A politics of placelessness? The limits of democratising memory in the Centro de Documentación e Investigación of Lima’s Lugar de la Memoria

The Centro de Documentación e Investigación (CDI) is an online archive which provides free access to over 20 collections on Peru’s internal armed conflict (1980–2000), a conflict which was distinctly shaped by racial and social inequalities. The digital nature of the archive is presented as an opportunity for democratising access to these historical sources and for promoting commemoration as a means of cultural reconciliation. However, there is a risk that pre-existing social geographies and material concerns will mean that the CDI replicates offline exclusions. This article argues that, whilst the CDI has made these documents accessible to a broader geographical audience, usage of the digital archive is still largely mediated through social hierarchies. Through its online archive and offline engagement activities, the CDI appears to have generated a more geographically distributed network of content producers, but one which remains biased towards university-educated participants in urban areas.

a documentary heritage that belongs to all Peruvians' (Lugar de la Memoria, la Tolerancia y la Inclusión Social, 2017a). 1 The project aims to centralise over 20 collections on the conflict, which have previously been housed in disparate offline locations, to make them freely available to a wider public. By achieving this, the team at LUM hopes to 'generate spaces of reflection' which would help to 'contest and broaden the horizons' of the official historiography (Lugar de la Memoria, la Tolerancia y la Inclusión Social, 2017a) of a conflict which was distinctly shaped by racial and social divisions.
As highlighted in the editorial to this issue, scholars of digital technology have long debated the spatiality of digital space, with the Internet often presented either as a boundless, liberatory space, free of offline exclusions (Barlow, 1996), or as a lawless realm in which social norms break down and legal protections from predatory behaviour are absent (Jakubowicz, 2017;Valentine and Holloway, 2001). Others have argued that there are stronger continuities between online and offline worlds than has previously been recognised (Cowan, 2005) and that this can either help to foster a stronger sense of belonging and community (Coles and West, 2016;De Waal, 2014), or to replicate and extend existing exclusions (Chaudhry, 2015;Tawil-Souri, 2012;Zook, 2012).
Bearing this in mind, could digital technology mitigate the centralisation of wealth, power and population in Peru's capital which has historically limited the ability of Quechua-speaking, indigenous populations to contribute to knowledge production (Kristal, 1987)? By collapsing the physical distance between Lima and the country's interior, the CDI is presented as a possible means of overcoming the vast social and economic inequalities in Peruvian society, and of offering communities who were disproportionately victims of violence the opportunity to challenge historiographical orthodoxies on the conflict. However, as Haskins (2007: 405) argues, while the Internet may appear to 'promise' greater interactivity between communities 'formerly limited in time and space', the 'very features of electronic communication. . .can also abet political fragmentation' (Haskins, 2007: 419) and therefore may only replicate existing place-based exclusions. This article will therefore use the CDI as a case study to ask if such projects are grounded in a 'myth of placelessness' (Cowan, 2005) and examine whether the digital archive can overcome the offline, material barriers to participation in knowledge production that remain present throughout Peruvian society and geography.
Below, the article summarises the existing historiography on the conflict, highlighting the importance of place in the construction of post-conflict narratives. Using documents published by the LUM, I outline how the LUM and CDI are designed to create spaces in which dialogue between dominant and marginalised communities can occur, emphasising why it is crucial that the voices of Quechua-speaking, indigenous communities are not excluded from this historiography. The article then approaches the question of whether the CDI has been successful in its mission to broaden participation in memory production (Lugar de la Memoria, la Tolerancia y la Inclusión Social, 2017a) with an analysis of geographical user data from the CDI's online platform. 2 This data gives us a broad sense of where the archive has been accessed from over time, but it does have several limitations. Firstly, the location is limited to administrative department level and tells us little about who is accessing the archive; conclusions about whether participation is diversified across different social groups are therefore limited. Furthermore, the data used is for the years 2017-2019, meaning that it provides only an early snapshot of how successful the project may be up to 3 years after its opening.
Nonetheless, the sharp skewing of this data towards the Lima region and other university cities does suggest that more work may be needed for the CDI to achieve its stated goals. Accordingly, the article ends by highlighting a series of offline engagement activities, including CDI user workshops and an essay competition, that may help to partially close the 'digital divide' through skills training and by forming dispersed researchers into a loose but connected network. While these activities represent an attempt to marry online and offline engagement to overcome Peru's exclusionary social and physical geography, their impact also remains mostly limited to universityeducated participants in urban areas. At this stage, the CDI appears to have generated a slightly more geographically dispersed network of content producers on the conflict, but is yet to achieve its aim of democratising knowledge production across a wider social spectrum.

The CDI, LUM and politics of memory in post-conflict Peru
To understand the importance of the CDI in relation to present-day struggles for memory in Peru, it is necessary to briefly summarise the history of the LUM and the evolution of Peru's post-conflict historiography. In 2003, the CVR revealed that there were approximately 70,000 deaths during the conflict, more than double the previous estimate, and that this number had been vastly underestimated because 'the tragedy suffered by the populations of rural Peru, the Andean and jungle regions. . .was neither felt nor taken on as its own by the rest of the country' (Comisión de la Verdad y Reconciliación, 2003: 316). The violence perpetrated by Sendero Luminoso and the Peruvian armed forces overwhelmingly victimised indigenous communities in the Peruvian interior, who have historically faced racial discrimination and socioeconomic exclusion from the country's whiter, colonial centre in Lima. Not only did these communities face a disproportionate level of violence during the conflict; these tens of thousands of deaths were obscured and barely noticed for decades.
In post-CVR Peru, two broad memory 'camps' (Drinot, 2009) have emerged: a state-military narrative which emphasises the role of 'terrorist' violence in promulgating the conflict, and a human rights narrative which is critical of both insurgent and state violence, while seeking to understand how structural racism shaped the conflict. For the latter, it has been important to uncover unheard testimonies from Andean and Amazonian regions, whilst pro-military narratives have often disproportionately focused on the car bombings and assassinations perpetrated by insurgent groups in Lima (Willis, 2018). There is therefore a 'locatedness' (Radstone, 2011) to knowledge production in post-conflict Peru that has shaped who is 'eminently grievable' (Butler, 2009: 24) and which perspectives remain ignored. In order to challenge the dominance of state-military and limeño narratives (i.e. those produced by Lima inhabitants, particularly whiter and more middle class groups), recent years have seen a number of studies examine how post-conflict narratives are inflected by the geographical perspectives (Willis, 2018) of their authors, as well as by their class (Greene, 2016), gender (Guiné, 2018) and ethnicity (Koc-Menard, 2014;Meza Salcedo, 2014). Scholars have also conducted hyperlocal studies on Andean and Amazonian communities (Del Pino, 2017;Meza Salcedo, 2014;Villapolo Herrea, 2003) to gather sources and perspectives that were previously unheard or ignored in the capital. However, while increasingly influenced by indigenous perspectives, scholarly writing on the conflict remains dominated by academics based in Lima or elite international institutions. The LUM and CDI therefore have a dual purpose: to act both as a 'pedagogical institution' that encourages a limeño audience to 'identify with affected populations' (Lugar de la Memoria, la Tolerancia y la Inclusión Social, 2014: 67-68), and to provide a more diverse set of actors with the resources to participate in knowledge production.
The LUM was officially inaugurated in December 2015, culminating a drawn-out process during which its opening had been continually postponed. Taking a lead from similar museums dedicated to promoting historical education, collective memory and human rights (e.g. Yad Vashem [Jerusalem], the Museo de la Memoria y los Derechos Humanos [Santiago]), the construction of a national memory museum in Peru was first recommended by the CVR which called for 'museums, monuments, publications' to commemorate the conflict (Comisión de la Verdad y Reconciliación, 2003: 99). Although originally conceived of as a civil society project, the LUM came to be managed by the Ministry of Culture in 2009 following a donation (which was initially rejected, but later accepted, by then President Alan García) from the German government to construct the museum (Milton, 2018: 153). Ongoing wrangling and criticisms from civil society and the Peruvian armed forces over the museum's content, and the removal of several LUM Directors, further delayed the LUM's completion (Milton, 2018: 184;Sastre Díaz, 2015: 128-138).
These contestations over the purpose, form and content of the LUM are crucial for understanding the logic behind the construction of the CDI. As Milton argues, the Peruvian armed forces, civil society and victims' families have all used public space as a battleground in which competing narratives of the internal armed conflict are curated, contested and negotiated (Milton, 2018). As it is situated in the 'gastronomic cordon' and 'tourist circuit' of middle-class Miraflores (Montalbetti, 2013) and located within a wider commemorative landscape which is deeply encoded with military and oligarchic symbolism at odds with the social inequalities that precipitated the conflict (Milton, 2018: 189;Willis, 2018: 303-309), the LUM's ability to act as a site for social reconciliation for all of these groups is limited.
The museum has also been criticised for how it treats indigenous communities in its museographical narrative, as Sastre Díaz (2015: 128) argues that it sits within a 'tradition of cultural discrimination' by state institutions. Indigenous communities, such as the Asháninka people of Peru's central jungle regions, appear predominantly as victims of violence, reinforcing stereotypes of passivity and victimhood. The LUM's architecture leads visitors in a linear manner from the bottom to the top of the building in a way that, Sastre Díaz (2015: 142) argues, reinforces homogenising narratives of the conflict in which there is one Peruvian nation united in grief (rather than numerous disparate regions divided by generations of structural violence and discrimination). Although efforts have been made by the curatorial team to disrupt this linear narrative (Willis, 2018: 306-307) there is still insufficient attention paid in the museum to this history of state violence against Peru's indigenous populations as is done in other museums such as Yalpani Wasi in Huancayo. Furthermore, González (2017) has argued that the LUM's use of digital technology risks alienating working class and indigenous Peruvians with less experience of digital engagement. For the CDI to achieve its aim of acting as a form of 'symbolic reparations' (Lugar de la Memoria, la Tolerancia y la Inclusión Social, 2017a) to marginalised communities, the archive therefore needs to create a space in which more heterogeneous, indigenous, non-Lima narratives of the conflict and structural violence can be developed, with users able to move easily between different sources and narratives outside of the limits of the LUM's architecture.
However, despite these criticisms and allegations of state censorship owing to the input of the Peruvian armed forces into the museum's content (Milton, 2018: 153-160;Sastre Díaz, 2015: 142), the LUM remains strongly associated with the human rights memory camp. In part, this is due to the involvement of numerous scholars and artists who have been critical of state violence (e.g. Alexandra Hibbett, Ponciano del Pino, José Carlos Agüero, Karen Bernedo, Miguel Rubio) in the LUM's curatorial team (Lugar de la Memoria, la Tolerancia y la Inclusión Social, 2014: 19). This team has stated that it was partly inspired by a 'never again' approach to cultural memory (as seen in post-dictatorship memory struggles in the Southern Cone) but that the LUM should also go beyond this by unpicking the structural racism and inequalities that predated the conflict, generating a 'more democratic, tolerant and inclusive country' (Lugar de la Memoria, la Tolerancia y la Inclusión Social, 2014: 24). Ultimately, this means that there are tensions, contradictions and ongoing contestations in elements of the LUM's design. The CDI, in this context, is all the more important as it provides a more fluid, adaptable space to share sources that may well have been censored or removed from the museum's permanent exhibit (although of course, there is the potential that sources could just as easily be censored or removed from the digital archive).
More importantly, whereas activists or victims of violence would previously have needed to travel to Lima to engage with the LUM's content, the CDI offers the opportunity to do this from one's own home or locale. By late 2017, the CDI comprised some 200,000 digitised files on its virtual platform. The collections include 'Public Testimonies to the CVR (2002CVR ( -2003'; a collection of music from Ayacucho (where 40% of the conflict's victims lived); and the 'Newspaper Archive 1981-2007' which contains 190,000 press articles (from newspapers and magazines including Expreso, La República, El Comercio and more) relating to the conflict over 25 years. 3 These sources are absolutely crucial for constructing new narratives and fresh perspectives on the conflict. For example, whilst the CVR's final report was heavily influenced by the public testimonies given to the commission, the CDI's collection allows researchers to return to the source material. Given that many of the testimonies were given in Quechua before being translated into Spanish, and that only one of the Commissioners spoke Quechua, it has been argued that the CVR 'reflected elite sectors of Peru least affected by the violence' (Heilman, 2018). By making this source material accessible to a far greater audience outside Lima (including a far greater number of Quechua speakers living in rural areas) the CDI offers an opportunity for Quechua-speaking researchers to produce alternative perspectives on the conflict which mitigate against some of the CVR's own institutional biases. It could be argued that it is only by the creation of the CDI that one of the CVR's key recommendations, that 'access to the final report, and the documentary evidence collected or produced by the CVR, be as broad as possible' (Comisión de la Verdad y Reconciliación, 2003: 106), has finally been fulfilled.

Logics of placelessness and connectivity
At this point, it is important to clarify the context in which I am deploying terms such as 'placelessness', 'connectivity' and 'democratisation' in relation to the literature on digital technology. Halstead (editorial, this issue) argues that early scholarly studies on digital technology 'tended. . . to depict the emerging Internet as a boundless, amorphous, and placeless virtual realm divorced from the offline world'. In this context, 'cyberspace' is presented as spatially unconnected to the material world, as a space that is free of the many material and imagined borders of real world geography. As Cowan (2005: 258) highlights, the experience of engaging with the Internet is built upon a series of metaphors that suggest digital technology takes us to a place that is 'somehow qualitatively different from the more mundane places from which we access the online world'. This generates what Kneale (1999) terms a 'fictional geography' in which digital space is divorced from the materiality of the offline settings through which it is accessed.
This 'placeless' reading of digital space can be interpreted, in a positive sense, as a means of collapsing the distance between people and allowing new communities to forge (Coles and West, 2016), or as a space in which new social relations can be constructed outside the potentially exclusionary boundaries 'of place, family and community' (Hoskins, 2011: 22). For communities that experience place through a series of marginalisations and exclusions, placelessness might offer liberation. Furthermore, by opening up content creation and knowledge production to a more socially, geographically and ethnically diverse set of actors, digital space can contribute to the democratisation of memory-making by disrupting the 'traditional hierarchy of author-text-audience . . .Instead of only official accounts . . . all kinds of stories can now become part of an evolving patchwork of public memory' (Haskins, 2007: 405-406). As Drinot (2011) argues, the 'de-territorialised nature of cyberspace' gives it its 'democratising potential'. In this sense, there is what I term a politics of placelessness at the heart of the CDI's design, using digital technology to dissolve social barriers to knowledge production and generate fresh analytical perspectives on the conflict.
Alternatively, the placelessness of cyberspace could be interpreted as a disconnection from one's identity, as replicating a sense of commercial exchangeability and anonymity (Relph, 2008: 143) or as the antithesis to notions of 'belonging' (Brubaker, 2005). The Internet can be used as a divisive tool to break apart communities by reflecting 'geographies of racist practice back onto the places from which they emerged' (Zook, 2012); in other words, by reinforcing and exacerbating pre-existing societal divisions. These contrasting readings suggest that notions of 'connectivity' are key; particularly for understanding how communities can remain unfractured and connected to other groups without excluding outsiders. Hoskins (2011: 20) suggests that society has undergone a 'connective turn' in which 'mobile media and social networking technologies' have enabled 'more complex and diffused connections within and between groups'. For Coles and West (2016: 45) the Internet invites users to 'imbue their pages with a sense of their personality and to establish personal connections with other online individuals'. In this sense, rather than dissolving the space between users and communities, digital media appears to connect users as nodes in a distributed network of content producers.
In the case of the CDI, the question is the extent to which such a network would replicate, or could overcome, real world geographies and inequalities. Digital exclusion is not a problem limited to Latin America (Van Dijk, 2005), but the continent's diverse topography and 'major deficits in basic infrastructure' (Galperin, 2017: 13) contribute to a particularly stark 'digital divide' (Münster, 2005) between urban and rural areas, with low levels of overall internet usage in the Northern Andes (Pérez Álvarez, 2014). Despite significant progress in Peru between 1994 and 2001, when it was estimated that over 1000 cabina públicas (public access points similar to internet cafés) were opened and the 'Peruvian model of Internet access . . . gained international recognition', only 2.8% of the total population had used the internet by the end of this period (Holmes, 2001: 1). Furthermore, digital exclusion is not just a matter of access; the knowledge and skills to navigate technology are also unequally distributed. In part, this explains the gendered inequalities in internet usage which have been identified in Latin America (Bull, 2016;Gray et al., 2016). By 2001, internet users still tended to be 'predominantly urban and located in certain regions, better educated and wealthier, young, and male' (Holmes, 2001: 26). By 2017, 48.73% of the Peruvian population were accessing the internet (World Bank, 2017). However, research suggests that significant inequalities persist, with only 14% of people in rural areas using the internet, and young people with medium to high education levels doing so the most (BBVA Research, 2017: 2).
The question, therefore, is how the CDI reckons with its own 'locatedness' (Radstone, 2011: 111) and generates a space for positive connectivity without replicating the offline exclusions and racist practices of Peruvian society. Uploading a large collection of sources to an online archive may not be, in itself, enough to guarantee a more democratic and inclusive politics of memory. However, the series of engagement activities developed by the CDI suggest a genuine attempt at 'distributing authorial agency among various institutions and individuals involved in the production of content' (Haskins, 2007: 406).

Placeless archive, or a distributed network of knowledge producers?
How, then, might we judge the CDI's success in democratising knowledge production on the conflict?

Form and design
The CDI brings together a wide range of sources previously housed in disparate locations such as the Pontificia Universidad Católica de Perú (PUCP) and Instituto de Estudios Peruanos (IEP). Both the PUCP and IEP are elite academic institutions with (quite literal) gatekeepers that limit access without official documents and payment (for non-members). Access to both institutions is mediated through a range of material and social contingencies that reduce the likelihood of their members and students being of a working class, indigenous or outer-urban/rural background. By contrast, the 'dissemination of these collections and the free download of hundreds of files (digital audio, video, press, etc.)' via the CDI is designed to generate 'new knowledge and . . . to foment the study of the [conflict]' (Abanto, 2017: 282). The CDI is presented as a 'virtual platform' which provides a 'space for research . . . free to access for the whole public', where 'neither user registration nor recommendation letters are required', 'there are no hours and the files are easy to download' (Lugar de la Memoria, la Tolerancia y la Inclusión Social, 2017a). The sources are available for 'present and future generations', with the platform facilitating, 'interaction, searching, and support' in order to 'disseminate information to a growing public and promote knowledge production' (Lugar de la Memoria, la Tolerancia y la Inclusión Social, 2017a). With this access, the CDI sees itself as providing the resources for victims and socially marginalised communities (including indigenous, working class and LGBTQ+ Peruvians, among others) to participate in a democratised and more collaborative process of memory-making.
However, given what we know about the digital divide in Peru, it is likely that the CDI's predominant audience will be young Peruvians living in urban areas, especially university students and those with the ability to access the CDI at home. Even where digital participation has been promoted and is increasing in some rural areas, the nature of the equipment being used (usually low-cost mobile phones [eMarketer, 2017]) is not conducive to the large amounts of data bandwidth needed to access some of the CDI's files. 4 This would run counter to the democratisation of memory production and would instead, in the words of Haskins (2007: 402), only replicate 'official ideologies of the ruling elites while claiming to speak on behalf of the people'. Although only open for a short period, it is possible to draw some early conclusions on whether this is the case by analysing CDI user data provided by the LUM, which gives some insight as to how placeless we might really consider the archive to be (see supplemental material). 5

CDI user data
This data, collected between July 2017 and March 2019 and broken down by geographical region, shows that usage of the digital archive has been heavily weighted towards the Lima region, major urban areas and university towns. For July-September 2017, the CDI was used in 46,250 sessions, an average of 503 per day, of which 93% were from within Peru (the US was the only other country to account for more than 1%). Within Peru, 48.85% of these sessions occurred in the Lima region, with 25.24% from Apurímac, and just over 5% in Cusco and Junín. Between January 2018 and March 2019, Lima's overall share of total sessions remained fairly consistent at about 40%, whilst Apurímac and Junín's share of the total represented only 5.79% (the highest region outside of Lima) and 2.95% respectively. Cusco was the second highest region outside of Lima, consistently attracting 5% of CDI users across each period. Arequipa had the third highest average outside of Lima with 4.26% average after a high of 9.61% in January-March 2019. Even when the region's population is taken into account (34.8% of Peru's total) it is overrepresented in the number of total CDI sessions (41.39%). This would suggest that, whilst the hypothetical placelessness of the archive offers the possibility of removing barriers to participation, usage of the CDI (as with digital participation as a whole across Peru) remains fundamentally shaped by offline social and geographical exclusions. It is little surprise that Cusco and Arequipa rank highly among other departments, given that both regions are centred around large cities with established universities, relatively good digital infrastructure and populations of roughly 1.5 m people. Similarly, Junín, La Libertad and Piura (with populations of 1.5, 2.2 and 2.75 m respectively) each account for roughly 3%-3.5% of the total CDI sessions. It is reasonable to hypothesise that the majority of these sessions will have been from their large capital cities with multiple universities (Huancayo, Trujillo and Piura). In this sense, as we might expect, access to the digital CDI appears to be deeply influenced by Peru's offline geography.
There are, however, some positives to be taken from the data, with approximately 65,000 CDI sessions having been initiated outside of Lima in regions which previously would have had limited access to key sources on the conflict. Taking even the reasonably well connected regions highlighted above, the documents held in the CDI have been accessed 37,542 times (in Arequipa, Cusco, La Libertad, Junín and Piura) by people who may otherwise have had to travel to Lima to have access to the same range of documents. Even across the more isolated jungle regions in the Peruvian Amazon and along the borders with Brazil and Ecuador, there were a combined 843 sessions initiated in Madre de Díos, Ucayali, Loreto, Cerro de Pasco and Amazonas. Of course, this tells us nothing about who is accessing the CDI from these places or for what reason. Yet, even whilst some of the sources have previously been available online elsewhere, that the CDI is being accessed in some numbers from across Peru suggests at least some degree of success. Whereas previously narratives of the internal conflict have been centred on the perspectives of limeños, the collections made available through the CDI offer the opportunity for Peru's disparate communities to reflect on their own local microhistories of the conflict.

Offline/online engagement activities
It is also important not to discount the 78,237 sessions initiated from Lima during this period as the LUM hopes not just to overcome geographical barriers to knowledge production, but social barriers too. In order to encourage greater participation with the CDI, the LUM has run a series of workshops in partnership with universities, called 'Memoria, Fuentes Históricas y Espacio Digital' ('Memory, Historical Sources and Digital Space'). These workshops were run throughout 2017 at the Universidad Nacional Mayor de San Marcos (UNMSM), Universidad Autónoma del Perú -Sede Central, Universidad Nacional Federico Villarreal, PUCP, Universidad de Ciencias Aplicadas (all Lima), Universidad Nacional San Cristóbal de Huamanga (UNSCH -Ayacucho) and Universidad Jorge Basadre de Tacna (Tacna). These sessions are run by María Elena Príncipe and Julio Abanto (both work on the CDI Project and are historians at UNMSM) and are designed to introduce the CDI and its collections to students and to train them in how to use the digital platform. While the workshops appear to focus predominantly on introducing the technical aspects of the CDI, ensuring 'correct use' of the data and training students in how to use the search function, they also act as an introduction to the study of violence and memory. 6 The LUM has also co-ordinated a series of specialist workshops, entitled 'Pedagogía y Memoria' ('Pedagogy and Memory'), which use the CDI collections to train secondary school teachers in how to approach personal and collective memories, or address issues, arising from the conflict in their classrooms. Further activities hosted at the LUM include film screenings, school visits (approximately 7000 schoolchildren visited the LUM in 2017), tours with members of the curatorial team, and public debates. 7 The LUM has also established a national essay competition that has resulted in the two anthologies on the theme of youth and the internal conflict. 8 Although not specifically tied to the CDI, the competition is described as being 'a part of LUM's aim to promote research and access to information [on the conflict]' (Lugar de la Memoria, la Tolerancia y la Inclusión Social, 2017b: 13) and aim to meet the need to 'have larger spaces of dissemination for new studies' (Abanto, 2017: 279). Of the twenty authors, six appear to be current university students, eight are graduates, three are or have recently been engaged in postgraduate study and three appear to be in paid scholarly work. Nine have either studied or worked at UNMSM, six at PUCP, one at La Cantuta university in Lima, three at universities in Cusco and Ayacucho, and two are currently based abroad. Whilst the entries do contribute some fresh, innovative perspectives on the conflict, analysing memories of the conflict through the lens of Peruvian rock music, private school scholarships and the perspectives of young soldiers in the post-conflict period, it is difficult to say that they reflect a more distributed form of knowledge production, coming overwhelmingly from graduates or employees of Lima universities. A fuller analysis of these publications is beyond the scope of this article; however, the contributions are perhaps not as diverse as one would hope and as such do not give the impression that the CDI has, yet, supported the emergence of a new generation of citizen researchers. In this sense, the LUM appears to have succeeded to some extent in developing a more closely networked community of researchers producing work on the internal conflict, but one drawn predominantly from sectors of society that have traditionally been engaged in knowledge production.
Without further data, it is difficult to draw too many conclusions about the effectiveness of the CDI workshops in broadening participation (for example, neither Tacna nor Ayacucho region registered more than 1% of total visits between July 2017 and March 2019), particularly without more detailed data on users or where the CDI is being accessed from. Nonetheless, these workshops demonstrate an awareness within LUM that the digital nature of the CDI archive alone is not enough to overcome the barriers to participation in memory-making. Rather than assuming that they are overcome purely by transferring the sphere of participation into digital space, these barriers can only be ameliorated through a recognition of their materiality and rootedness in Peruvian geography. Whilst the CDI user data does not suggest that the LUM's workshops and other engagement activities have yet achieved the establishment of a broad-based network of content producers on the conflict (outside of universities at least), they are at least based in the recognition of these social distances.

Conclusion
The case of the CDI suggests that its digital nature is not enough to remove the social, geographical and technical barriers to participation in historical memory-making which exist in Peru. The CDI platform does appear to mitigate many of the deeply embedded social exclusions which permeate Lima's offline archival geography, and in this sense supports the idea that digital archives can provide a degree of freedom from the 'social bonds of place, family and community' (Hoskins, 2011: 22). However, the user data collected by the CDI in the past 3 years suggests that access remains highly centralised around Lima, large urban areas and cities with a large student population.
In this sense, the CDI's politics of placelessness is both insufficient and incomplete as the archive does not collapse the space in between individuals or communities. However, the museum's team has recognised the limitations of its own project and has provided some participants with the sources and necessary digital skills (a key feature in Latin America's 'digital divide') to make connections with others as part of a distributed network of content producers. There remain significant limits on the geographical reach of this network; yet it is arguably already more geographically and socially diverse than those who would typically make use of Lima's offline archives. Furthermore, the CDI's potential to provide better access to historical sources to rural communities in the Peruvian Andes and Amazon is yet to be reflected in the LUM's publications.
The CDI is therefore an important demonstration that 'placeless memories' are rarely free of exclusion in practice and remain deeply shaped by forms of social, technological and cultural geography. Rather than seeing digital archives as inherently democratic in nature, similar projects would do well to learn from the CDI's example, taking account of these exclusions and acting to mitigate them. Only through such work can the democratisation of knowledge production be truly achieved.