Between Light and Dark Archiving Annet Dekker Post-doc Research Fellow, London South Bank University annet@aaaan.net Abstract Some people argue that the digital archive is an oxymoron (Laermans and Gielen 2007) or more akin to an anarchive (Ernst 2015, Zielinski 2014). Derrida used the word anarchive to signal that “what remains unvan- quished remains associated with the anarchi.” Ernst relates it to the digi- tal archive and describes the anarchive as something that cannot be or- dered or catalogued because it is constantly re-used, circulated, and ex- panding, and thus only a metaphorical archive. Similarly, Foster describes how the anarchival is about obscure traces rather than absolute origins, emphasising the incomplete, which may offer openings to new interpreta- tion, or ‘points of departure’ as mentioned by Foster (2004). These vari- ous descriptions imply that digital archives, and in particular Web-based archives, function less as a storage space and more as a recycling centre in which the material (the archival document, if one can still use this term) is dynamic. In other words, the default state of the digital archive is re-use instead of storage, circulation rather than centrally organised memory, constant change versus stasis. How to capture and retrieve all this data, information and documentation, but more importantly, in what way does archiving take place on the Web? In what follows I examine projects by artists who in various ways explore the challenges of online archiving. These examples show how information and data is captured and archived on the Web. In particular, how it becomes a networked environment, or performance space characterised by the transition from objects to pro- cesses. This new situation, I argue, means moving between dark and light archiving, and it’s the place where a new method of networked co- archiving emerges. Keywords Online archive, networked data, curation, preservation Between Light and Dark Archiving 134 A tension between light and dark, or heavy and light Archives are shaped and defined not merely by their content, but by the nature of their structure and the systems that are used to document and archive records. This applies to archives in any medium whether they are paper, video, or photography, but it is most clearly visible in digital ar- chives, specifically those that are created and evolved on the Web. In 2013 Olga Goriunova reflected upon the role of the online curator as a tension between heavy and light curating: Curators and museums working with the new computational mate- riality are compelled to remain lightly operational, responding to the creation of new aesthetic value, whether by artists or those be- yond-artists. Confronted lightly and omnipresently with the new aesthetic values ceaselessly churned out by the operations of computational matter, the curator’s or art institution’s work is heavy. (2013: 28) In other words, a curator involved in the computational intricacies of the Web sits between the lightness of applications and interfaces and their aesthetic values, and the ‘heavy’ precision of human–technical creation and systems involved in the production of cultural works. She continues: Today, the emergence and further unfolding of aesthetic value is open to intervention and meaning-making in a manner that is un- foreseen. Parties of different kinds and orders partake in producing, elbowing each other and dipping into such openness. Here, artists and curators are actors amongst many others, whose aims may be far less generous or conducive to future imaginative openings and the excitement of living. (2013: 28) This means that curating on the Web is not merely concerned with ob- jects, but that curation functions within a wider ecology of social and technical relations. Applied to online archives this means that archiving on the Web also needs to take into account a complex interrelated net- work of dependencies and contexts that are often invisible, but which are of crucial importance to the making and archiving of culture. In such a scenario, and especially when taking into account algorithmic processes, archiving is becoming darker as less and less is known about the condi- tions and outcomes of the computer processes. This process is well exem- plified in the project Dark Archives (2015) by Erica Scourti. In general, the term ‘dark archive’ is used to indicate a repository for 135 Dekker information that can be used as a failsafe during disaster recovery – it is a copy of an archive that consists only of meta-data and is not for public use. However, Scourti is interested in another type of dark archive: the information in an archive that cannot be seen. For example, Amazon.com could be seen as a very ‘light’ archive. Their business model is based on retrievability, which means that everything can be found and is accounted for. Amazon.com has to battle against the forces of darkness, which threaten to make things in the archive un-findable: the photos and videos that somehow evaded classification, the false negatives, the misclassifica- tions, or the media that fell outside Google’s definition for that search term. It could also be spam or things with very similar titles, which hap- pens more and more with algorithmically produced content. So, there is need to keep things retrievable otherwise the content of the archive can fall into darkness: the items are available but you cannot find —or sell— them anymore. With Dark Archives Scourti wanted to examine how visi- bility and invisibility relate to online archival platforms such as Google Photo to explore the idea of what eludes classification in an era of increas- ingly intelligent auto-classification systems and in what ways these sys- tems affect data of individual users. Her project Dark Archives consists of two phases; in the first, Scourti uploaded her entire fifteen-year personal media archive of daily (digit- ised) photos and images, videos and drawings to Magisto, an online auto- editing and archiving app. Magisto’s algorithms analyse videos and pho- tos that are uploaded and breaks them down on three levels: visual analy- sis, audio analysis and storytelling. Essentially Magisto searches for simi- larity in the images and sounds. For example, most people take several photos of the same moment or object, because the light was not right, or the smile not genuine enough. The algorithms trace all these instances, which often end up somewhere in the dark corners of hard drives, and based on the results edits everything together in a short video. Scourti’s videos showed some of the footage but not all. For the second phase, and inspired by the results of Magisto, Scourti turned to Google Photo. Next to archiving photos, Google Photo uses ‘Assistant’, a similar algorithm to Magisto that generates short animations or panoramas by stitching simi- lar images together. Of course none of the users really know how these algorithms search for and categorise the images, what the exact parame- ters are or what the algorithm looks for. This ambiguity, or randomness, of tracing, retrieving and creating new content was of interest to Scourti, and in an attempt to address the ‘missing media’ she asked five authors – Between Light and Dark Archiving 136 Jessica Bunch, Christina Chalmers, Sandra Huber, Linette Voller and Jo- anna Walsh – to search her Google Photo page with keywords of their choice and then speculate on and write captions for what they imagined to be the missing set of media for that search term. By asking the writers to imagine the way an algorithm works, the project tried to get at the core of what a non-human thought process or logic could be. Scourti then matched their captions with the existing media from her archive, and cre- ated a final series of videos. Scourti’s Dark Archives shows how online content is always ambigu- ous and unstable—at least when using automated editing systems or even certain platforms. Content moves between ‘light’ and ‘dark’ archival sys- tems, surfacing occasionally, while at other times it may present itself in a completely changed context. And even though the content might stay the same, it is not static; it is always added to, and depending on the search terms, the context changes. It could be argued that any collection, or ar- chive, has potentially limitless constellations within it and the meaning of it changes depending on the context in which it is used. However, with semi-automated algorithms an additional, ambiguous, unfinished and semi-fictional quality enters into the equation that is unknown, unpre- dictable and invisible. Dark Archives points to the issues that emerge when image archives can be parsed, and potentially monetised, once in the hands of corporate platforms. It also explores how new technologies inscribe knowledge in different ways and how they record and archive the lives of their users. In other words, how identity and memory are con- structed. Perhaps every document creates (rather than describes or illus- trates) the event; every search creates an archive, and every archive gives rise to a different reality. Search queries both create an archive and are potentially archival material in themselves (as the still on- going fascination with Google’s auto-complete attests to) and, as Derrida says, the archiving itself is productive of events, historical and otherwise. (Scourti, in Dekker 2016) How to re-assert agency and control when creating online performative or dynamic archives? One possibility I explored in several online projects was to see how such spaces are used as both sources for information and provide possibilities for new enactments, projects or performances. 137 Dekker Circulation In the project One Terabyte of Kilobyte Age (2011–ongoing) Olia Lialina and Dragan Espenschied developed new archival methods that reflect the way archival content was created: the captured universe of Geocities. Ge- ocities was a free Web hosting service founded in July 1995. It soon emerged as one of the most popular and inhabited places on the Web and remained so until the late 1990s. At the peak of the dot.com fever in Jan- uary 1999 Yahoo! purchased Geocities—for 3 billion dollars. However, Geocities soon became synonymous with old-fashioned aesthetics and bad taste. At the same time people drifted to social network profiles. In April 2009, Yahoo! announced that it would shut down Geocities in six months. During these months the Archive Team and about 100 people managed to rescue almost a terabyte of Geocities pages. And on 26 October 2010, marking the first anniversary of Geocities’ closing, the Archive Team re- leased a torrent file archive of 641 GB, containing approximately 1.2 mil- lion accounts. As mentioned by digital archivist Jason Scott: Geocities arrived in roughly 1995, and was, for hundreds of thou- sands of people, their first experience with the idea of a webpage, of a full-colour, completely controlled presentation on anything they wanted. For some people, their potential audience was greater for them than for anyone in the entire history of their genetic line. It was, to these people, breathtaking.1 As a symbol of the ‘amateur’ Web, Geocities is a trace of how the Web was used at the time. This was one of the main reasons why on 1 November 2010 Lialina and Espenschied bought a 2-TB disk and started download- ing the largest bit torrent file of all time.2 They started unzipping the first files in January 2011, a process that ended in March 2011. After down- loading, storing and sorting the 16,000 archived Geocities sites, which took another year, they started to redistribute screen captures of the Geo- cities homepages through the Web. As Espenschied remarks: ‘Content’ that is isolated, de-contextualized and shuffled around in databases of social networking sites is the main form of communica- tion, to be useful an artefact has to work as a ‘post’, it has to become impartible and be brought into a format that is accepted every- where. And that is a screenshot. (Owens 2014) 1 See http://ascii.textfiles.com/archives/2720. 2 For more information about their research and findings, see Lialina (2017). Between Light and Dark Archiving 138 The circulation was done in different ways: they opened an automated Tumblr blog that every twenty minutes uploads a new screen shot of a Geocities homepage; the screen captures are liked and reposted by the Tumblr followers; and the most reposted or liked are then presented next to related research on their blog One Terabyte of Kilobyte Age while at the same time distributed through Twitter.3 The Geocities archive became a spiral in which Lialina and Espen- schied reflect on the Tumblr archive of the torrent archive of the Geocities archive, people reblog, retweet, like and save the posts, and it just keeps going on.4 While Geocities was almost a forgotten world on the Web, due to several enthusiasts and thousands of followers and users it became not only visible but also an important marker in the Web’s history, and through liking, sharing and redistribution Geocities keeps circulating and popping up in new contexts.5 Next, the project introduced a whole new folksonomy through tagging—for example, ‘alive’ and ‘under construc- tion’—of how this new archival material could be categorised and analysed (Lialina 2017). Lialina and Espenschied’s project provides all kinds of in- formation on how Geocities was used and misused, in terms of frames, banners, navigation elements, GIF’s etc. One Terabyte of Kilobyte Age provides a means of archiving over 500,000 screenshots of homepages, and (re-)viewing the home pages through contemporary interfaces says much about the humour that drives online culture—at least in those days. Instead of purely collecting the ma- terial for the purpose of preservation the project became about question- ing what ‘archive’ could mean in the context of making work accessible. The artists specifically choose to represent Geocities’ history as a dynamic and still-evolving project, rather than have it exist as static ‘back log’ of data. As described by Espenschied this form was explicitly chosen as a cu- ratorial and conservation method, because: 3 See for the Tumblr page http://oneterabyteofkilobyteage.tumblr.com/, the blog http://blog.geocities.institute/, and the Twitter page https://twitter.com/geocities_txt. 4 Interestingly in 2013 Yahoo! which previously shut down Geocities (and also quickly brought down the popular sharing site Delicious after they bought it) bought Tumblr. With such a track record, it’s not unlikely that One Terabyte of Kilobyte Age will outlive its cur- rent host. 5 Often circulation is described as a process of distribution and liking (see, for example, Steyerl 2013). Although this is certainly part of the process, circulation is also inherent in the system through hyperlinking, sharing, and execution processes that move beyond per- formativity. For more information about this type of computational circulation see, for example, Mackenzie (2005). 139 Dekker Digital Culture is Mass Culture, it is also more about practices than objects. In order for artefacts to survive culturally, they need to be- come useful again in contemporary digital culture. (Espenschied in Owens 2014) Interesting questions arise concerning traditional concepts such as prove- nance and authenticity. As Espenschied also acknowledges, the screen- shots have ‘authenticity issues’, but he goes on to say, ‘this is greatly out- weighed by their accessibility and therefore impact’ (Owens 2014). The other way to experience the online archive would be to emulate the Netscape browser, but this would be costly and require complex emulator setups. While accepting the losses, applied in a generative and circulatory way, One Terabyte of Kilobyte Age keeps creating new forms and inter- pretations, at times pushed forward by humans, other times by bots (Twitter is supposed to have millions of bot accounts, comprising at least 15% of all their accounts). In this way, the digital archive becomes a care- fully designed mass re-enactment. The light interface allows for easy re- circulation, bringing to the surface more forgotten moments and new ex- periences. The project is driven by a desire to use technology as a tool to make vis- ible and open up content or conduct that is neglected, forgotten, discard- ed or deliberately concealed. In this way, it makes sense – as also pointed out by Bethany Nowviskie, director of the international Digital Library Federation at CLIR—to ‘take the notion of cultural heritage not as content to be received but also as technology to be used’. This means that artefacts and events are no longer merely about the past, but are tools that can be used to imagine alternative pasts and futures (Nowviskie 2016). Similar attempts can also be seen in other areas and disciplines, for example, in the conservation of public art. If, as proposed by conservator Glenn Whar- ton, it is done in collaboration with local people such ‘participatory con- servation has the potential for critical dialogue on how material culture is preserved, presented and used (Wharton 2005).6 Or in the methods used by Shiobhan Davies’ dance company for their archival project RePlay. Launched in 2009 the elaborate site includes a comprehensive range of audiovisual content, text-based materials and other content that had not been seen before in the public domain, and offers users digital representa- tions of the dance-making process and tools to collect and organise 6 See also the concept of ‘proliferative preservation’ that is suggested by Rinehart and Ippolito (2014) – a method that allows copies of a work to circulate and mutate to better ensure their survival. Between Light and Dark Archiving 140 searches for research, teaching and general interest. The archival site started in 2005 as a simple online portal; however, it turned into an ex- tension of the artist’s work, reflecting back on her creative methods and also influenced her—as well as other peoples—future projects. Rather than purely collecting the material for the purpose of preservation, RePlay queries what ‘archive’ could mean in the context of making work accessi- ble. Similar to Lialina and Espenschied, they were interested in—whether the circulatory characteristics of a digital archive would change how a work was read and in what ways it would provide new meanings of work that were otherwise lost or at least less accessible. Rather than existing as a static ‘back log’ of material, RePlay became a dynamic and still growing site that changes over time: […] it involves artists in ‘dancing the archive’ by exploring archives as source material for new performances and novel methods of art- ists documentation, rather than for faithful re-enactment or recon- struction. (Whatley 2017) These efforts emphasise the critical context of something that is transi- ent and often seen as ephemeral. Or as the creators of RePlay put it: ‘the archive reveals not merely explains something: it probes, questions and provokes’ (Whatley 2017). Used in a generative way, the archive keeps creating new forms and interpretations of a work. Networked co-archiving Based on some of the outcomes of her ongoing research into online cura- tion, curator and PhD researcher Gaia Tedone argues that the new tech- niques of many online platforms have shifted the meaning of the term ‘curating’ in which a new ‘networked co-curating’, and I would argue also a ‘networked co-archiving’, is taking place that is formed by a strategic alliance between curator or archivist, users, objects and machine (Tedone 2017). As she points out, this change is characterised by a collision of dif- ferent interests driven by economic, cultural, and socio-political agendas, and can be framed as a new space of performativity, signalling a move from object and project to process and intervention (Tedone 2019). In other words, ‘networked co-archiving’ shifts attention from what is pro- duced to how it is performed and processed. In this scenario, taking into account the algorithmic processes described earlier, ‘networked co- archiving’ is becoming darker as less and less is known about the condi- 141 Dekker tions and outcomes of computational processes. In an attempt to counter such darkness artist Harm van de Dorpel developed the project Deline- ar.info (2014–ongoing). Delinear.info sits in between a sketchbook, a social platform, and an archive. The content that is uploaded by each user simultaneously func- tions as navigation. Different users take part and connect associatively through images, sounds or texts. Van den Dorpel developed Delinear.info because he believed that many existing social media platforms are too chronologically driven and are not enabling enough freedom for expres- sion, as the systems were too rigid and pre-set. Similar to Davies, Van den Dorpel sees Delinear.info as a studio, a place to try things out. As he em- phasised he is […] particularly interested in how we can connect information in new, meaningful ways. In this context I mean meaningful as aes- thetically surprising. I do not believe that knowledge is embedded in documents, just as that beauty is not embedded in objects. Beauty and knowledge arise through a game between the creators, viewers, contexts, historical stories, et cetera. (Dekker 2015) Van den Dorpel approaches the network as a recursive environment, which is about organisational structures, conceptualisation and making relationships. Emphasising that something is never finished, that results are not fixed and that things can always change, Delinear.info shows how a work of art or an image doesn’t create meaning or value by itself. It is the different relationships that—in this case—determine the identity of an image. This way of thinking could be said to be prevalent in the history of art, but in the last decades it is also gaining more attention and relevance in the field of archiving. However, what is slightly different and pushing the current ‘boundaries’ of archival thinking is Van den Dorpel’s insistence in searching for complexity, which he finds in the layers between different systems that are connected to each other. Whereas most methods for or- ganising information are focused on reducing complexity, he doesn’t try to simplify or reduce, but to encourage complexity. Most advanced search and database engines are still mainly based on text; hence there is a need to reduce complex matters to a brief summary, ideally one word. This way data can be classified and interpreted mechanically. As Van den Dorpel mentions, the interpretation from something complex to abstract qualifi- cation hinders alternative interpretations (Dekker 2015). Based on visuals in Delinear.info the ‘object’ links, it stands on its own and connections— Between Light and Dark Archiving 142 or relations—arise from the operations that are performed on it. As a user you trace information more intuitively and aesthetically to encounter new information through links, instead of following expected routings. Rather than a smooth fabric or relatively ordered story, images, words, sounds and sentences seem to dance around, move chaotically, form chains, and assemble into a ragged and unbounded multiple. Going against structured searches in database or archival systems, De- linear.info shows the performativity of a database and it articulates the medium-specificity in an aesthetic gesture. Such performance resembles a thickening, a saturation of relations that are intensified, doubled, made similar, made serial, and it is from these operations that something else arises. Embracing unpredictability and the indeterminate It could be argued that Delinear.info is far removed from conventional archival practices that are concerned with original records, objects or documents. Indeed, online archiving has many of the characteristics pro- posed by the anarchival. In an environment in which data are constructed, and can only be experienced in relation to other data, code, software and hardware, which may have no obvious relation with the ‘object’, there is a need to consider alternative ways of thinking about archiving. In order to culturally and socially survive, an online archival practice should reflect the environment in which it’s ‘records’ are created, embedded and con- structed. While an anarchival position proposes itself as a counter- example to conventional archiving—stressing the (re)use and production of new work, as a creation event,7 with the term ‘networked co- archiving’—I would like to stress the importance of rethinking the archiv- al, not merely as a counter-strategy, but as a concept to reconsider the in- fluence different systems (human and machinic) exert on archival practic- es, while introducing alternative ways of thinking about what an archive could mean. The examples I described emphasise an undermining of what could be considered an archival monopoly to classify, document, display and pre- serve, and the preoccupation with the archival record. Online archiving is performed in and through human and technical elements, relations and 7 For more information about anarchive in the sense of ‘creation event’, see the detailed explanation on the Senselab site: http://senselab.ca/wp2/immediations/anarchiving/anarchive-concise-definition. 143 Dekker interactions, which are active in and are simultaneously organised through the platforms and systems that are used. The result can be framed as a ‘networked co-archiving’, which is not only concerned with objects, but how different elements function within a wider ecology of cul- tural and socio-technical relations that participate in creating and archiv- ing culture. It is by embracing these relations and moving between light and dark that a new model of thinking, and archiving, emerges that trig- gers new relations and creative associations. This new model is open to diverse interpretations, and by following paths that were not yet consid- ered, something new could arise. Such a method attributes new values to computational ‘objects’ and creates new keys by enacting a specific human and machinic aesthetic relationality. References Dekker, Annet. 2016. “Archiving Our (Dark) Lives. Interview with Erica Scourti.” Het Nieuwe Instituut, January. http://archiefinterpretaties.hetnieuweinstituut.nl/en/archiving-our- dark-lives-interview-erica-scourti. Dekker, Annet. 2015. “Choosing Complexity: Harm van den Dorpel.” In Metropolis M, n. 6 December/January. Ernst, Wolfgang. 2015 [2002]. Stirrings in the Archive. Order from Dis- order. Translated by Adam Siegel. London: Rowman & Littlefield. Foster, Hal. 2004. “An Archival Impulse.” In OCTOBER, Vol. 110, Fall 2004, pp. 3–22. Goriunova, Olga. 2013. “Light Heavy Weight Curating.” In Speculative Scenarios, or What Will Happen to Digital Art in the (Near) Future, edited by Annet Dekker. Eindhoven: Baltan Laboratories, pp. 25-31. Laermans, Rudi and Pascal Gielen. 2007. “The archive of the digital an- archive.” In Image [&] Narrative. Online Magazine of the Visual Nar- rative. Issue 17, April. www.imageandnarrative.be/inarchive/digital_archive/laermans_giele n.htm. Lialina, Olia. 2017. “Still There. Ruins and Templates of Geocities.” In Annet Dekker (ed.) Lost and Living (in) Archives. Collectively Shaping New Memories, Amsterdam: Valiz, pp. 193-210. Between Light and Dark Archiving 144 Mackenzie, Adrian. 2005. “The Performativity of Code: Software and Cul- tures of Circulation.” In Theory, Culture & Society. Vol. 22/1, pp. 71- 92. Nowviskie, Bethany. 2016. “Speculative Collections.” 27 Oct 2016. Web- site http://nowviskie.org/2016/speculative-collections. Owens, Trevor. 2014. “Digital Culture is Mass Culture”: An interview with Digital Conservator Dragan Espenschied.” 24 March. Website https://blogs.loc.gov/thesignal/2014/03/digital-culture-is-mass- culture-an-interview-with-digital-conservator-dragan-espenschied. Rinehart, Richard and Jon Ippolito. 2014. Re-Collection: Art, New Media, and Social Memory. Cambridge, MA: MIT Press. Steyerl, Hito. 2013. “Too Much World: Is the Internet Dead?” e-flux Jour- nal #49, 11. http://www.e-flux.com/journal/too-much-world-is-the- internet-dead. Tedone, Gaia. 2019. “Online Curating – Interview.” In Annet Dekker (ed.) What We Talk About When Talking About Online Curating, Brescia: Link Editions (forthcoming). Tedone, Gaia. 2017. “Tracing Networked Images: an Emerging Method for Online Curation.” In Journal of Media Practice, Vol. 18/1, pp. 51-62. Wharton, Glenn. 2005. “Indigenous Claims and Heritage Conservation: An Opportunity for Critical Dialog.” In Journal of Public Archaeology 4, pp. 199-204. Whatley, Sarah. 2017. “Archiving the Dance: Making Siobhan Davies Re- play.” In Maaike Bleeker (ed.) Transmission in Motion: The Technolo- gizing of Dance, London and New York: Routledge, pp. 62-9. Zielinski, Siegfried. 2014. “AnArchive.” In Eckhard Fürlus, Claudia Gian- netti (eds.) AnArchive(s) - A Minimal Encyclopaedia on Archaeology of the Arts and Media. Idea by Siegfried Zielinski, Köln / Oldenburg: Verlag der Buchhandlung Walther König / Edith-Russ-Haus für Me- dienkunst.