Architectural Forms of Collective Memory

memorative monuments have attracted interest from a variety of disciplines. In history of course, but also in geography, sociology, anthropology or political science, there are many works devoted to them (Rowlands & Tilley, 2006). However, these constructions also have a close link with memory and from this point of view it is surprising that social psychology has not paid more attention to them. This is even more disturbing since, for several years, the notion of collective memory (Halbwachs, 1950) seems to have taken a significant place in this discipline (Laurens & Roussiau, 2002; Liu & Hilton, 2005; Tavani, 2013; Rateau, 2009; Wagoner, 2015). Moreover, since Nora’s (1984) work, it must be acknowledged that commemorative monuments also have a link with the identity of those who recognize them as ‘places of memory’. These findings therefore suggest that it may be appropriate for social psychology to consider commemorative monuments as worthy objects of study. This work is an attempt in this direction by examining the possible links between the memory of a past event and the architectural forms of the commemorative monuments to that event. In 2015, the Moscow City Hall and the Gulag History Museum launched a competition for the creation of a commemorative monument dedicated to the victims of political repression in Russia from the period of the 1920s to the 1950s. The artists, designers or architects wishing to participate in this competition had to summarize their proposal in the form of a computer-generated image, a drawing or a photograph of a model showing their project as it would appear once completed. They were asked to write a text presenting or explaining their project, its symbolism and its links with the theme of the competition. These images and texts were posted on a website dedicated to the competition. The public interested in this competition gave its opinion by voting for one of the presented projects. We were able to obtain these images and texts and to analyze them in order to answer two questions: Do the texts written by the competition participants tell us anything about the collective memory that they share about the political repression that took place in their country? How has this collective memory been expressed when it is translated into concrete and visible commemorative monument forms?

the Council of People's Commissioners (Yakovlev, 2003). Therefore, as in the case of the Holocaust, millions of people were repressed, deported or executed not because of the acts they committed, but simply because of the categories to which they belonged (Applebaum, 2003). However, unlike the Holocaust, these categories were neither stable nor absolute. An ad hoc category of ' enemies of the people' was created, the contours of which varied throughout history. In fact, more than 70% of the victims of Stalin's terror were ordinary people who were neither part of the Soviet elite nor members of the Communist Party (Cohen, 2010).
The Great Terror corresponds to the period from 1937 to 1938. There were 1.7 million people arrested in only 17 months (1937)(1938), more than 700,000 of them were killed and 300,000 to 400,000 were sent to Siberia, Kazakhstan or other remote regions of the Soviet Union (Cohen, 2010). Even today the total number of victims of the political repression remains uncertain. According to historians, it varies from 5 million to 30 million (Etkind, 2013) or from 12 million to 20 million during the Stalin period alone (Cohen, 2010). Therefore, it is a historical event whose magnitude probably exceeds that of the Holocaust. While Nazi Germany's Holocaust was directed towards the destruction of the ' other', the Jew, Stalin's repression was directed, like suicide, against the Russian people themselves. This specificity has probably disrupted the implementation of the resilience mechanisms that should have occurred in a society that has experienced such a disaster (Etkind, 2013).
The speech on the cult of the personality (of Stalin) done by Khrushchev, the General Secretary of the Central Committee of the Communist Party of the Soviet Union in 1956 marked the beginning of the rehabilitation of political repression victims. But this rehabilitation process was interrupted during the time of Brezhnev (1964Brezhnev ( -1982, because the aim was to remove the issue of political repression from the public discourse. This question came back at the end of the 1980s, under the impetus of President Gorbachev, whose two grandfathers were victims of Stalin's era. This period was marked by the appearance of victims' testimonies in public spaces, the release of several films on this issue and the public actions dedicated to the memory of victims ('week of consciousness' in 1988). But it was not until 2011 that President Medvedev ordered the creation of a working group to formulate proposals to implement a program to perpetuate the memory of the victims. Finally, in 2014 President Putin ordered the creation of a monument for which the competition mentioned above was organized. It should be noted that the Russian authorities initiated the rehabilitation process that was just briefly described. We will see later that this point probably has its importance for the memorial construction based on a project from the analysed array.
Even now the attitudes towards the period of repression and towards the main political figures who may have been responsible for it still divide the Russian society. For example, the 'Last Address' movement is calling to construct the commemorative plaques engraved with the names of victims. In Moscow, the 'Return of the Names' action regularly brings people together around the Solovetsky Stone, 1 a monument erected in 1990 in front of the NKVD headquarters. But at the same time, several surveys show that the number of those who think that Stalin's repression was an 'inexcusable crime' is decreasing, while the number of those who claim to know nothing about that time is increasing. In addition, the opinion about Stalin developed favorably between 2001 and 2017. 2

Collective memory
We owe the term ' collective memory' to the sociologist Maurice Halbwachs (1950). Analyzing the formation of individual memory, this author emphasizes first of all the eminently social character of the human subject. Then, he deduces that most of our experiments take place in 'social frameworks' that we share with others (social belongings, social practices, language, etc.). But these social frameworks are also benchmarks that we use when we recall an event, and that individual memory is necessarily effected by the seal of social inserts. Hence, collective memory can be defined as the product of the convergent activity of individuals who remember the past as members of a group or a community. This work of recollection is not a mere restitution. It is a reconstruction in the present (Fraïssé, 2008), aimed to make the past comprehensible at the time of its actualization and compatible with the social identity of the group. That is to say that with the image that this group wants to maintain of itself. From this point of view, the collective memory of a group maintains a distant relation with the factual reality of the past insofar as it rests on mechanisms of selection and transformation (Bartlett, 1932) that allow its' adjustment to group identity.
In his writings on collective memory, Halbwachs (1950) repeatedly uses the notion of representation. When he proposes examples illustrating his purpose, we find the expressions such as 'representation of an event' or 'representation of the past'. It is interesting to note here that at the time when Halbwachs wrote it, he only had the notion of ' collective representation' (Durkheim, 1898). This notion presupposes the existence of world views shared by all members of a society, but Halbwachs attempts to explain that the different groups that constitute a society carry different collective memories. Since the theory of social representations (Moscovici, 1961;see Rateau et al., 2011 for reading in English) had not yet been formulated, Halbwachs could not see that the collective memory he tried to theorize is very close to the phenomenon of social representation and that the two notions can be easily assimilated (Licata, Klein & Gély, 2007;Liu & Hilton, 2005;Wagoner, 2015).
Initially Moscovici (1961) defines social representations (SRs) as 'universes of opinion' related to objects in the social environment of individuals. More precisely, the content of a SR can be expressed in terms of information, opinions or beliefs shared by members of a group about a given object (Moliner, Rateau, & Cohen-Scali, 2002). These constructions are the result of the socio-cognitive processes that individuals use to understand their social environment and to give it meaning. They are also the result of different communication processes (interindividual and collective) that lead to the emergence of consensus among social groups. But according to Moscovici, among the main processes involved in the emergence of a SR, is the ' anchoring' process that responds to a principle of cognitive coherence. When a SR emerges within a social group, it necessarily enters the cognitive universe that precedes it. In other words, a social group develops a new representation by ensuring that it remains compatible with its pre-existing SRs, norms and values. For example, when Moscovici studied the SRs of psychoanalysis (1961), he noted that among Catholics this psychotherapeutic practice was regularly compared to confession.
The first function of SRs is a function of interpretation and comprehension of the social environment. However, for several authors (Abric, 1994, Jodelet, 1989, Moscovici, 1961 they also fulfill an identity function by allowing members of a social group to differentiate themselves from other groups (we are different because we do not have the same representations as the others). But the SRs are also identity regulators (Deschamps & Moliner, 2012). It happens that social groups make them evolve in order to maintain a positive image of themselves. For example, Mugny and Carrugati (1985) found that if the parents of an only child consider that intelligence is an asset resulting from a process of parent-child transmission, that this is not the case for parents with two children. The latter prefer to believe that intelligence is an innate gift. It is obvious that, faced with the inevitable disparity in the intellectual capacities of individuals, parents with several children prefer to adhere to the theory of gift. This allows them to minimize their responsibility when they find that one of their children is less intelligent than the other.
According to Abric (1987), SRs are the culmination of a process of collective reconstruction of reality. But this work was carried out with respect for the pre-existing history, norms, values and beliefs of the social group (anchoring process). It was also conducted with the concern to maintain a positive social identity of the group (identity function). We can think that the 'representations of the past' which Halbwachs discussed may also be social representations. This is the thesis that Viaud developed by identifying three characteristics common to both phenomena. First, they emerge and are updated through social interactions that involve '… ordinary communication processes' (Viaud, 2003: 15). Moreover, they are determined by group affiliations that orient their contents according to identity logic. Finally, the two phenomena seem to be intimately linked to language, a natural language different from that of science, but a language equally capable of producing '… symbolic realities for individuals and social groups' (Viaud 2003: 15).
Let us add that among the various articulations between collective memory and social representations suggested by Viaud, there is one that has a particular interest for us. According to him, the memory could also be considered as …a practice on the past intended to give a representation of it (ibid: 24). From our point of view, the practice of the commemorative monument is in line with this logic, because while re-presenting an event from the past, it shows the memory that a social group wishes to preserve or wishes to impose on it.

Commemorative Monuments
Today there is a great amount of literature on commemorative monuments (Rowlands & Tilley, 2006) that cannot be summarized here. However, it is necessary to extract some contributions that make it possible to identify the specificities of this type of construction. One of the oldest ones distinguishes between monuments and historical monuments (Riegl, 2015). The former are the result of an explicit intention to reify the memory of an event, a person or a group (eg. monuments to the dead) while the latter are the fruit of competition of circumstances making that a construction of the past could cross time and that individuals give it a testimony value. In order to clarify the issue, Debray (1999) proposed to distinguish between 'monument-form', 'monument-trace' and 'monument-message'. The first type of monuments corresponds to the category of historic monuments suggested by Riegl. The only difference is that these constructions are characterized by their remarkable aesthetic qualities. Monuments of the second type may be documents, sites or constructions that do not result from any aesthetic intent or motivation for testimony, but which have '…a strong value of evocation, emotion or restitution' (Debray, 1999: 34). Finally, the specificity of the 'monument-message' is to fulfill a symbolic function assigned to an event of the past. Strictly speaking, this is the commemorative monument. Therefore, it is clear that, according to Debray, the commemorative monument is not characterized by its aesthetic qualities or its historicity, but by the message it delivers. It is also understood that to properly fulfill its symbolic function, the memorial must be the subject of a thoughtful design by the person who imagines it. In other words, the development of a commemorative monument is more a communication process than an aesthetic or creative process.
With the notion of 'place of memory ' Nora (1984) further clarifies the symbolic function of the commemorative monument. It is of course a question of keeping the memory of the past present in the memory, but the memory in question here is not an individual memory scattered in the consciousness of each person. Explicitly quoting Halbwachs, Nora explains that this memory is collective, specific to a group. In this sense, what is restored by places of memory -including commemorative monuments -is quite different from history, in the scientific sense of the word. More precisely, the place of memory exists only in the interaction between memory and history. It is because an event has entered into history that we feel the need to crystallize its memory in a place.
There are many works that have focused on commemorative monuments including the ones to the dead (Shanken, 2004). But it must be said that very few of them have questioned the possible connections between the architectural forms or decorations that could be seen on these monuments and the meanings that they could convey. However, we should mention the pioneering works of Prost (1977) or Agulhon (1986).
In a complementary perspective of 'historical psychology' (Meyerson, 1948), another question about commemorative monuments has been suggested. Could their study allow us to update the way in which the designers of the monuments represented the commemorative event? Indeed, Meyerson writes: 'The examination of the works, far from being the sole preserve of the historian, must constitute the main subject of the research of psychologist. In the succession of works the psychologist can find the spirit that built them, with its various aspects, levels and specific contents, and their transformations; through the history of the works, he can thus constitute a history of the spirit, of psychological functions ' (1949: 81).
For Meyerson, the works of the human mind are divided into broad categories such as language, myths, religions, but also visual arts. Commemorative monuments fall into this latter category and, following Meyerson's proposal, their study could therefore be a pathway to memory content related to a past event.

Research Objectives
Our research was concerned with an array of 329 monument projects dedicated to the memory of the victims of political repression in Russia. Each project is represented by an image (computer-generated image, drawing or photograph of a model) and it is accompanied by a text written by its designer. As it was mentioned in the introduction, these projects were proposed in response to a competition organized by the Moscow City Hall and the Gulag History Museum for the construction of a commemorative monument. First of all, it concerns a particular type of object obviously linked to the memory of the past. Our first objective was therefore to find out whether in the texts there were elements relating to the collective memory that the designers of these monuments shared about the political repression that had taken place in their country. Moreover, since the visual appearance of these monuments is intended to convey a message (Debray, 1999), we wanted to know if it was possible to find traces of this collective memory in their forms. Finally, we wanted to know if it was possible to identify the links between the texts, which are presumed to express certain aspects of the memory of their authors, and the formal characteristics of the monuments.

Study 1: Analysis of the Presentation Texts Method
The 329 texts accompanying the project images were collected to form an array of 53710 words, including 4607 different words and 1983 hapax. This array was subjected to an Alceste-type classification analysis (Reinert, 1983) using the Iramuteq 3 software.

Results
It is not surprising that the most frequent words were those that exactly reflect the objective of the competition (monument: 500 occurrences; repression: 277; victim: 217; politics: 207). The Alceste procedure identified 1495 text segments (sentences) and managed to classify 1392, or 93.11%. Figure 1 presents the result of this classification, which shows two main classes, themselves subdivided into sub-classes (for a presentation in English of the Alceste method and the classification procedure see Chartier & Meunier, 2011).
At the top of the dendrogram there were three subclasses that corresponded to the technical descriptions of the monuments. Here, the designers referred to the environment in which the monument would be installed (space, area, entrance, complex, environment), the space defined by the monument (interior, light, path) and the elements or materials with which it would be built (sculpture, inscription, fence, granite, bronze, concrete, etc.) or its dimensions and technical characteristics (height, meter, shape, frame, etc.). At the bottom of the dendrogram there were two sub-classes that referred to explanations related to the symbolism of monuments.
The examination of Table 1 suggests that the symbolism of monuments was evoked from two points of view. The first refers to the historical dimension of the event.
Here political repression seems to be evoked as a period in the country's history (tragedy, country, world, history, era), situated both in time and space. The second refers to the human dimension of the event, which seems to be evoked more through the people it has affected (man, woman, child, people).
This partition between the historical dimension and the human aspects of an event is reminiscent of the distinction made by historians between 'positivist' or 'methodical' history, essentially concerned with the description of facts (Offenstadt, 2017) and ' event' history produced or suffered by humans (Ricoeur, 1992). As far as the texts in question are not the work of historians, they should be analysed and understood in reference to the collective memory of their authors. From this perspective, it seems that this collective memory is organized according to the two dimensions, one of which would refer to the temporal and geopolitical nature of repression while the other would refer to its human consequences. The question then arises as to whether these dimensions constitute the same representation of the past or whether they reflect the existence of two different representations. It is difficult to answer by taking into consideration the available data. At most we can observe the absence of a significant correlation between the number of text segments relating to the historical dimension of the event and the number of segments relating to its human dimension in all texts in our array. This suggests that these could be independent dimensions.

Method
The first step in this analysis consisted of composition of an inventory that referred to all shapes presented in the 329 images of monuments. A shape can be defined as a combination of lines and colors that reproduce more or less accurately an object that could be perceived directly. In order to be as objective as possible in description of each monument, we chose to retain only the most unambiguous and identifiable shapes in the images. All monument projects were coded by using a series of notes that correspond to the presence (note = 1) or absence (note = 0) of each shape. Finally, the shapes that were presented on at least 10% of the images were selected for the analysis. Another criterion corresponding to the space defined by each monument was added to the inventory of shapes. Given the theme of these monuments and its obvious link with the notion of confinement (Gulag), it seemed to be useful to consider the fact that a monument has or has not an autonomous space in which a visitor could enter. In order to ensure the objectivity of this first coding, we asked another evaluator to code independently the 329 images. After that, we calcu-lated the convergence between the two codings for each of the selected shapes (Cohen's kappa test). As it can be seen in Table 2, the convergence between the two codings was strong (>0.61) or very strong (>0.81) for all the considered shapes. Based on this observation we retained the initial coding. Table 3 presents the number and percentage of monuments where the seven selected shapes were identified. Figure 2 shows six examples of monuments and their coding. As it can be seen, two items clearly stand out from the others. A significant proportion of the projects showed a human form (46.50%) or defined an autonomous space (40.00%). However, these two characteristics rarely appeared together on the different projects (Phi Pearson = -0.25, p < 0.001). They appeared simultaneously in only 41 projects, or 12.46% of the whole array. On the other hand, 112 projects showed a human form without showing an autonomous space, or 34.04%, while 92 projects showed an autonomous space without showing a human form, or 27.96% of cases. Thus, 62% of the projects alternately showed one of these two shapes. These observations therefore suggest that these are two important characteristics that distinguish monument projects quite clearly. Overall, almost two-thirds of them either showed a human form or they showed an autonomous space.  It can be seen on Table 4 that when monuments define an autonomous space, they show less often barbed wire (Chi square = 4.20, ddl = 1, p < 0.05) and more often walls (Chi square = 17.03, ddl = 1, p < 0.0001) whereas when they show a human form, they show more often barbed wire and less often walls.

Results
On the basis of these results it can therefore be assumed that the criterion of autonomous space and the human form constitutes two dominant shapes of the analyzed monument projects. Indeed, these two shapes are found in 74% of projects, rarely together and more often alternately. Moreover, as we have just seen, they are associated in different proportions with the other forms we have identified. One can then wonder about the messages that the designers of these monuments wanted to deliver. In our opinion, the monuments that defined an autonomous space aimed to suggest to the visitor the experience of confinement, while those that showed a human form aimed to recall the suffering of the Russian population during the period of repression.

Study 3: Collective Memories of Repression and Forms of Monuments
At this stage of the analysis we were interested in knowing if the explanatory registers (historical vs. human) used by the designers in their presentation texts were linked to the forms present in their monument project. Taking into consideration the previous studies, it could be expected that projects accompanied by a text evoking the human dimension of repression would show human forms more often.

Method
In order to test our hypothesis each project image was assigned two notes based on the accompanying text. The first means the number of elements in the accompanying text that belong to the Historical dimension class, highlighted by the Alceste analysis (See Study 1). The second corresponds to the number of elements of the accompanying text that belong to the Human dimension class. Two sub-groups of monuments were then considered: the subgroup of monuments accompanied by a text with only elements of the Historical dimension class (n = 89) and the sub-group of monuments accompanied by a text with only elements of the Human dimension class (n = 57).   Table 5 shows two typical extracts from these accompanying texts. Table 6 shows the frequencies of the different shapes occurrence in the monuments proposed by the two subgroups of designers. As it can be seen, depending on whether it is referred to political repression in historical terms or to its human aspects, the designers did not made identical use of the different architectural forms that we were able to identify. In particular, those authors who favored the historical aspect proposed monuments that had an autonomous space and inscriptions more often, while those who favored the human aspect proposed monuments where human forms and barbed wire were seen more often. We can obviously think that if the designers who explained the symbolism of their project by using terms referring to the human, it is precisely because they had imagined a monument showing a human form. But it is more difficult to apply the same reasoning to those who explained their monument using terms referring to history. Indeed, as it was mentioned above (see Study 2), the symbolism of autonomous space refers rather to the theme of confinement. However, if the observed differences were only tautological, the project designers who evoked confinement should have deployed a vocabulary related to this theme (prison, cell, Gulag, etc…). But this was not the case, the terms: prison, cell and Gulag were absent from the class to which their texts belonged. It therefore seems that the way in which all these designers explained the symbolism of their project, with reference to political repression, was not merely a discursive reflection of the formal properties of the monuments. It rather reflected their representation of these events and this representation of the past guided them to develop the message that they wanted to convey in their monument.

Discussion
If we accept that the analyzed texts reflect a part of the collective memory of their authors, the presented results lend credence to the thesis that there is a link between socio-cognitive constructs and visual representations (Moliner, 2016). Indeed, according to this thesis, individuals who do not share the same socio-cognitive constructs about a given object produce different visual representations of that object. As we have seen, depending on whether artists, designers or architects have a history-oriented or human-oriented memory of political repression, they did not propose the same commemorative monument projects. This observation, if it is considered to be valid, calls for several reflections on the propagation, anchoring and objectification of collective memory.
As it was mentioned above, the commemorative monument projects analyzed in this research were proposed as a part of an initiative taken by the Russian authorities, the initiative as a part of a rehabilitation process that began after Stalin's death. However, this process is still a subject of debate in the modern Russian society (Koposov, 2010(Koposov, , 2018Morenkova, 2014). It would be difficult to present the all the elements of this debate in detail here, but we can try to sketch out the main points. On one hand, the 'progressive' trend (Morenkova, 2014) initiated by Gorbachev's policy would like a complete rehabilitation of the victims of Stalin's repression, accompanied by the denunciation of the crimes committed, probably associated with a forgiveness enterprise involving the responsibility of the Russian State. On the other hand, during the 1990s, there emerged a 'patriotic' trend (Morenkova, 2014) around the narrative of the 'Great patriotic war' (Second World War) which '…serves to eclipse another memory, that of Stalin's terror, and to convince Russians of the positive role of the State in national history' (Koposov, 2010: 51). It seems that for a while this story has successfully fulfilled its function. But the action of some media and opposition parties, the debate among historians and the public discourse of former victims of repression or their descendants will give renewed vigor to the denunciation of the crimes of Stalinism (Morenkova, 2014). This Table 5: Excerpts from typical texts of the Historical dimension and Human dimension classes.

Excerpt of a typical text from the Historical dimension class Excerpt of a typical text from the Human dimension class
'Monument to the memory of the victims of the political repression. The mad history of the twentieth century, of absolutely ideologized century, is freely readable in the context of mythological and at the same time dogmatic consciousness. Without reconciliation, it is impossible to rethink our common past and to understand the present, and to build the future'.
'The rock devouring human fates, with chained four prisoners, symbolizes the torments of the repression. The Angel Savior in the form of a female figure symbolizes the experiences of loved ones, families, it is a symbol of hope, a source of strength that helps to endure the pain of imprisonment'. may explain the initiatives taken by Presidents Medvedev and Putin to rehabilitate the memory of the victims of repression. However, these initiatives, and particularly the one relating to the construction of a commemorative monument, can be seen as an attempt to propose that Stalin's repression is represented in an acceptable way by the greatest number of people. In other words, it seems that this project was part of a 'propagation' strategy in the sense defined by Moscovici (1961). It should be recalled here that according to Moscovici, propagation refers to a communication strategy aimed at proposing an image of a social object that is compatible with the values of the target group. However, by organizing the ' commemoration of the victims of political repression', the Russian authorities are focusing attention on the victims, to the detriment of their executioners. In this way, it attempts to satisfy, at least partially, divergent aspirations in Russian society by finding a way that is compatible with 'progressive' and 'patriotic' positions. There are indications that artists who have proposed monument projects have been influenced by this propagation strategy. Thus, for example, in their proposals there are only six (evening 1.82%) that show a person who can be considered a soldier, a guard or an executioner in the pay of Stalin's power. The assimilation of the concept of collective memory with the concept of social representation of the past presupposes that a collective memory emerges and perpetuates itself notably through the anchoring process. In fact, the first anchor of a collective memory lies in the social identity of the group that carries it, since one of the functions of that memory is precisely to preserve the image that a group makes of itself. But beyond this ontological anchoring, it seems that a social representation of the past can also be anchored in other frames of reference, likely to be combined with the initial socio-identity anchoring and to modulate its effects. With regard to the case we have studied, that of political repression in Russia, we have seen that these frameworks refer to the history of the country or the experience of its people. We could then think of two ways to consider an event from the past depending on whether or not we choose to integrate individual trajectories of ordinary people into it. Thus, the preference to represent repression in a rather historical or human dimension would refer to different conceptions of the past, which, for some, would be a succession of interacting events, and for others a tangle of individual stories. An illustration of this dichotomy can be found in recent research on the memory of the First World War (Bouchat et al., 2018). In this work, the authors asked 1347 students to associate five words with the event. The collected associations were the subject of an analysis similar to that carried out in our Study 1 (see above). This analysis revealed seven classes of discourse, some of which included terms referring to the historical and political dimension of the event (Treaty, Versailles, Entente, etc.), while others contained terms referring to the first actors in the conflict, i.e. the combatants themselves (soldiers, Poilus). However, among the participants, psychology students produced, among other things, more terms from the class citing combatants, while history students produced more terms from the class citing the political consequences of the conflict. As a part of our research we have no information on the designers of the analyzed projects. It is therefore difficult to explain the differences in anchoring that we observed. But there is perhaps one road to explore that suggests that social identity may not be the exclusive anchor point for social representations of the past.
Finally, the assimilation of the concept of collective memory to the concept of social representation of the past also implies that a collective memory can be objectified. Clearly, the commemorative monument is one of the most remarkable modes of objectification of this type of representation. But it is a mode that, unlike the one we observe for social representations, goes beyond the sphere of individual consciences and the mental images they contain. By this we mean that the 'monumental' objectification of a social representation of the past is necessarily introduced in the public space. As a result, its product enters the field of the visible. But for this materialized objectification to achieve its objective, it is probably necessary that it echoes or at least does not clash with the various representations of the past that coexist in a given society. For example, when in 1990 the Russian authorities accepted the construction of the Solovetsky Stone, they did not take much risk. As mentioned above, this monument is summed up as a stone whose symbolic origin (the Solovki Islands where the first Gulag was implanted) can only be clearly understood by initiates, that is historians or descendants of the repressed involved in the militant activity in favor of the rehabilitation of victims. Anyone who passes this monument today without knowing the details of this history cannot understand that the monument in question is dedicated to the memory of the victims. To some extent, it seems that the monument that was finally chosen in the competition at the end of the 2015 and inaugurated in 2017 in the presence of President Putin follows the same logic. This monument brings together many shapes that we have identified in our research, including the two dominant shapes of the array. It is indeed a curved wall of about thirty meters, thus drawing an elliptical space.
This wall is made up of intertwined human silhouettes and the inscription 'remember' is engraved in 22 languages. This monument was officially dedicated to the victims of political repression, but it is difficult not to see in it a universal symbolism that refers to the victims of all repressions. It can therefore be assumed that, by choosing this monument, the Russian authorities have completed their work of propagation by introducing into the public space the materialized representation of a socially neutral collective memory that is well aware of any explicit denunciation of the Stalin period.

Conclusion
It is clear that the research we have just outlined has several limitations. Probably the most important of them is the lack of information about the artists, designers and architects who proposed their monument projects. For example, it would be very useful to know whether they had any victims from the time of the repression in their families (parents, grandparents, etc.). This information might have helped for a better explanation of the observed differences between monuments with human shapes and those with other shapes. In the same way, it would be interesting to know how these designers positioned themselves in relation to the 'progressive' and 'patriotic' currents that are now passing through Russian society. Here again, it can be assumed that these positions could be linked to the different shapes identified in the proposed monuments. Despite these limitations, it seems to us that this research suggests that the study of commemorative monuments could provide a privileged access to the collective memories of certain past events. We hope that we were able to convince the reader of this.

Notes
1 It is a simple stone from the first Gulag at Solovky. 2 Levada Center, Russian Public Opinion 2017: h t t p s : / / w w w. l e v a d a . r u / c p / w p -c o n t e n t / uploads/2018/05/2017Eng.pdf. 3 Ratineau, 2008, 2014: http://www.iramuteq.org.