Paradoxes of curating colonial memory

In the last decade, many initiatives were taken to digitize colonial archival legacies. In this article, we analyse Dutch policy and a number of Dutch initiatives in this field with the aim to find answers to our central question whether digitization of colonial archival legacies offers possibilities to decolonize these archives. The aspiration to decolonize colonial legacies seems to be a paradoxical statement since there is something innately colonial in the recordkeeping systems that cannot, and should not, be removed. But digitization of archives means creating new recordkeeping infrastructures, and these new infrastructures shape new interfaces between the documents which were created in the past and the users of today. We argue that decolonizing these archives can be based on a comprehensive understanding of the complexity of the variables which shape the new digital archival infrastructure. Inspired by the third-space perspective and the concept of (de)coloniality, we explore the possibilities to develop archival infrastructures that contribute to decolonizing colonial archival legacies in the sense of offering multivocality, multiple agency and multiple provenance. We conclude that what we call third-space infrastructural frameworks create promising opportunities to contribute to the decolonization of colonial archival legacies.


Introduction
More than 30 years ago, student Frank Westerman, now a well-known Dutch journalist, visited a small museum in the Spanish town of Banyoles. Among the taxidermized animals which were exhibited in the museum, he discovered a display case 1 3 containing a body of a small black man. The text explained that it was a 'Bushman from the Kalahari'. The postcards the museum sold with the picture of this macabre curiosity mentioned that he was 'Bechuana' (member of the Tswana peoples in Botswana). With 'a diffuse sense of shame', Westerman sought out who this man was and how he ended up as an object in the museum: '[t]he lack of clarity about his background marked his dismantling as an individual (…). He no longer had a name, and nobody knew his date of birth or death. All his characteristics had been lost' (Westerman 2019, pp. 20-21). Westerman's search resulted in a book, published in 2004, in which he describes the history of the exhibited stolen body after it was shipped from Africa to Europe. In 2000, 170 years after the unknown man was dug up by Jules and Éduard Verreaux, he was reburied in Gaborone, Botswana, the country in which he presumably was born and died. In 2019, Westerman published a revised edition of his book in which he explained that newly discovered sources have shown that the man was reburied in the wrong country. The Verreaux brothers travelled no more than 100 km from Cape Town, South Africa, which is almost 1000 km from Gaborone. Westerman explains that due to 'improved digital bore technology' it is possible 'to penetrate the thickest layers of archives' and discover information he did not know in 2004. Thousands of collections have been digitized in the past few years, and he received more and more alerts which were related to the digital lure of a combination of words of 'Verreaux' and 'Betjouana' in many different spellings (Westerman 2019, p. 247). It is a horrific and painful story of colonialism, and it is definitely not an isolated case. It also shows how archives and digital technology may help to find and reveal the facts behind the colonial behaviour and activities. Digitized archives as instruments, as sources, to facilitate and contribute to justice, ethical atonement and maybe in some respects even to facilitate decolonization. But it also raises the question of what the position of the records, archives and archival institutions is in the debate of decolonization and whether and how to value new digital infrastructures in this respect?
In 2017, the Dutch National Archives and Leiden University organized a conference entitled 'Rethinking the Dutch East India Company'. The conference aimed to discuss new research perspectives and opportunities which emerged from digitizing large parts of the 1.2 km of shelves filled with records of the former trading company which are kept by the Dutch National Archives. The more fundamental question which was addressed by the organizers of the conference was whether and how the Dutch East India Company (VOC) records can be decolonized. Or to put it differently, whether digitization and online availability of former colonial records may contribute to decolonizing these archives? And even more importantly: what does it mean to decolonize archives? In more general terms, these questions were also addressed in Bastian (2019). For the Dutch debate, see Dresscher (2018); Karabinos (2019). These questions are part of a much broader debate in which disciplines and memory institutions critically examine their role in maintaining and transforming colonial perspectives and whether and how they can become more inclusive. Until now, the debate in the Netherlands on decolonizing memory institutions and their collections was mainly conducted by museums and barely by archival institutions. See for instance the Studio-i initiative of Stedelijk Museum, Amsterdam and Van Abbe Museum, Eindhoven, to promote diversity and inclusiveness. Concrete

3
Archival Science (2020) 20:199-220 products are glossaries in which the relevance of language is explained and alternatives for problematic terms are suggested, for example Venir and Lundin (2016) and "Woorden doen ertoe" [Words Matter] (n.d.) published by Tropen Museum, Afrika Museum, Museum Volkenkunde and Wereldmusuem, which contains sensitive and problematic terms which are used to label items in (colonial) collections. For the international debate, see for instance Wright (2019). Furthermore, we notice that the debate is predominantly framed as seeking inclusiveness. We are very critical about the use of this concept of inclusiveness since inclusion does not necessarily dismantle dominant power structures and it is even likely that in such reasoning the deep memory infrastructures remain untouched.
Against this background, we critically investigate and discuss in this article some expectations and possible implications of digitization of colonial records. A critical stance is important, since, as Roopika Risam reminds us, digitization of archives is 'often heralded prematurely for their contributions to the (…) project of decolonization' (Risam 2019). It means that we need to investigate in what circumstances digitization may offer new possibilities to decolonize the former colonial archives or when it only reaffirms the dominant power structures in a different form. In order to formulate meaningful answers to these questions, we will not limit the concept of decolonization to the repatriation of colonial records. Digitizing colonial records means creating a new space that brings forth new problems and issues. In any debate on decolonization, we believe it is necessary to first firmly portray the colonial, and therefore we include a discussion on the coloniality of archival infrastructures through the pervasiveness of the colonial cultural archive. Furthermore, we focus on the mechanisms of digitization and its epistemological implications with respect to the aim of decolonizing the records. After all, digitization of paper files is much more convoluted than just changing the form by which information is made accessible for users. The records are parts of a larger whole and from a traditional archival perspective inextricably connected to the recordkeeping system. Recordkeeping systems are representational systems. Elizabeth Yakel reminds us that '[r]epresentational systems are both manifestations of a culture as well as the infrastructures to support that culture' (Yakel 2003, p. 6). Recordkeeping systems are not neutral. They were developed to serve specific informational needs of institutions and bureaucracies. It also means that, due to the fabrications and exclusions in the process of colonial record creating, the authority of the historical archive is not in the least self-evident and should be challenged (Parry 1995, p. 37).
What interests us here is getting a better understanding of the implications of digitization as it relates to creating new representational systems of records which were created in past centuries. How do newly created representational systems relate to the 'original' systems? How can new perspectives be facilitated without erasing previous structures? Such questions are applicable to representational systems in general, but are particularly relevant to colonial archival legacies which are subject to a critical and intensive societal debate. We should keep in mind, as Verne Harris has emphasized, that '[l]egacies are never received; they are only ever made and re-made' Harris, in turn, relied on Derrida, who claimed that '[t]o inherit is to select, to sort, to highlight, to reactivate' (Harris 2011, p. 117). It also means, as Nathan Mudyi Sentance recently argued, that archives and museums have a large 1 3 responsibility in dealing with colonial pasts. They cannot hide anymore behind a wall of innocence by stating that they are acting as neutral intermediaries between past and present. Inaction by memory institutions is not neutral either. Inaction means that current oppressive structures are supported (Sentance 2018).
This paradox of colonial archives is central in our work-that there is something innately colonial in the recordkeeping systems that cannot, and should not, be removed, and, yet, must also be decolonized. That it is innate comes from the colonial structure that created the records and the system. That it should not be removed is based on the idea that mishandling the removal process of the colonial nature of the structure would further entrench it. Not only would it be hiding the colonial past, it would take away the ability to continue to learn from and about the colonial period, it would be a disservice to those who suffered under colonialism and would misrepresent both the past and how information was created, stored and accessed. The paradox comes in that the colonial archive must undergo some sort of transformation, something that many would describe as decolonization. The information must be more accessible, more open and used in such a way as to promote decoloniality.
We argue that digitizing colonial heritage is a conscious act of (re)activation. By just digitizing colonial archives, representations of disputed societal (read: colonial) structures are reactivated. Physical paper-sites of colonial memory transform into digital websites of colonial memory which have not only conquered successfully the boundaries of time, but also of space. It might be a next step in a process of what we label as archival colonization unless the recordkeeping community finds ways to develop infrastructures that facilitate multiple perspectives, permits different voices, enables different forms of agency and meaning-making. We found inspiration in the concept of what Bhabha and Licona have termed a third space. This third-space concept is based on the notion of cultural difference, which Bhabha describes as 'the process of the enunciation of culture as "knowledgeable", authoritative, adequate to the construction of systems of cultural identification' (Bhabha, 2011). It means that third spaces are sites of meaning-making which go beyond dichotomous opposites of either/or and create a both/and perspective (Bhabha 2004;Licona 2005;Licona 2012;Lee 2017). Third spaces 'constitute the discursive conditions of enunciation' and 'ensure that the same signs can be appropriated, translated, rehistoricized and read anew' (Bhabha, 2011). We investigate whether and how archival infrastructures could be based on this concept of third spaces as potential solutions towards archival decoloniality.

Colonial archives
The year 1595 can be seen as the starting point of Dutch colonial expansion. Four ships with 240 crew members under the command of Cornelis de Houtman left the Netherlands and sailed via the Cape of Good Hope in Africa to Bantam on Java. Over two years later, in August 1597, three ships with only 87 crew members returned to Amsterdam. Despite the many difficulties, losses and hardships, the voyage demonstrated that Dutch ships were able to sail to Asia. A few years later, in 1602, the United Dutch East India Company was established to organize the trade to the East Indies in a symbiotic alliance between state and commercial interests. In 1621, a similar organizational model was used when the West India Company was founded to regulate Dutch trade between West Africa and the Americas. Both companies became agents of colonization of areas in Asia, Africa, South America and the Caribbean. In the late eighteenth century, both companies were dissolved and the territories they held reverted to the Dutch state and colonization, now under the direct responsibility of the Dutch state, entered a new stage.
The Dutch East and West India companies were not only major players in exchanging goods and people between the continents, but also large information producing and exchanging entities. It is estimated that the Dutch East India Company produced roughly 100 million pages of records. Approximately 25%, or circa 25 million pages have survived: 2500 m are kept in the National Archives of Indonesia, 1200 m in the National Archives of the Netherlands, 450 m in the Western Cape Archives and Records Service in Cape Town, South Africa, 310 m in the National Archives of Sri Lanka and 64 m in the Tamil Nadu Archives in Chennai, India (Guleij and Knaap 2017, p. 8). The numbers of created and preserved records of the second phase of colonization, after the bankruptcy of the Dutch East India Company, are even more impressive. Although it is difficult to demarcate the category colonial records, we know that approximately 9 km of records created by Dutch colonial institutions after liquidation of the Dutch East India Company are kept in the National Archives of Indonesia. The records of the Dutch Ministry of Colonies in the National Archives in The Hague occupy almost 3.5 km of shelves, and another 150 m of records are kept which were transferred from the Dutch East Indies to The Hague (Bos-Rops et al. 1982;Otten 2004). The archives in the Netherlands are a hybrid collection. From an institutional and traditional provenance perspective a large part of the archives which were created by the Ministry of Colonies are indisputably Dutch. The policy object of the ministry was the colonies, and, in many records, local indigenous knowledge and information, although usually not inscribed by locals but by Europeans, is recorded. Parts of the archives can, however, be framed as colonized Indonesian archives, such as the archives and documents looted and seized by intelligence services during the war of decolonization.
In 1996 and 1997 preservation, or better the lack of preservation, of what was regarded as overseas Dutch cultural heritage suddenly received political attention in parliamentary debates in The Netherlands. After having visited some remnants of Dutch colonial presence in India, member of parliament Van Middelkoop was disconcerted by the poor condition of overseas Dutch colonial heritage. He criticized the lack of interest and care in the Netherlands for its overseas cultural heritage 'as narrow minded and based on feelings of shame and guilt for its colonial past' (Tweede Kamer 1997). His alarmist view resulted in additional funding from the government to safeguard Dutch overseas cultural heritage. A first inventory of Dutch overseas heritage not only included built heritage such as buildings and ships constructed or used by the Dutch but also records relating to Dutch activities overseas or written in the Dutch language (Tweede Kamer 1996Kamer -1997. This new political attention was a starting point for a cultural policy based on the concept of what was labelled 'mutual', 'shared' or 'common' cultural heritage (Finieg et al. 2008, p. 26).

3
This new policy resulted in, among other things, the worldwide recognition of the value of the aforementioned 25 million pages of VOC records since they were put on the Unesco Memory of the World Register in 2003, with the archives of the West India Company following in 2011.
At first sight, the concept of mutuality may be attractive, but the question is what is meant by shared, mutual or common heritage. It suggests a shared view of established values of the cultural remains between the Netherlands and the former colonies where many of the vestiges are situated. In many cases, the assumption of shared value appeared to be a misconception and maybe even the result of overrating the importance of Dutch presence in the collective memory of the former colonies and in their traditions of writing and telling national histories. Following Van Maanen and Ashworth, it would be more appropriate to admit that what was labelled as shared heritages are 'in fact different heritages that are being experienced although emanating from the same resources' (2016). Dutch historian and archivist Jinna Smit, for instance, ascribed the negligence of the 64 m of Dutch East India Company records in the Tamil Nadu Archives in India to the irrelevance of the Dutch episode to the history of India (Smit 2012, p. 184). Despite their Unesco status as documentary heritage of 'world significance and outstanding universal value', the Dutch interest to preserve and keep these documents was and is very different from the perspective of the Indians. It illustrates, as Australian heritage scholar Laurajane Smith has pointed out, that all heritage is inherently intangible, since 'we are engaging with a set of values and meanings, including such elements as emotion, memory and cultural knowledge and experiences ' (2006, p. 56). It means that emotions such as repugnance, hatred or more mildly indifference can be powerful sources that define engagement with the legacies from the past. Nevertheless, the new policy framework of mutual (later officially changed into shared) cultural heritage gave a powerful impetus to digitization of former Dutch colonial archives.
In the same period of emerging political interest in preserving colonial heritage, scholarly interest in investigating the colonial past was growing. In 1999, the Dutch National Archives and Leiden University launched a programme "Towards a New Age of Partnership" (TANAP), which was 'a Dutch/Asian/South African programme of cooperation based on a mutual past' (TANAP 2001) and fitted perfectly well in the newly defined policy framework of mutuality. The aim of this programme was not only to preserve and to optimize accessibility to the Dutch East India Company archives which were kept in the Netherlands, Indonesia, Sri Lanka, South Africa and India, but also to train students from Asian and African countries in the (language) skills which were required to be able to study these archives for writing their own histories. The TANAP-website displays more than 20 titles of PhD theses, written by predominantly non-European students, making use of the Dutch colonial sources (TANAP 2007). It was one of the successful results of the TANAP programme.
Providing 'digital access and linking of the VOC archives in Asia, South Africa and the Netherlands' was one of the early ambitions of the project. Ten years after the start of the ambitious and successful TANAP-programme, the Dutch National Archives organized a conference on Mutual Cultural Heritage. Nine directors of national archives and other repositories in the world holding valuable archival 1 3 Archival Science (2020) 20:199-220 materials labelled as shared cultural heritage from Brazil, Ghana, India, Indonesia, the Netherlands, South Africa, Sri Lanka, Suriname and the USA of America came together in The Hague to discuss archival cooperation. They agreed that accessibility of the archives of their 'shared past' should be improved through digitization and the creation of a 'single worldwide digital MH [mutual heritage] collection that is linked together and searchable in its entirety' (Smit 2012, p. 181).

Digitizing the Dutch colonial past
In a feverish pace, millions of documents which once were part of governmental recordkeeping systems have been digitized in the last few years by archival institutions all over the world. Finally, after being incarcerated for so long, the documents seem to be liberated from the galling bonds defined by static paper files stored in repositories and arranged based on classification logic of decades, sometimes even centuries, ago. Digitization of archival legacy and publication on the web is widely recognized as being the best way to expand and facilitate public access of records and maximize their value to the nation and the world (National Archives and Records Administration 2014). The Memory of the World programme of Unesco, for instance, not only aims to facilitate preservation of the 'world's documentary heritage', but also encourages making digitized copies available via the Internet (Unesco Memory of the World n.d.). According to figures received from The National Archives in December 2018, of the 125 linear kilometres kept in the repositories of the Dutch National Archives, 3800 linear metres, which is nearly three per cent of the analogue collection, has been digitized. But when we look at different subject areas, the efforts to digitize records which resulted from Dutch colonial rule of the past 400 years have been very successful. Of the total amount of 3800 linear metres of digitized paper records from the Dutch National Archives, 1650 linear metres, or 43%, are the so-called colonial archives. For information about the Dutch East India Company [VOC] records, see the Unesco Memory of the World Register Nomination form (2002).
The Shared Cultural Heritage Policy of the Netherlands has also resulted in (still ongoing) initiatives to digitize colonial archives kept by countries that were previously colonized by the Netherlands. The financial resources that are made available by the Dutch government and by private initiatives to preserve and digitize archival material is the lubricant to be able to execute digitization projects and to bridge divergent interests between the states for carrying out such projects. The Dutch National Archives started projects in cooperation with, for instance, the Tamil Nadu Archives in Chennai, India, the National Archives of Sri Lanka and the National Archives of Indonesia to digitize parts of the archival holdings which are of interest for Dutch colonial history. Furthermore, between 2010 and 2017, the Dutch National Archives digitized the documents of 40 different archival fonds created by agencies in the former colony of Suriname, in total 800 m of shelf length. This digitization project is an interesting and, so far, unique example of how two sovereign states with a long colonial history have dealt with their archival legacy. From the early twentieth century until the independence of the Republic of Suriname in 1975, the colonial 1 3 government decided to send several batches of records from agencies in the colony to the National Archives of the Netherlands 'for safekeeping and on loan, with the explicit stipulation that the archives remain the property of the colony' (Hassankhan 2006). The main reason for transferring them to the Netherlands was the lack of proper housing conditions for records in the colony to prevent further decay due to destructive climatological conditions. Since the Republic of Suriname is considered the legal successor of the colony of Suriname, the explicit condition that the records were sent on loan and remain the property of the colony became the motive for the government of Suriname to insist that the archives should be returned. Finally, in 2009, after many years of sometimes difficult negotiations, the Dutch and Suriname governments agreed on the conditions for returning the 800 m, in total 5.5 million pages, of archival documents of the former colonial institutions from between the seventeenth and the twentieth centuries. Digitization of these records and online availability were one of the conditions of the Dutch state. After a stay in the Netherlands of more than one century, the 800 m of original documents were sent back to the legal owner of the archives, the Republic of Suriname, in 2017. This Dutch restitution of archives to Suriname often receives appreciation. Recently, James Lowry (2019a) quoted this as an example of literally decolonizing archives.
There are more examples of realized projects of digitization of colonial archives. The Dutch non-profit organization The Corts Foundation digitized about 885,000 pages of Dutch East India records in the National Archives of Indonesia (Sejarahnusantara, n.d.), and the Dutch National Archives have started to digitize large quantities of nineteenth century colonial records kept by the Indonesian National Archives.
Digitization is almost always presented as a solution (for instance for preservation, providing access, digital repatriation) and rarely considered as a socio-technical space that raises new issues and questions. Although there are many reasons to critically examine digitization, a critical approach to the processes of creating digital surrogates is still more the exception than the rule. Recently, Zinaida Manžuch gave an overview of the many ethical issues which emerge from digitization, such as biases in selection of what to digitize, dominance of metadata schemas of the Western world, neglect of the needs and values of indigenous peoples and neglect of different, non-institutional perspectives on giving/restricting access to digitized archives (Manžuch 2017). Others, like Ala Rekrut and Charles Jeurgens, have paid attention to the specific values of materiality and the dilemmas of digitization (Rekrut 2014; Jeurgens 2013b). Using such a critical approach, we will examine the concepts of coloniality and decolonization and their relationship to digitization and its creation of new digital archival infrastructures.

Concepts of decolonizing
Examining concepts and definitions of decolonization means being confronted with a cacophony of views and approaches. In the European context, the most straightforward meaning of decolonization can be defined as a transfer of power by which colonies become independent of a colonizing country (Encyclopaedia Britannica

3
Archival Science (2020) 20:199-220 n.d.). But it would be far too simple to limit decolonization to the change of regime and change of political dominance.
Scholars like Quijano, Mignolo and Walsh have shown that 'decolonizing' is a convoluted term and argue that 'the answer to the question 'what does it mean to decolonize?' cannot be an abstract universal'. It depends on who, where, why and how (2018, p. 108). See also Bailkin (2015) and Mir (2015), in which they emphasize that decolonization is an ongoing process and looks very different from a formal state/political perspective or from a perspective of individual experience and social interactions. It is not our aim to give a complete overview of the different concepts of decolonizing, but it is important to emphasize that in our approach in which we focus on colonial archival legacies, decolonizing these archives is seeking ways to dismantle the hegemony of the custodial institutions in archival knowledge production and acknowledging that archival infrastructures are the key to agency. After all, the archive is, as Aljoe et al. have argued, 'a knowledge event' (Aljoe et al. 2015, p. 259). To put it simply, when we use the phrase decolonize in this sense, we are referring to dismantling that which is colonial in the archival infrastructures, which is why we must include a discussion on the coloniality of archives.
The current scholarly debate on decolonization beyond the perspective of political regime change is conducted most intensively in the context of break-away settler colonialisms where cultural resurgence of First Nation, Native, Indigenous and Aboriginal peoples in these areas is a strong impetus for this debate. Finding a balance in deeply divided societies where indigenous communities suffered tremendously from different forms of colonial oppression and the necessity to transform the knowledge infrastructures that acknowledge multiple voices and perspectives is one of the important motives behind the current decolonization debates (McKemmish 2017). This is fundamentally different from the debates taking place in countries with a colonial past which was not based on massive settler migration but on extracting resources from the colony for the benefit and enrichment of the home country. In this respect, the reality of Suriname, like many countries in the Caribbean, is even more complicated. They are examples of what we could best describe as colonies based on forced settler migration of enslaved people in order to extract resources for the benefit of colonizers and enslavers. After the demise of these colonial empires, archival issues between the former colony and colonizer were usually limited to the well-known issues of ownership which sometimes resulted in repatriating or copying claimed archives (Auer 1998;Lowry forthcoming). Bearing in mind the large-scale digitization of colonial archives and making these available online, the time has come to conduct the debate on these archives not only from a viewpoint of ownership, but also from an epistemological perspective.
The Dutch examples that we draw upon can be regarded as either colonial or colonized archives. We define colonial archives as those that were created by former colonial institutions in the era of colonization, such as those of the East India Company, West India Company and the Dutch colonial administration which began in the nineteenth century. Part of these records was created in the former colonies and part in the former metropole. For colonized records, we refer to records which were originally created, owned and used by local institutions and people but were collected, looted, bought or copied and shipped to Europe. These colonized archives 1 3 are those which James Lowry acknowledges 'may literally lead to the decolonisation of collections through repatriation' (2019b). Again here, decolonization is linked to the return of archives. Though nuanced and different, the settler and bilateral state decolonization debates begin to merge in the shifting mentality we have mentioned, one that transforms our thinking and which can be in line with what Ghaddar and Caswell describe as a 'decolonial archival praxis' (2019).
The debate on decolonization has much common ground with the discussion of post-colonialism. Like post-colonialism refers to much more than just designating the period after colonization, the term decolonizing encompasses more than the political process of dismantling the domination of a state over distant territory and its people. The New Zealand scholar Simon During once characterized the postcolonial aspiration as 'the desire of decolonized communities for an identity' and that identity should be 'uncontaminated by universalist or Eurocentric concepts and images ' (2003, p. 125). Archival scholar Jeannette Bastian argues that although the term 'postcolonial archives' is in the first place an intellectual space and a state of mind to combat master narratives, it might also be 'a real place, a place where archival theory is thought about differently and in non-traditional ways' and refers to the importance of recognizing cultural expressions beyond the traditional textual focus (2013, p. 127). Although gradually more attention is given to different archival traditions, infrastructures and perspectives (Falola 2017) such as community archives (Stevens et al. 2010) and practices of non-documentary forms of knowledge and memory formation by historically marginalised groups (McEwan 2003), and despite continuum concepts having gained ground, archival theory is still highly dominated by European concepts which are rooted in the ideas as formulated in the Dutch Manual, and by archivists such as Jenkinson and Schellenberg. They all have a strong institutional and government-centred focus and a rather limited and one-dimensional view of provenance (Bastian 2013, Wright 2019. Although poststructuralist critique may have raised the awareness of archival power and to a certain extent even dismantled meta-narratives as the large unifying frameworks for interpretation and understanding the past, we argue that this does not mean that memory institutions have become critical spaces that interrogate their own role in maintaining or breaking neo-colonial power structures or invest in and facilitate a multidimensional, pluralized use of the colonial heritage they keep. The 'deep memory systems', by which Verne Harris meant the national archival and heritage structures (Harris 2011, p. 117), are not very susceptible to change.
Matthew Kurtz wonders under what conditions it would be possible to create a 'postcolonial' archive, and we could expand that question by asking what options we have to decolonize colonial archival legacies. Kurtz contends that 'a paradox echoes through the construction of an archive that strives towards "the postcolonial"', and he shows this paradox with a project in arctic Alaska that aimed to recover indigenous memory by creating an 'oral history archive'. He concludes that 'the institutional practices of building an oral history archive (…) re-inscribes many practices and aspects of colonialism ' (2006, pp. 65, 89). This paradox refers to the root of the problem we are confronted with in discussing decolonization of archives. Decolonization was originally the struggle of the native population to oust the colonizer from their colonized lands with the aim to form a sovereign state or community.

3
Archival Science (2020) 20:199-220 Decolonization did not, however, terminate the cultural, epistemic and hermeneutical principles upon which these colonial structures were built. Peruvian sociologist Anibal Quijano introduced the concept of coloniality of powers, which refers to the continuity of colonial power relations even though the political/administrative form of colonization has ended. This concept of coloniality is the underlying logic of all Western colonialisms. In that sense, decoloniality is in the first place the capacity of unravelling the underlying epistemes and ontologies and next the capacity to create multiple epistemologies and ontologies which is not just 'setting up a "new" school of thought within Western cosmology' (Mignolo and Walsh 2018, p. 228). Coloniality is the epistemic basis of colonialism and colonization. Understanding this mechanism is vital for heritage institutions that want to decolonize their holdings and infrastructures by digitization projects. Authors such as Simon Gikandi, Roopika Risam and Nicole Aljoe have emphasized that colonial violence is often perpetuated in the digital record. Colonialism was firmly based on obsessive recordkeeping. The colonial archive was the nerve centre of colonial rule. But as Spivak reminds us, the colonial archive was first and foremost a hall of mirrors: colonial administrators constructed a selective image of the social order and through these fictions they governed colonial society (Spivak 1985, p. 249), or as Gikandi argues, writing and record keeping 'was a form of violent control; the archive was (…) the key to the ideology of white power' (Gikandi 2015, p. 92) which positioned enslaved people as non-subjects and property. Aljoe et al. emphasize that these processes of obfuscation, erasure and eradication produce 'nonknowledge' which does not change by making the archive digital. The paradox of the digital archive is that it 'aspires to knowledge and transparency, but in its coloniality it enacts erasure and violence' (Aljoe et al. 2015, p. 260). And this is exactly the problem archives are facing.
But it does not mean that the digital cannot contribute to removing the coloniality in newly created archival infrastructures. Since coloniality perpetuates after formal decolonization, it means that efforts of decolonizing the archive begin with decolonizing the underlying logic, the Western knowledge structures and epistemologies. Preventing perpetuating coloniality in newly designed digital archival infrastructures requires 'epistemic reconstruction' or, as Walter Mignolo argued, it entails being 'epistemically disobedient'. It means that the categories and limitations which are defined by Western thought are not taken for granted (Mignolo 2011). Decolonial thinking, Mignolo explains, 'strives to delink itself from the imposed dichotomies articulated in the West, namely the knower and the known, the subject and the object, theory and praxis. This means that decolonial thinking exists in the exteriority (…). It exists in the borderland/on the borderlines of the principles of Western epistemology, of knowing and knowledge-making' (Mignolo 2017, p. 42). Decoloniality means epistemic and ontological pluriversality. Institutions like universities, museums and archives play an important role in perpetuating coloniality. Western museums and archives as 'houses of knowledge' were also institutions which collected, categorized and displayed non-European artefacts (stolen, looted, purchased, etc.) (Jeurgens 2013a) and tell 'the single and all-encompassing narrative of western civilizations caged in western museums' and archives (Mignolo 2013). But at the same time this means that such institutions can play an important role in changing the narrative. The Early Caribbean Digital Archive (ECDA) founded by Nicole N.

3
Aljoe and Elizabeth M. Dillon from Northeastern University (Boston) is an interesting example of an experiment in which digital means are used to decolonize colonial archives. This initiative is based on the 'writing back paradigm', in which digitized records from existing archives are remixed and reassembled and new materials are added. For instance, slave narratives are extracted from texts written by Europeans and these are remixed to shape a new digital collection of narratives that speak to one another in new contexts. This 'epistemic disobedience' offers possibilities of re-archiving 'which invite (if not require!) us to disrupt, review, question and revise the colonial knowledge regime that informs the archives from which we draw most of our materials' (ECDA n.d.).

Coloniality of archival infrastructures
In archival science, archives are defined as places, institutions, on the one hand and interrelated documents produced or received in performing activities on the other. If any debate on the potential of decolonizing archives is to take place, then the coloniality of archival infrastructures must be at the forefront. Ellen Ndeshi Namhila defines colonial archives as 'both archival records and archival institutions that were created and maintained under colonial rule, that is, in the political context of a territory that is not sovereign but ruled by another country in a colonial situation' (Namhila 2016, p. 114). Traditionally, archival institutions supported the 'ideological construction of national histories' and as such they 'have played a significant role as 'agents' or 'infrastructures' of colonial oppression' (Friedrich 2018, p. 421).
Namhila stresses that colonial archival heritage survives the political frameworks that were responsible for unequal relationships. Based on her critical empirical research concerning the content of colonial archival holdings in Namibia, she found out that there is a huge discrepancy between the availability of person-related records of whites and indigenous people in the country. Estate records of whites have been preserved and indexed by the National Archives of Namibia, while the 'corresponding "native" records present a picture of widespread neglect, non-transfer, unauthorized destruction and lack of processing and indexing' (Namhila 2016, p. 122). She warns against the danger of maintaining the colonial recordkeeping legacy in the current archival practices. It is precisely this epistemological legacy, the forms of categorization, the language used in the descriptions and other metadata, the conscious but often even unconscious dominant white Eurocentric perspective of the archival interfaces which are taken for granted that make these colonial infrastructures so powerful. The observations made by Namhila emanate from a critical disciplinary self-examination which is the methodological yield of the 'archival turns' in several disciplines by which archives became epistemological sites, cultural practices and metaphors for storehouses of knowledge (Ketelaar 2017). Investigating coloniality in current archival infrastructures and subsequent attempts to decolonize archives as places/institutions and interrelated documents of the archival turns is highly indebted to these archival turns.
In this respect, the concept of the cultural archive as defined by Edward Said might be helpful to get a better understanding of the relationship between coloniality 1 3 Archival Science (2020) 20:199-220 in thinking and how this thinking is materialized in currently shaped archival infrastructures, which in turn is directly connected to the discussion of decolonizing archives and their infrastructures. The concept of the cultural archive refers to the impalpable storehouse of emotions, ideas, classifications and stories which determine people's worldview. Said reflects on the self-referentiality of this mechanism, since '[i]n your narratives, histories, travel tales, and explorations your consciousness was represented as the principal authority, an active point of energy that made sense not just of colonizing activities but of exotic geographies and peoples. Above all, your sense of power scarcely imagined that those 'natives' who appeared either subservient or sullenly uncooperative were ever going to be capable of finally making you give up India or Algeria'' (Said 1994, pp. xxiii-xxiv). We argue that the structures, classifications and interfaces of the material archive are for a large part the result of and built on this cultural archive. Recently, the concept of the cultural archive has received new attention by Dutch scholar Gloria Wekker. She uses the term to describe the still dominant thoughts, emotions and points of reference within a society regarding race and ethnicity and she shows the subtle power of the cultural archive on meaning-making processes and conceptualization of the self and the other in daily life. Partially quoting Ann Stoler, the cultural archive refers to '"a repository of memory" in the heads and hearts of people in the metropole, but its content is also silently cemented in policies, in organizational rules, in popular and sexual cultures and in common-sense everyday knowledge, and all of this is based on 400 years of imperial rule' (Wekker 2016, p. 19). Because the cultural archive is implicit, persistent and omnipresent, it is often extremely difficult to accept the inconvenient truth of the results of excavating the codes of the cultural archive and to disclose the mouldy substance on which the cultural archive is built. As a knowledge system, the cultural archive is 'imbued with an 'attitude' and a 'spirit' which assumes a certain ownership of the entire world, and which has established systems and forms of governance which embed that attitude in institutional practices' (Smith 2001, p. 56). Because of the persistent and ubiquitous nature of the cultural archive it is, especially for those who are part of it, difficult to identify and acknowledge how their previously uncontested pattern of thinking may keep oppressive knowledge infrastructures alive. Decolonizing archival infrastructures starts with contesting the colonial logic and a thorough understanding of the implicit and explicit beliefs and practices of the cultural archive, the cultural system of classification and representation on which the material archival infrastructures are built. Like mastering, reading along the grain is required before one can start thinking of reading against the grain, and recognizing coloniality of thinking is required before one can start thinking of decolonizing archival infrastructures.

Signposts of coloniality drenched in good intentions
In this section, we discuss some Dutch examples of what we call 'signposts of coloniality'. How can we recognize coloniality in the existing archival frameworks and infrastructures? In the next section, we make some suggestion of how 1 3 these examples could have their coloniality removed or mitigated, or essentially, be decolonized.
Between 2001 and 2007, a team of 17 Indonesian archivists, with assistance of three archivists from the Dutch National Archives, worked on improving accessibility and material condition of the 2500 m of Dutch East India Company archives which are held in the National Archives of Indonesia. The project resulted in, among others, compiling an inventory of the archives of the Dutch East India Company and the local institutions in Batavia (Jakarta). The introductory texts in the inventory, which contain an institutional overview of the record-creating agencies and describe the history of the archives and give justification of the choices made in compiling the inventories, are published in English, Bahasa Indonesia and Dutch, but are exclusively written by Dutch archivists and historians. The voluminous inventory itself, containing the descriptions of the series and records of 15 different archival fonds, is only available in Dutch. The Dutch descriptions were produced by Indonesian archivists under supervision of the Dutch archivists (Balk et al. 2007). The inventory is the ultimate key to get access to the 2.5 km of Dutch East India Company records which are kept in the National Archives in Jakarta. Researchers who want to use the inventory need to be able to read the language of the former colonizer. The same is the case with the inventories related to the colonial period after the Dutch East India Company was dissolved. The National Archives of Indonesia contains approximately 9 km of shelves filled with records of colonial institutions operating between 1800 and 1942. Almost all inventories of archives of former colonial institutions that have been produced by the National Archives of Indonesia in the last 25 years are written in Dutch. Although the introductory texts are in Bahasa Indonesia, the descriptions are in Dutch, which sometimes leads to peculiar and odd situations. For instance, in the Inventaris Arsip Keresidenan Soerabaja (inventory of the archives of the residency of Surabaya), which was compiled in the 1990s, a description in Dutch reads 'ingekomen brieven van de Panumbahan van Pamekassan (translated in English 'letters received from the Panumbahan of Pamekassan, 1831-1832') with a notification, again in Dutch, that the letters are 'written in Malay' (National Archives of Indonesia, inventory number 350). The often-heard argument used to defend the Dutch language choice is that it follows the logic of the dominant language of the records, which are for a large part written in Dutch: if a researcher wants to investigate the documents, knowledge of Dutch language is required anyhow. It would be a somewhat balanced professional reasoning if, for instance, the calendars of medieval archival documents (the calendars contain descriptions and short summaries of the content of medieval charters which are often written in Latin) made by Dutch archivists were compiled in Latin instead of Dutch. It may be not a surprise that these descriptions and summaries are written in the Dutch language, which makes sense even for researchers who do not master Latin. Based on the descriptions in the inventory, they may decide whether it is worthwhile to organize assistance for translation. The same happened with the many records, pamphlets and other documents which were seized in the 1940s by the Dutch military forces and intelligence services operating in Indonesia and which are kept by the Dutch National Archives. Although most of these documents are in the Indonesian language, the descriptions of the documents in the inventories are in Dutch (National Archives of the The inequality of power becomes even more embarrassing when digitization of such archives is at stake. Digitization is expensive, requires facilities and knowledge and above all, and the digital infrastructures are designed to and based on the premise to cross borders and to easily distribute and exchange information. Therefore, interoperability has become a key concept in mass digitization projects. Interoperability is not only a technical concept, but 'emerges today as an infrastructural logic, one that promotes openness, modularity and connectivity' (Thylstrup 2018, p. 68). A logic which is defined by those who have the means, power and interests, which is one of the negative sides of what Trudy Huskamp Peterson has called 'crosscultural digital copying projects'. She describes what happens if an institute from a rich country has identified archives in less wealthy parts of the world which are of interest to researchers it serves. Under the pretext of preserving records that are not stored in good conditions, projects are launched to digitize these records and to ensure that the information in the records will survive. The acquiring institution pays the staff to digitize, donates some equipment used in the project, provides a digital copy to the donor and takes the digitized material home and puts it on its website (Peterson 2010). Different perspectives on ownership, sovereignty, legal and moral authority of these (digitized) archives are regarded as hampering for the 'universal' network logic of interoperability and connectivity. Keith Breckenridge has critically addressed the problematic aspects of digitization projects in South Africa in which joint digitization programs between South African institutions and foreign, western institutions were felt and perceived as projects of digital imperialism, heritage theft and cultural pillaging. Digitization of archives perpetuates the unequal relationship between the global North and South and only because of temptation of financial aid is a new form of imperialism reinforced (Breckenridge 2014). The national archivist of Suriname hints at such patterns of coloniality in covert terms in an interview with a Dutch professional journal, when she refers to the requirement of the Dutch government to digitize the 800 m of Suriname records before they could be returned. Although Surinamese legal ownership of these records was not contested, and the Suriname government initially was not in favour of digitization and even less of online publication, it was conditional for restitution. Digitization was regarded as the only way to guarantee future access to these archives for people in the Netherlands (De Nijs 2017). Sometimes there is disagreement between the institution that owns the records and the institution that digitizes the records on the mandate to publish the digitized records online. The owners of the records, the sponsors of digitization and the subjects of the records may sometimes have very different concepts of what they consider as 'sensitive' (Leopold 2013). But even if parties agree, it does not automatically lead to a balanced relationship and to equal opportunities to make use of the digitized records. Asymmetric availability of required technical infrastructure can be a serious obstacle to have equal access to the records. Furthermore, although distributing digitized colonial records via the web may facilitate the availability of such records, it does not necessarily improve accessibility in the sense of usability. The Indonesian website https ://sejar ah-nusan tara.anri.go.id/id/, which is the result 1 3 of cooperation between the Dutch non-profit organization Corts Foundation and the Indonesian National Archives, can serve as an example. As in the paper-based infrastructure discussed earlier, the digital infrastructure is also based on the Dutch language. Inventories which are put online contain introductory texts in Bahasa Indonesia and English, but the archival descriptions are predominantly only in Dutch and sometimes in English.
Colonialism also continues to exist in more subtle forms. An example which recently received some attention in the Netherlands may illustrate this. The Dutch East India Company offered jobs to thousands of people as sailors, soldiers and servants working in the trading posts. Since there was always a risk that people would die while serving the VOC, many officials went to one of the VOC-notaries in Batavia or the Cape of Good Hope to notarize their will. Copies of these wills were sent to the headquarters of the Company in Amsterdam. Nowadays, these wills are important sources providing insights into the private lives of common people. Recently, the Dutch National Archives made available online an index with names of more than 10,000 testators and digitized the corresponding handwritten documents (Oost-Indische Testamenten, n.d.). The index of names was made by archivists in the nineteenth century. Although husband and wife frequently went together to the notary into draw up a will, and in spite of the fact that in such cases both names-husband and wife-are equally mentioned in the testaments, the nineteenth century index only refers to the male testator. The nineteenth century view of male dominance and obfuscation of female presence in the archival infrastructure is preserved and continued in the twenty-first century search infrastructure. 1 Examining the notarial deeds even more carefully, it appears that in many instances not only the names of husband and wife, but also the names of local, indigenous people (men and women) in many different roles are mentioned, such as beneficiary, housemate, creditor, debtor or enslaved property. Although their voices were recorded, a nineteenth century archival world view, materialized in the compiled index, not only obscured women, but also indigenous people. The real problem in the context of our discussion is the perpetuation of this nineteenth century archival worldview in a new, revitalized digital archival infrastructure: it renders women and indigenous people invisible and is an example of epistemological inertia. To repeat what Nathan Mudyi Sentance has argued: inaction is not neutral. Inaction supports the oppressive archival infrastructure from the past.

Possible direction: decolonizing via third spaces
In the perspective of our debate on coloniality and decoloniality, one of the key issues is the monopolistic power of memory institutions such as archives and museums in defining, describing and communicating the archives they have in custody as sacred end-products. Most archival institutions in the world position themselves 1 3 Archival Science (2020) 20:199-220 outside the process of archives creation. They receive, describe, keep and give access to the records which are created by an agency or person, conceptualized in the archival fonds. The legitimation is found in a short-sighted interpretation of the main archival principle of provenance. The archival concept of provenance traditionally puts all emphasis on identifying the single organization or person that created or received records in performing their activities. Despite early notions (since the 1960s) that this one-to-one relationship between administrative actor, activity and document is too simple and too narrow and despite strong arguments to broaden the concept of provenance in which, for instance, the subject of the record is included, successful attempts to translate such concepts in archival infrastructures are hardly known. In recent years, the concept of provenance has received even more bitter critiques such as being 'a relic of the colonial and imperial era' which emerged 'at a time when most people were structurally if not legally excluded from ownership; ownership of their own bodies, minds, labour, property and records' (Drake 2016). Chris Hurley suggested parallel provenance (Hurley 2005a, b); Tom Nesmith introduced the concept of societal provenance (2005), and recently Gracen Brilmyer proposed an assemblage approach to records to do justice to the 'multiplicity of their subject's experiences and many co-creators of records' (Brilmyer 2018, p. 103). Even though records continuum thinking has moved archival scholars away from the fixed one-dimensional relationship between creator and the record, we also must conclude that current archival systems still do not have the ability to mirror the above-mentioned complexity and multidimensionality of records. Renaming and redescribing records, as the National Archives in Singapore did with some digitized parts of the Migrated Archives, 2 could be a way to 'decolonize' descriptions, but it is again a one-dimensional way to deal with the problem and is certainly not a thirdspace perspective.
In a recent article, anthropologist Kate Hennessy (2016) explored the role of digital technology in redefining the cultural relationship between heritage institutions and originating communities in Canada. The case she describes is the MacFarlane collection, named after Roderick MacFarlane, who was in the 1860s a Hudson Bay Company trader. At the request of the Smithsonian Institution MacFarlane collected thousands of objects of the Inuvialuit community. The collected objects are scattered across various museums in Canada, USA and Europe. Recently, the whole collection was digitized and included into the Reciprocal Research Network (RRN), which is an initiative of the Museum of Anthropology at the University of British Colombia and three Indigenous communities (Musqueam Indian Band, Stó:lō Nation/Tribal Council, and U'mista Cultural Society) to create innovative forms of access to cultural artefacts. What interests us most here is the different perspectives that can be used to present artefacts within RRN (Rowley 2013). Users of the site can, via a provided link, view the original item record from the custodial museum, while also 1 3 reading a more in-depth description, item history and cultural context provided by members of the RRN community. RRN is an infrastructure which is not designed as an end point which exclusively presents 'authorized' descriptions but is designed on continuum principles. The infrastructure provides opportunities for communities to compile meaningful sub and parallel collections based on the items that are available within RRN, to make their own descriptions without approval from the museums that have the objects in custody and can decide to whom they give access to see and use their digitally curated objects and collections (RRN n.d.). RRN can be seen as a promising example of a third space, which creates opportunities that go beyond the traditional dichotomy of being either this or that. Expanding upon strategies implemented by RRN is a part of our ongoing research, which we plan to develop further in a future article using specific colonial archive cases held in the Netherlands.
We imagine such networks to hold the ability to undo some of the coloniality we have encountered in the archival examples provided. With searchable community-generated descriptions, seized archives could be highlighted or given more and different context(s) than what was known by the intelligence officer creating the attached metadata, or the names of women and indigenous people could be added to item descriptions-all this while preserving the 'original' description as created by colonial authorities or later archivists to record the changing contexts. An obvious community addition would be translations from Dutch or English into Indonesian for the aforementioned inventories. Such an action would not be an end; however, and knowledge of the archives' existence would still need to be spread, among other work. This is a fundamental barrier to archival access and a prerequisite for decolonizing the archives. Users must know where to look, what terms to search. Digitization alone cannot bring decolonization, but in tandem with designing new, third space infrastructures that facilitate multivocality, that offers possibilities to add new content that provides new contexts-preserving and yet deconstructing the old recordkeeping systems-the process of decolonization can continue.

Concluding remarks
The central question we wanted to investigate in this article was whether digitization of colonial archival legacies can contribute to decolonizing the archive. We moved away from the prevailing debate in post-colonial societies with substantial indigenous communities by taking the Dutch colonial archival legacy as an example of a different type of colonization which did not result in a dominant European settler community in the former colonies. We also did not take the dominant lens of decolonizing records by focusing on forms of repatriation. By taking the lens of archival infrastructures as representational systems of cultural knowledge, interests and views, we were able to illustrate that digitization in not merely an innocent and neutral act of changing the technical form of documents. Instead, we illustrated, without exhausting all possible aspects, that digitization of colonial and colonized records is a complex interplay with many different variables at stake which ultimately determine and form a new archival representational infrastructure. We argued that a possible solution to escape the paralyzing paradox of decolonizing archives is located 1 3 Archival Science (2020) 20:199-220 in rethinking and reshaping the archival infrastructures, by which we mean the conceptual and technical structures which shape the interface between the documents created in the past and the user of today. The third-space perspective, based on understanding the concept of coloniality of powers, might be a fruitful attempt. Archival institutions that have colonial and colonized records in custody cannot be held responsible for the processes in which these records were created, but they are responsible for the interfaces, the archival infrastructures, the representational systems they create to define, manage, categorize and give access to these records. Decolonizing these archives means investigating and understanding the variables which shape the archival infrastructure, such as the conceptual scope of the principle of provenance, selections that are made what to digitize, forms of description, or forms of agency in giving access. The way each of these variables are operationalized in the infrastructural framework may bear traces of coloniality. Awareness and understanding of the complexity of these mechanisms may contribute to the process of decolonizing the archive. In our search for promising examples of what we consider as third-space infrastructural frameworks, we think that the RRN is a valuable example for archival institutions. It offers interesting opportunities to combine digitization of documents with different communities using their agency to define and communicate their perspectives without interference from the archival institutions that have the records in custody.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/. 1 3