt allegory of the passing of time visualised as the accumulation of sand, both the shape of the hourglass and its invertible nature can be used as a representation of our approach. At the top of the hourglass lie the historical, political, and social contexts (macro). This is a large area that combines an ensemble of predominant collective notions, beliefs, and perceptions of the past. In the centre, the narrow passage between the two bulbs represents the particular case of family memory (meso). It acts as a filter, a meeting point between the official (or public) history and the lived (or private, intimate) history. Its filtering capacityrepresented by the width of the neckchanges depending on the peculiarities of each family, but also the evolution of the role of the family across different generations. At the bottom rests the individual (micro). However, the individual is not an empty receptacle. In this broad space lies the individuals' idiosyncrasies, such as their personal beliefs, attitudes, or social identifications, all of which will influence the way they apprehend and reconstruct the past. But each grain of sandeach memorycollected will have a certain impact, direct or indirect, on the individual. This neatly reflects Halbwachs' notion of memory (1994Halbwachs' notion of memory ( [1925), where the individual is always dependent on the collective and the social framework in which it evolves.
Conversely, inverting the hourglass allows us to put the individual and their family at the top. After all, it is the individuals themselvesnot the groups to which they belongwho remember, and it is the weight of their memories that is brought to bear on the way history and societies represent the past. When discussing memories of the First World War, Winter has insisted that political leaders and academic scholars do not shape historical narratives; rather families are the ones doing the work of remembrance (Winter 2014, 166).
Finally, turning the hourglass on its side reveals its resemblance to the infinity symbol (∞), hinting to the never-ending loop between the two sides. Top-down and bottom-up influences interact continuously (Rothberg 2009), through discussions but also through various media, both traditional and digital. Memory is never set in stone. It evolves and develops, depending on the availability of information, the perception of the generation, the identity, the needs, the goals, the functions, and the roles of the individual and the collective. As Wang said: 'Just as one cannot separate a dance from the dancer, so one cannot separate a collective memory from the collective and its individual members who are the creators and carriers of the memory' (Wang 2008, 305).

Collective memory studies: Across disciplines and across generations
The hourglass metaphor suggests that, in order to obtain a well-rounded understanding of the interactions between the different levels (macro, meso, and micro), top-down and bottom-up approaches must be examined side-by-side. We argued above that the best way to tackle this challenge is to combine the expertise of multiple disciplines. For instance, whereas psychologists have excelled in developing paradigms to investigate the processes of memory through a bottom-up perspective, other fields of research in the humanities have established the importance of the roles played by the media, the elite, culture, and context in shaping collective memory and, thus, have mainly focused on top-down analyses.
Within the next section, we review what history, political science, and psychology can bring to the study of collective memory through the hourglass metaphor. It is important to note that by analysing these three branches of memory studies (that correspond to the disciplines of the authors), we do not wish to exclude any other discipline. Indeed, the contribution from other disciplines is highly valuable and welcome, but it would be pretentious of us to attempt to describe their expertise in the field of memory studies within the scope of this paper.

Looking through the hourglass: top-down approaches
Top-down approaches are generally interested in the content of collective memories shared across members of broad groupsthe most studied one being the nation (Liu and Khan 2021)but also in the way these memories impact society and its individuals.
They focus on what is common to the community depending on a group's characteristics and its cultural and historical context. The transmission of these collective memories generally occurs through traditional media, where the press can widely broadcast prevailing public opinions. Furthermore, collective memories can be reactivated or given more prominence depending on their presence in the media, for example, during times of commemoration (Kligler-Vilenchik et al 2014). This approach has been used extensively within history and political science, but also within social psychology.
Historians have argued that memory differs from history, but the former should not be viewed in direct opposition to the latter. In principle, history attempts to maintain an objective distance from its subject matter to explain past events through the scholarly analysis of mentalities and contexts. Memory, on the other hand, is inherently associated with social perceptions and beliefs within a group (Nora 1984). Its objective is not to recount events as they happened, but to create or maintain an identity (Bedarida 2003;Roekens 2006). This does not mean, however, that historical scholarship is not informed by memory and vice versa (Noiriel 2004). The selective and fluctuating character of memory should not be considered a negative attribute. It is, on the contrary, inherent to any situation in which a person resorts to memory: narratives of the past are constantly updated and adapted to present circumstances. Thus, following Pierre Nora's pioneering work, historians have interrogated the history of 'collective memories' as social representations of the past that have been maintained by communities of varying sizes. Using top-down approaches, they have endeavoured to identify the actors or social mechanisms responsible for shaping collective memories. For example, historians tracing 'the history of representations' or those working on 'new cultural history' examine how groups understand the world and events at a given time, and how their collective memories then impact the emotions and attitudes of group members. They analyse the power relations by reviewing how groups' representations, practices, and discourses influence the identities of the 'dominant' and the 'dominated'. In other words, they investigate how representations can be imposed from the top and also how individuals can modify and reclaim them (Chartier 2003;Veyne 2010). They do so through the analysis of a broad range of sources, including written, iconographic and artistic sources, and, more recently, online sources, such as social media.
Other scholars prefer to focus on official memory agents. In their seminal book The Invention of Tradition, Hobsbawm and Ranger (1983) discussed how the elites of the nation actively manipulate collective narratives for their own purposes. Political authorities can attempt to structure memory transmitted privately through textbooks, monuments, or commemorations (Citron 1987;Namer 1983;Warland 2019).
This instrumentalist approach to collective memory finds a similar resonance within the field of political sciences. Officials, including foreign policy actors, constantly reshape the meaning given to past events according to varying national and international circumstances (Rosoux 2001(Rosoux , 2019. Of course, the facts of the past cannot be altered. No one can undo or alter what has been done. But the meaning of what happened is never completely set in stone (Ricoeur 2000). A key concern in political science approaches has been the role of memory in conflict resolution. Most authors working in this field consider that only a change in the meaning given to the past can ultimately alleviate the emotional burden of painful past events (Bracka 2017;Krondorfer 2018). In the absence of such changes, some groups appear to be characterised by 'unfinished grief' (Baussant 2006;Tison 2011), reliving past events without end. What matters then is not to establish the truth with a capital 'T', but to elaborate narratives that have the potential to bring different groups closer together. The objective is to tell the story in another way, in other words, to tell it from the other's point of view (Ricoeur 2000). In this way, a plurality of memories can be established, turning the work of memory into the work of memories (see Irwin-Zarecka 2017;Olick 2003;Zerubavel 2003). This effort to open the past up to a plurality of representations appears to be the only way to elaborate a narrative that may not be shared by all but is ultimately 'shareable' (Passerini 2014).
Indeed, political scientists and, more broadly, experts in transitional justice often stress the importance of forms of recognition for past crimes, either through requests for forgiveness, official apologies or truth, and reconciliation commissions (Brooks 1999;Brophy 2006;Daase et al 2016;Lind 2008). Some argue in favour of a 'reconstructive approach' between representatives of various national groups in order to bring former adversaries together (Ferry 1996). This approach is based on actors assuming historical responsibility for past crimes in a critically measured manner. The official representatives of formerly war-torn communities are called upon to describe the wrongs suffered and committed, so that future generations can turn over the bloody pages of the past and work towards a shared future. From this perspective, the different types of media that reach and influence people widely (such as radio, television, and social media) are often seen as critical forms of communication that can help gradually develop a vision of the past which is acceptable to all sides (Gilboa 2009).
Although much research has been carried on the subjects of transitional justice and reparation politics (e.g., Butt 2008;de Greiff 2006;Kritz 1995;Thompson 2002;Torpey 2006), one particular element remains under-analysed: the long-term effects of a critical assessment of the past. The negotiation processes that lead to the recognition or nonrecognition of historical responsibility have been the subject of numerous analyses. But little in-depth research has been devoted to the impact of such gestures across not only one, but two or even three generations, even though all case studies demonstrate that the transformation of relationships between former enemies has to be understood in terms of generations, and not of years (Rosoux and Anstey 2017).
Finally, psychologistsand more specifically social psychologistshave also applied top-down approaches to their studies of collective memory mostly through the application of the social representation theory developed by Liu and Hilton (2005). According to these authors, large groupssuch as nations, political, cultural, and religious communitiesdefine themselves through shared representations of their history. Collective memory not only provides a strong social identity for groups, but also influences their beliefs, perceptions, attitudes, and behaviours (Pennebaker et al 2013;Tajfel and Turner 1986). In order to investigate these relationships, researchers using top-down approaches map out the social representations of history of specific groups by surveying a large sample of their members and then correlate these representations with the behaviours or attitudes expressed by the participants. For example, researchers have examined how the perceived victimisation of a group in history may still encourage its members to act in certain ways or hold negative attitudes towards other groups (e.g., Bar-Tal and Antebi 1992;Bouchat et al 2017).
Studies on social representations of history assert that collective memory serves three principal functions Liu and Hilton 2005). First, as already mentioned, sharing similar interpretations of the past helps create and maintain a feeling of belonging to the same group by providing examples of what the group is and how it should act. It thus serves the essential function of building social identity. The second function is social at its core. Through their shared representations of history, groups keep track of other groups they interact with and define these relationships: Are they friends or foes? What are the intergroup politics? The last function of collective memory is not only to provide frames of the interpretation of the world but also to help us make 'better' decisions in the future by learning from past mistakes.
Overall, communities draw on their histories to elaborate narratives that can unify them and guide their behaviour. In order to thrive, these narratives should have memorable characteristics, such as the presence of narrative patterns or distinctive protagonists (heroes, villains, or victims), and they also should be plausible, although factual accuracy is less important (László 2008;Olick 1999). They need to be widely broadcast through traditional media, while social and digital media have also become increasingly significant in allowing for the rapid and wide-ranging transmission of narratives. Some important events, such as foundational myths, can even become 'historical charters' that account for a group's historical mission but can be negotiated and amended over time often by the elites (Liu and Hilton 2005). It is, thus, through these top-down studies that we can perceive, evaluate, and investigate how the choice of the past weighs on the members of the group.

Looking through the hourglass: bottom-up approaches
Bottom-up approaches start with the individual and tend to focus on the processes that influence the creation, distribution, and the evolution of collective memories, as well as on the consequences of such memories have for the person. They question how the small, the insignificant, the discrete manages the weight of the past but also, at their own level, influences it. Especially since the rise of social media, individuals have found new platforms to counter dominant memories broadcast by mainstream media and foster new perceptions of the past (Birkner and Donk 2020).
Political scientists and historians have observed and analysed these discrepancies between official interpretations and individuals' perceptions of an event (Margry and Sanchez-Carretero 2011). In this regard, the plurality of representations of the Algerian War of Independence is an emblematic example. This war led to the definitive secession of an integral part of the French territory and the exodus of an estimated one million refugees of European descent alongside pro-French Muslims to the other side of the Mediterranean Sea. From 1954 to 1962, more than two million French soldiers served in Algeria. Approximately 20,000 of them died, while hundreds of thousands of Algerians were killed during the war (Pervillé 2004). These figures show the tragic impact of the war on individuals from both sides of the Mediterranean. However, in the French strategic narrative, this devastating conflict remained a war without a name for almost four decades (Gacon 1994). In refusing categorically to recognise the fighting in Algerian as a war, the French government caused a gap to develop between the sanitised narrative that was emphasised officially and public awareness of the violence that characterised the conflict. This example reminds us that, even if institutional actors can attempt to modify narratives of the past to fit their own objectives, the dissemination of a sole, monolithic historical interpretation cannot simply be imposed from above. Citizens exposed to official discourses are not merely empty vessels waiting to imbibe state-sponsored narratives without hesitation. Rather, they co-construct the messages conveyed to them.
In order to bring these alternative narratives of the past to light, oral historians have focused on what Assmann and Czaplicka (1995) termed 'communicative memory', which are collective memories based on everyday communications. These memories are neither formal nor stable, and they evolve and exist within individuals and their idiosyncrasies. Collections of oral testimonies can provide fascinating insights into how topics are represented and discussed within the broader population, as well as the specific emotions associated with them. According to Descamps, oral history is a useful tool to deal with the tension between memory and history, as it provides 'a vivid, psychologically and affectively sensitive memory' of an event alongside clear and rationalised historical knowledge of that same event (Descamps 2019, 85). This approach is clearly inspired by the pioneering work of scholars such as Thompson (1963) and his concept of 'History from Below' and Ginzburg (2015), who advocated for historians to reduce their scale of observation and to favour 'individual appropriations' and 'microstoria'. And now that these stories can be easily shared through social media, historians increasingly refer to the digital sphere to include these new sources in their analyses (Clavert 2021).
Experiences of trauma and tragedy have been a particularly fertile ground for the research of both historians (Audoin-Rouzeau 2013; Conan and Rousso 2013; Tison 2011; van Ypersele 2010) and political scientists (Cobb 2013;Cobban 2007;Hayner 2001;Pouligny et al 2007). In public debate, issues related to the legacies of war are often described in a binary way, i.e. remembrance versus oblivion. Nevertheless, one of the most fundamental interrogations common to this topic is not whether to remember or forget, but how to remember and forget in order to move forward. From this perspective, the critical question is not only 'what happened?', but alsoand above all -'what shall we do with the past?'. The challenge is immense. Despite the rich literature devoted to the topic, there is no standard list of specific techniques and ingredients for processing problematic past events and correcting historical injustices. However, there is a list of constraints that inevitably limit the room for manoeuvre. The most challenging obstacle is the irreversible nature of the crimes that were committed in the past. From this vantage point, it is worth questioning the notion of 'post-conflict'. This concept is widely used in handbooks, dictionaries, and encyclopaedias, yet the duration of so-called 'post-conflict' contexts remains uncertain. What are the basic criteria that determine when a conflict is over? How long do the notions of victim and perpetrator make sense? When does the victors/vanquished dichotomy lose its meaning? Until when are the labels 'occupiers/occupied' relevant? All these questions intersect with individual and collective levels of memory. Case studies indicate that the answers given to this series of interrogations determine, to some extent, individual perceptions (in terms of self-esteemor lack of self-esteemand otherness), as well as political positions (in terms of national interest).
Beyond these important questions, cognitive psychologists have focused their research on understanding the processes that explain how individuals can influence and construct collective memories. In order to do so, they have applied methodologies previously used to evaluate interpersonal communication (generally at the dyadic level) to bigger groups of people. They postulate that psychological processes which occur at the micro level can have repercussions all the way to the macro level. Four main processes are generally studied from a bottom-up perspective: how memories are reinforced (usually through repetition across time, e.g., Roediger et al 2009); how they are shared across networks (Coman et al 2016); how they are manipulated (consciously or not) by other people or contaminated by information from other sources (Maswood and Rajaram 2019); and how they are forgotten (Abel and Bäuml 2015;Cuc et al 2007). For example, researchers have applied a well-known psychological method used to evaluate how particular silences in individual recall can promote forgetting (the method is called retrieval-induced forgetting) in socially shared contexts (Coman et al 2009). Stone et al (2020) even tested this theory in a real-life situation by examining how the omission of particular topics from the speeches of public figures (in this case, the Belgian king) can actively promote collective forgetting.
One field of research that has been well developed in psychology using a bottom-up approach is that of flashbulb memories. Flashbulb memories are detailed memories of the reception context where one learns about an important, surprising and consequential public event (Brown and Kulik 1977;Luminet and Curci 2018). One of the main reasons why we form these flashbulb memories is essentially to be part of the collective. They allow us to align our lives with that of other members of the group and share memories even though we were not present at the event (Hirst and Meksin 2018). Social identity thus influences the formation and the retention of these particular memories , which, when shared, foster bonding across members of the groups. However, flashbulb memories can also be modified by the collective and the media. One famous example is that of George W. Bush's memory of when he first learnt about the 9/11 terrorist attacks (Greenberg 2004). On separate occasions, he described watching the footage of the first plane hitting the tower on television before walking into a classroom, yet this memory cannot be true as there was no such footage available at the time. A likely explanation is that he integrated the vivid and memorable images of the first plane crash, which were widely broadcast in the media in the days following the attack, into his own memory. This example shows how suggestible our memory can be, but also how the media can influence it (Schacter 2022).

Looking through the hourglass: families and generations
One strength of the hourglass metaphor is to suggest an additionalyet importantlayer to the study of collective memory: families and generations. Although families were central to Halbwachs' model of collective memory (1994Halbwachs' model of collective memory ( [1925) and the Assmanns' notion of communicative memory (Assmann 2009(Assmann , 2011, they have received far less attention than groups such as nations. Yet, they are important hubs of information and representations. Notably, through the sharing of family memories, parents, and grandparents convey to their (grand)children important moral and social values that are essential for the formation and maintenance of a cohesive family identity (Green 2019;Merrill et al 2019;Pratt and Fiese 2004). These values can act as filters for family members, which may lead them to question, reinterpret or reject information they receive. However, this filtering capacity of families depends on the perceived importance of their role in establishing values. Thus, they will differ depending on the culture of the family (Wang et al 2017) as well as on the evolution of the role of family through time (Radica 2013).
The notion of 'generation' has been studied by several sociologists (Mannheim 1952;Pichler 1994). These sociologists generally focused on the concept of 'social generations' (also called cohorts), which included groups people born around the same time on the premise that because of their common historical-biographical past, they would have a shared world view and a generational consciousness (Scherger as cited in Timonen and Conlon 2015). Yet, the concept of 'family generations' has been the subject of far less scrutiny in the field of collective memory, except for the large body of literature focusing on the multigenerational transmission of memories of the Holocaust. According to Hirsch (2012) to take only one examplethe notion of 'postmemory' describes 'the relationship that the "generation after" bears to the personal, collective, and cultural trauma of those who came before' (Hirsch 2012, p. 5). So far, however, work on postmemory has mainly focused on literary accounts of intergenerational transmission rather than systematic assessments of the processes and mechanisms underlying such transmission. In recent years, scholars have begun to look beyond the Holocaust as an area of study of intergenerational mnemonic transmission and undertake more systematic studies of other events that may be less emotionally charged. Some of this work has included personal events from times of conflict which are clearly historically relevant. For instance, Svob and Brown (2012) found that, in Croatia, transmission of memories between generations was more likely to occur if the events described took place during a war. In his study of 'third' generation Germans, Welzer (2005) found that grandchildren knew personal events from their grandfather's Second World War experiences, but they often misremembered them. This handful of studies demonstrates the originality and considerable potential of multi-disciplinary approaches to the intergenerational transmission of memories, especially given the current dearth of literature in this particular field.
In psychology, research on intergenerational transmission has thus far been restricted principally to values and trauma (Albert and Ferring 2012;Nauck 2001), with a few exceptions. Robyn Fivush and her team have closely examined family reminiscing and how it influences the way autobiographical memories are transformed into narratives of the self (Fivush 2008). They argue that 'children are born into storied worlds' (Fivush and Merrill 2016, 308) and that family stories, informed by cultural master narratives, influence the way children and adolescents come to construct their view of themselves and the world (Fivush and Merrill 2016;Fivush et al 2011). Similar to our perspective, this approach is interscalar. They propose that family stories can be divided into three ecological systems: micro (co-constructed narratives of shared events), exo (communicative family narratives) and macro (master narratives and cultural history). These different systems constantly interact and influence the construction of autobiographical memories. Yet, few other researchers in cognitive sciences have included family narratives within their studies of memory.
Finally, a few recent psychological studies have examined the intergenerational transmission of historical memories Stone et al 2014). They found that overall, the transmission of family historical memories of Second World War-related events was very limited and that, by the third generation, most family memories were lost or stripped back to the general gist without many specific details. Yet, even if detailrich memories are not transmitted easily, feelings of victimisation can seep from one generation to the next (Bouchat et al 2017).

Conclusion
The duration of the transmission processes observed raises several methodological questions. How can we assess the 'effects' and 'after-effects' of an important event for the community? How can we measure the transformation of representations from one generation to the next? How can we detect emotional and even unconscious processes? These questions underline the need to work with interdisciplinary teams of scholars, so that future research on collective memory can be studied from numerous angles, because as of now, different disciplines have mostly excelled in their own approaches. For a long time, historians have mostly focused on the macro and meso scales, leaving generally little space for the everyday person's role in the construction of collective memory. Political scientists have highlighted the functional role of memory for the community and shown how memory's malleability is a tool that can be used to achieve set goals. Psychologists, on the other hand, have mostly limited their scope to individuals and their personal memories, generally ignoring the broader cultural and societal contexts. In this paper, we have suggested to consider memory work as an hourglass, with the collective and the individual at two possible entry points and the sand of memories passing from one to the other, filtered through family memories, values, and representations. This metaphor poses several challenges for future research.
The first challenge is the need to examine side-by-side research from top-down and bottom-up perspectives. Indeed, as an hourglass needs to be turned upside-down from time to time, so does our understanding of the processes, content and context of collective memory. Interactions between the one and the plural are reciprocal: individuals always evolve within social frameworks, yet the collective is nothing without its members. One way to overcome this challenge is to work on long-term research projects that provide the necessary time to run different types of studies as well as to analyse their results side-by-side.
The second challenge relates to the importance of the passing of time in memory studies, and that this time should be counted in generations. On the one hand, social generations allow researchers to situate a person in time by specifying the social, political, and cultural context in which they have lived. It forces the researchers to take a nonnormative approach by considering the evolution of the norms and the emotions associated with a particular event through time. On the other hand, family generations provide an estimation of the personal importance of the event for the individual. For this reason, we believe that studies on the intergenerational transmission of memory are essential to developing our understanding of collective memory.
The final challenge is to bridge discussions across disciplines in order to take advantage of the diversity of expertise in the field of memory studies. Scrutinising the creation, transmission, and evolution of collective memory requires a deep understanding of psychological processes, while being able to ground these within broader political, social, and historical contexts. It requires not only time, but also a form of intellectual curiosity open to examining a memory from other vantage points and, to a certain extent, a form of intellectual humility that accepts the theoretical and methodological shortcomings of one's own discipline.
Funding. The authors are the members of a project titled 'The transmission of memories related to stigmatisation: Official and family memories related to collaboration and colonisation in Belgium' (RE-Member), funded by an UCLouvain Concerted Research Action grant . Valérie Rosoux and Olivier Luminet are both supported by the Fund for Scientific Research (FRS/FNRS).
Disclosure Statement. The authors declare that this manuscript is an original work and have not been published before, not is it being considered for publication elsewhere either in printed or electronic form. However, the original idea for this paper was presented in a grant proposal for an UCLouvain Concerted Research Action project titled 'The transmission of memories related to stigmatisation: Official and family memories related to collaboration and colonisation in Belgium' .
Competing Interest. The authors Aline Cordonnier, Valérie Rosoux, Anne-Sophie Gijs and Olivier Luminet declare that there are no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.