The internet will not remember you: curation of autobiographical online materials in Russia in Spring 2022

Abstract This paper is based on research conducted in February–April 2022. It describes and illuminates what was happening with tech-savvy educated people between 20 and 40 years old in Russia, while their usual digital tools and places for the autobiographical process were changing in the spring of 2022. Facing censorship of platforms, surveillance, and the inability to pay for services, people who were keeping important memories of their lives online were deleting their profiles, migrating to other platforms, censoring themselves, and creating archives of autobiographically meaningful materials. The paper examines these disruptions as a case that illuminates the role of online platforms in autobiographical memory and expands some concepts within autobiographical memory studies, such as evocative objects and autotopography.


Introduction
'Instagram for me is an album of memories' was written in a comment published in the Russian online news outlet Meduza on March 13, 2022.That month Instagram was blocked in the Russian Federation.A month before this publication, on February 24th, Russian army forces invaded the territory of Ukraine.The invasion was accompanied by a variety of repressive actions, starting from arrests of people who publicly criticised the government, and ending with censorship of online content as an attempt to control the representation of the invasion.At the same time, companies from other countries started leaving the territory of the Russian Federation, limiting access to online services and payment methods.
Forced changes in the Russian online environments in this paper are not approached through an ethical or political perspective but are viewed as a case of memory practices at risk, which highlights the role of platforms in the ways people curate autobiographical memory.Based on the works of Brockmeier (2015), I understand autobiographical memory as a process of constant creative interpretation of experience with the usage of cultural means.As other scholars show (e.g., Dijck 2007;Heersmink 2018), this autobiographical process requires a regular remembering with the help of materials that connect a person to their past (photos, diaries, etc.).The internet plays a significant role in this process, as it provides spaces and tools to save and take care of materials of all sorts.Therefore, when it fails to provide a safe and stable environment, what happens?
This paper describes and analyses changes that were happening to the curation of autobiographically meaningful materials, for people who could not use the internet in the same way as before the start of the main censorship acts.The main question of this research was: How have people in Russia curated their autobiographically meaningful online materials during the media censorship in February-March 2022, and how does this case illuminate the role of online platforms in autobiographical memory?With archival research and eight in-depth semi-structured interviews, I investigated changes while they were happening.
The results of the paper could be used to deal with these challenges in the future, both on the personal level, on the level of platform design, and in policies (how to protect autobiographical data in case of challenging events).While the production of guidelines is beyond the scope of this small research project, my analysis of disruptions illuminates both the challenges and possibilities that people may face.
This paper also expands on vocabulary to describe the role of online environments in personal memory curation.I suggest differentiating 'potential' from 'curated' evocative objects: autobiographical materials that are part of an ongoing flow of self-description online and are not meant as memory objects, and curated archives that are purposely organised to return to and remember with.I also explore the notion of autotopography, briefly suggested by other authors, and show how it can work in application to the internet.

Evocative objects and autotopographies
Remembering is shaped and mediated by material objects.Turkle (2007) applies the notion of evocative object to describe this role.She understands evocation in a broad sense, drawing attention to the ability of things to evoke feelings and thoughts.Heersmink (2018) further argues that autobiographical remembering is embodied and distributed (shaped by objects we interact with).Sometimes to remember, a person must hold onto objects (e.g. a pencil); their shape, texture, and visual appearance can navigate thoughts and actions.
Materialistic approaches to online environments can be used to view different digital entities as objects (Boomen 2014;Casemajor 2015), thus, allowing us to approach internetbased memories as evocative objects too.To avoid unnecessary distinction between analogue and digital autobiographical materials, José van Dijck coins a term 'mediated memories'.José van Dijck emphasises processual, creative, interpretative remembering, and claims that autobiographical memory is always mediated, while media and memories mutually shape each other.It would be wrong to say that either human brains/bodies or objects store memories, or that some 'raw' memories get distributed to the artefacts that, in turn, mediate them.Instead, as van Dijck argues, people adapt material things to their remembrance needs and adapt ways of remembering to existing tools and formats.The way memory is enacted and mediated when a person opens a photo album, can also happen with online albums or blogs that were carefully styled and organised by their users.Digital pictures, texts, emojis, and even parts of the interface can evoke stories of the meaningful past.
Memories online are not only mediated but also social and interconnected, networked (Hoskins 2014).In Beyond the Archive Jens Brockmeier writes that 'social media make autobiographical process more public and shared, blurring the line between personal and collective memories, as they are "constitutively networked"' (Brockmeier 2015, p. 39).
Observing mementoes in households, Petrelli and Whittaker (2010) noticed that digital memories are less integrated into active processes of remembering, they do not become a matter of daily routines or conversations, 'they cannot express symbolic meanings through new uses, instead they are constrained to simple representations of events and people' (p.166).
As I show in this paper, there is a difference between networked memories and locally stored memories, which makes the disposition between 'analogue/digital' an oversimplification.Autobiographical materials stored online can also easily be a subject to creative re-appropriation and collective remembering.The issues with integration into remembering practices arise when people try to take online mementoes out of social networks.These two factors create challenges like described by Petrelli and Whittaker (2010).
In the same research, authors distinguish between more and less accessible evocative objects (what Heersmink (Ibid.) also calls active and passive).Active evocative objects are placed in plain sight and are easily available for interactions.In turn, passive objects are stored in boxes and not displayed.With respect to this observation in offline settings, I find it unfitting to the realms of the internet.Compared to households, online environments are networked and full of multimedia artefacts that are publicly observed and used.Still, I suppose that similar distinctions might be applicable to autobiographical online materials.Later in this paper, I introduce a distinction between potential and curated online evocative objects that helps to capture the tension between self-writing as a performative set of practices, and as a curation of the self's past for future reminiscence.
The network of these material things (Heersmink 2018), following the work of Gonzalez (1995), addresses as autotopography.Petrelli et al. (2008) used the notion of autotopography to describe networks of mementoes at people's houses: what objects are purposively saved to remind of the past, where are they placed, how are they stored, how are they differentiated.Quoting Gonzales, they define autotopography as an arrangement of mementoes that constitute 'a physical map of memory, history and belief'.Practically that means 'studying how people choose significant memory objects and how they arrange and use those objects in their living space' (p.53).
Autotopography did not yet become a common term within autobiographical memory studies, or within research on digital culture.Nevertheless, I find it useful to situate the autobiographical process in a material world without losing focus on objects.The notion of autotopography can help to look at how people structure different evocative objects, organise, or classify them.During online media censorship, the usual autotopographies were changing, which allowed to structure and describe autotopography characteristics that became visible through this transformation.

Interlocutors
This research was based on eight online interviews with people who stored their autobiographically meaningful materials on Facebook, Instagram, VKontakte, or other platforms and made adjustments during February-March 2022.All my interlocutors were between the ages of 20 and 40, digitally literate.All people at the moment of interviewing were criticising the actions of the Russian government, the invasion, and the related political repressions.They were well informed about possibilities to avoid censorship consequences, and I expected them to show a wider and more professional or creative range of actions.
Transcripts were anonymised; personal details that could identify the interviewee were deleted, and each person was assigned a pseudonym and neutral pronouns 'they/ them'.Gender was not defined as a useful category in this research, as it has never been mentioned as one by the theoretical frameworks and research literature, nor did it show up in the conversations as a topic.Thus, the pseudonyms of the characters are not based on their gender.For privacy protection purposes, they were chosen without any implied connection to appearance, personality or other biographical facts that could identify these people.This is the list of people I talked to: Albertine was a cultural scientist.About nine years prior to this research, Albertine had an accident and lost their memory.Their earlier Instagram posts and VK photo albums were a way for them to access that past.
de Norpois was a social and internet researcher, a professional writer, and author of several biographical online blogs.
Basin was a developer and UX researcher, interested in different formats of digital diaries and lifelogging, and experimented with a multitude of ways of storing and accumulating autobiographical meaningful information online.
Swann was a researcher in a technology-related museum, a sound designer, and a fiction writer.They have been dealing with topics of web archiving in their professional activities.They have been using VK for blogging and communication since it appeared.
Robert was a researcher, who launched one of the organisations for internet studies in Russia.They used VK and Instagram for personal blogging and communication.
Leonor was a philologist, translator, poet, and political activist.One of their close relatives is an internet pioneer.They used Instagram and VKontakte for personal blogging, communication with people and had a small VK blog for his poetry.
Andrée was an archiving professional.At the moment of the interview, they have been working with web archives of digital-born heritage.They used VK to post about life and connect to people.
Léa was a cultural studies student, writer, and political activist.By the start of the invasion and following changes in the online environment, they had been under home arrest, with legally restricted access to the internet by which they did not abide.Léa used Instagram and other SNS for extensive autopoetical writing and activism.

Methods of collecting and analysing materials
The main way of gathering empirical data was through semi-structured online interviews.This method allowed to get access to sophisticated archiving strategies, meanings, the dynamic of the archiving process, peculiar challenges and problems of archiving, and biographical contexts from which remembering practices emerged.All my interlocutors tried to acknowledge when they were not sure if they remembered things correctly.This reflection helped to identify limitations and uncertainties in further analysis of interview materials.Questions were touching upon three topics: the role of social media in selfremembering, interlocutors' interest to archiving, its history and practices, and transformations that came around the time invasion and after.
Interviews were conducted and recorded via Zoom.Interviews were semi-structured, with variables in a particular order of topics and phrasing of questions.For the interview transcripts, thematic analysis was used (Lapadah 2010).Categories were based on inductive coding.Additionally, small archival research of articles published in Novaya Gazeta between January 1st, 2018 and March 28th, 2022, was conducted.This period was chosen based on information from interviews; the year 2018 was often mentioned as important for interlocutors' memory curation.Archives added broader political and cultural context to the experiences my interlocutors were referring to.

Changes in online memory curation in spring 2022
On February 24, 2022, The Russian Federation invaded Ukraine, which made millions of people leave their homes in search of safety, left people mourning their relatives or bombed houses.In turn, in Russia was a rise of censorship, surveillance, and transformation of political alliances.The internet was going through changes too.
Bans of platforms was the most obvious change.The changes were implemented by The Federal Service for Supervision of Communications, Information Technology and Mass Media (Roskomnadzor).During February and March, access to Facebook, Twitter, and Instagram was blocked (Glukhova and Efimova 2022;Karev 2022;Zepeleva 2022).This was a big change for the Russian media ecology.In March 2022, Facebook and Instagram were among the biggest social media in the Russian Federation; in October 2021 Instagram had more than 38 million active authors, Facebook had nearly 3 million.In comparison, another popular Russian social network VKontakte had more than 23 million, Odnoklassniki had more than 5 million, and Twitter had less than 1 million active Russian authors (Chjornyj 2021).This made both networks the core of the Russian online landscape.Presumably acknowledging this, Roskomnadzor gave users 48 h notice before the changes applied to Instagram were implemented.
My interviews showed that while people were insecure about the future, they did not always struggle with a lack of access.This censorship was easily ignored with the usage of VPNs (Virtual Private Networks).This is how Andrée, an archival activist, was describing it: As by the time of the interview, there were no legal acts that punished individuals for private usage of these platforms, people like Andrée did not feel the risk of 'breaking the law', and continued with their practices, protecting themselves with back-up archives.
Fear of surveillance was the second change, mainly linked to the biggest Russian platform VKontakte (VK).VK was created in 2006, with its name meaning 'Staying in touch'.By February 2022, platform allowed users to create a profile, add friends, write private messages and public posts, add audio and video from a huge user-generated library, create photo albums, and launch groups and pages for collectives.Platform had settings to adjust privacy.This variety of features created rich cultures of memory practices for groups and individuals on the platform.In 2014 VK became mostly owned by Gazprom, the majority state-owned energy company (Kolobakina 2022).At the same time, new policies were implemented, stating that VKontakte is obliged to collect, store, and make available to Russian law enforcement authorities, data about users' personal correspondence, voice messages, etc. (Ibidem.)Slowly VK became known as an unsafe SNS for people to criticise the government.As one of the commentators in Novaya Gazeta newspaper phrased it, 'for me, and for many users, it is simply an archive of digital life from the late nineties to the mid-tens' (Ibidem.).
No sanctions have targeted this network.Nevertheless, it was no longer an option for some people to continue using VK due to governmental surveillance and political propaganda.For example, independent Russian media DOXA agitated people to delete their accounts (DOXA 2022).
A particularly important year in the history of VK for my interlocutors was the year 2018.It was often mentioned as a period when they started worrying about their online profiles.This is Andrée's experience again: Archiving for VKontakte became relevant in 2018, when a lot of criminal cases started appearing … I didn't dare to delete the account, but I did clean my feed, news feed, posts, likes … I didn't do anything illegal on VKontakte, but because of the [legal] cases, it seemed that they [police] could be provoked by even the smallest things.
For Andrée, the motivation for archiving was a rise of politically motivated legal cases in 2018.Seven people from different Russian regions were taken to court because of meme pictures saved or shared on VK years ago.Choice of accused people seemed to be random, while the punishment was disproportionate; some people faced six years of jail (Vasil'chuk and Dokshin 2018;Vasil'chuk 2018b).In September 2018 it became clear that VK was already accessing private information of users by police requests (Satanovsky 2018).In this regard, March 2022 did not bring radically new anxieties, but new factors that escalated these fears.In Spring 2022, people considered once again whether they could keep their archives in VK.
Finally, the third factor that influenced memory practices was the lack of access to paid online services that people used to store autobiographically meaningful materials.For example, de Norpoi had a paid subscription to Google Drive that provided a big cloud storage.After Mastercard and Visa payment systems stopped providing services to Russian banks (Prokofiev 2022), de Norpoi was not able to pay for subscriptions and had to quickly minimise their cloud archives.
Reactions to these changes were diverse.Based on the interviews, I distinguished four overlapping reactions: deletion of profiles, slow migration from platforms, self-censorship, and back-up archiving.
Deletion of profiles in this social group was a political decision.People who openly criticised the government started feeling less safe with what they discussed and posted online.This, for example, happened to Leonor: When the war started, I had to think about each of my social media accounts, public or private, what I wanted to keep or remove … There was a certain technical purpose in deleting VK: to remove my data from Russian servers in order to prolong their work in finding compromising evidence, but I do not see that [risk] in [my usage of] the rest of the social networks.
So, the decision to delete VK was based on a desire to make police work more difficult in case they will be investigating Leonor's actions.At the same time, VK was very important to Leonor as a personal archive of poetry and correspondence, so they tried to archive what was possible.In this context, Leonor was not thinking solely about what they wanted to save from digital networks for the sake of memory, their decisions were based on a reassessment of the digital traces as an evidence for others.The role of social networks as mediators of memory shows its dubious and even harmful side here.
Slow migration from platforms was also common, but a less radical reaction to surveillance, presumably conducted by people who did not feel at imminent risk but did not feel safe either.For Albertine after 2018, VK slowly stopped being the main social network.VK remained Albertine's main source of news about local musical events and a tool to store, as they put it, 'artefacts of past years that I want to save somewhere else, but don't know where'.For everything else Albertine used other platforms.
Self-censorship was another reaction to changes, and it was two-folded.On the one hand, people were deleting already published information that could draw the attention of the authorities.For example, Albertine stored most of their photo albums in VK, but they have deleted an album with pictures from a pre-invasion trip to Ukraine.Many of the other albums were made private; only Albertine could see them.
Basin had to think more cautiously about things they post.In 2014, they escaped army forces that were supposed to be sent to Ukraine.From back then, Basin has a photograph in which they showed a middle finger to one of the Russian army war machines.After March 24th, Basin wanted to publish a post in VKontakte with this picture, but hesitated: I was scared … it's probably not super adequate, but I felt like they could take me [to the court] for this, send me to jail, just for the picture and the music [attached to the post].I thought, if I am sh**ting my pants about posting a picture, that's beyond any limits.That was the moment when I realized there was no freedom of speech on this social network.
On the other hand, people self-censored on social networks where they used to post about lifestyle, daily routines, and happy moments, as it felt inappropriate.This might be related to the idea of giving more space to discussions about the situation in Ukraine by reducing your own online presence.Personal involvement in the Russian-language internet sphere allowed me to witness those discussions.However, often such statements were not preserved, as they were posted in Instagram stories.They were not originally considered as part of the research.Traces of such discussions can be found in an online blog article about psychological aspects of using social media, which names examples of texts that ask people to not post anything unrelated to the invasion (Afanas'eva 2022).Or it could be just a personal lack of need and desire to post.Pauses in posting due to selfcensorship are nothing new.If your relative died, your friend has become a political prisoner, your home city is going through emotionally intense elections; such events can all prevent people from posting lifestyle posts and selfies for some time.But my interlocutors were unsure if the usage of online platforms would ever feel "normal" again at all.Already published memories suddenly became more of an archive and less of a part of an ongoing self-performance as the usage of social networks in self-writing has been disrupted.Invasion's consequences, both within and outside of the aggressor's country, might be lasting for decades.The question that arises is, what cultural formats and in what public/private regimes will be socially and personally justifiable?
Back-up archiving was an overlapping practice that could coexist with all three mentioned above.Archiving was aimed to ensure that information is not dependent on only one platform but can be accessed by other means, e.g., through a hard drive.The hard drive was one of the main solutions.A couple of people said that they have bought hard drives, but didn't use them, mainly because the very process of transferring information was not easy.They were hesitating to explain why it was so, referring to general laziness or lack of time.
My hypothesis is that de-networking of materials changed their role in people's lives, their meanings, and affordances.The next chapter addresses these issues further.

De-platforming and de-networking online memories
The question that has already become commonplace 'What would you grab if your house was on fire?' is also applicable to digital media.Changes that occur to the online environments can easily remind one of the burning houses from which people try to run while taking all their precious possessions.Partially based on this metaphor, I believed that the choices that people made, would highlight a certain perspective on what online memories are more valuable.But it was not that straightforward; my interlocutors were confused about the changes that happened to their materials while they tried to save data.Challenges arise when one wants to transfer remembering practices beyond the platform: either while trying to save certain media formats, or when downloading the whole profile pages.

Saving media formats
Not all data that was registered on platforms works as autobiographical data that is used and cared for by the authors.For some people only texts are important; for othersonly images or contact lists.This was also a clear archiving strategy among my interlocutors; people relied on media formats that were dearer to them.For example, Albertine attributed the greatest value to pictures, especially digitised film photographs that they made: I have been photographing on film since 2011, and for me such photographs are special objects of memory, I have a special emotional attitude to them … But film doesn't last forever, it deteriorates, so I digitize (I don't print) … computers are not a very reliable thing (one of mine broke down), so everything stayed in VK and FB.And at some point I decided that these social networks would be my personal archives.
Swann, a sound designer and musician, attached a lot of evocative value to audio from VK, as it reminded them of previous online collaborations or particular states of mind that they captured in music: First, I downloaded all the [audio] files from the private messages.This is a very emotional process; music is an emotional thing.When you replay the music, you are immersed in what was happening back then.It even depressed me a little, I felt depressed, because there are all sorts of sad experiences and memories.
Prior to infrastructural changes, these people already had a particular preferred way of capturing and remembering their lives.These media were closely linked to professional activities or hobbies, but also served as powerful evocative objects.This corresponds to the observations made by Petrelli and Whittaker (2010): the organization of digital mementoes is done around specific media: photos on external hard drives, videos on DVDs … This organization contrasts with that seen with physical objects which often follow the logic of time (objects of the same era), or topic (objects about the same event or person) tend to be co-located (p.164).
Nevertheless, it would be wrong to treat social networks as just containers in which people 'put' media, and then, if needed, can 'take them out'.Platforms combine a diversity of possibilities: both to save and to share, to be and to act, to publicly create a masterpiece, or serve as a medium for a stream of consciousness.
Spontaneous digital traces of actions also contribute to interface-mediated autotopographies.That means that when a text is extracted from a social network, it loses some of its qualities.For example, Leonor, who has been publishing poetry on VK, noticed that the interface added extra meanings that are lost after extracting texts for their personal archive, There is a certain period of poetry work over 3-4 years, preserved in VK.And this group page, owned and administered by me, is not preserved in the archive.As it's not part of my account [, it can't be saved automatically].Of course, all my lyrics are also stored locally, but those are two different things.Just because I have them on my hard drive doesn't mean that nothing is lost when I lose access to that group … I have a physical memory of which lyrics provoked reactions from my friends and which did not, and which I stopped to like because of that, and which I sometimes return to.
What Leonor misses is the networked nature of online materials.Writing on the networked memories, van Dijck again uses the metaphor of a box full of printed photographs, saying that 'we can no longer keep the lid on the shoebox we used to store in our attic: its pictorial contents will increasingly spill out into the virtual corners of the World Wide Web, where it seamlessly blends in with our collective pictorial heritage.Once again, pictures of life will become living pictureseven if unwittingly ' (2007, p. 121).In my study, the process is reversed; what once was meant to be a 'living picture', becomes isolated and stabilised.
Extraction of media formats means de-networking them, losing their status of memory sites, shared with other people, and media format's ability to work as communicative tools on their own, just by the fact of existing in a public online space and waiting for a reaction.De-networking of digital personal memories means that they are only owned and perceived by one person (or a small circle).
Another thing that is lost when people extract media formats from an online environment is interface.The notion of 'physical memory' that Leonor mentioned in the quote above shows that the interface itself can play a role of an evocative object.The poem that is posted in VK is a multimodal object that includes likes, comments, and dates.Later, Leonor mentions that dates that were automatically added by the platform under each post also contributed to the posted poetry, I really like the practice when a poet writes dates under the poem not according to when the poem was written, but just as an important date that was meaningful to them and the addressee/s, or a circle of friends.And obviously, the date is also a full-fledged part of the poem's text … In VK there was the date of the post [added automatically], and that's also an important piece of information: when did I publish this?When I look at the date of the post, I instantly connect it to the particular events during my school year, a particular season or a state of mind.It's an important piece of information that gets lost too [after manual archiving].
Extracting media from the platforms means treating the platform as a storage tool, which is problematic as bits and pieces of its interface have actually become a matter of creative and autobiographical processes.We know this also from Web History studies of nostalgia, as 'long-time users mourn the loss of particular programs, practices, services, or interfaces.Favourite sites shut down, loose network connections fade away, daily habits are disrupted' (Driscoll 2020, p. 27).In Leonor's story, the interface is not an evocative object per se, as a representation of the history of media, but it complements a memory artefact (in this case, poetry) to the extent that content without interface features changes its structure and meanings.
High dependency of mementoes on platforms is problematic, as it highlights how platformisation leaves people too little control over memory curation in crisis situations.Once networked, media genres cannot be extracted without losses of meaning, but must be reappropriated or curated in a manner that is sensitive to the transfer of meanings between different environments.

Saving whole profiles: getting lost in data
When my interlocutors were not committed to one media format, they turned to automated archiving of the profiles themselves.Wouldn't it be easier not to extract media, but to save them all together as an HTML or other multimedia format?By the time of this research, Instagram and Facebook offered this option, and for VK, there were grassroots archiving scripts available.One could choose what exactly to archive (e.g., only 'Messages', 'Groups', or 'Posts' from Facebook).
Most of my interlocutors did not know what exactly they needed and did not have any references to compare to, so they achived everything.This made archives 'dirty', containing plenty of undesired and unnecessary information.Information was saved as data, artefacts for a machine eye, and not as a set of materials for human perception.Interlocutors who turned to this solution said that they saved profiles 'just in case', but never opened them or did not understand how to read them.For example, Swann comments on their preservation of music, saying that, … there is a lack of understanding of what to do with these materials: some of these [audio] things can probably be cut, published … But for now, I'm just collecting them all together in one place.
Beyond the lack of vision, people felt frustrated about unusual formats as they did not understand how those could be converted into memory practices.Leonor stressed out this transformation while describing their archived message histories: I registered [in VK] in 2010, and it means that there are 12 years of correspondence.I've looked at one of the most copious message histories, with my partner.[The start of] it was two years ago, for a year and a half we were communicating in VK every day, so there are about 100.000 messages … [when I realise] that this old correspondence with people I was in love with is stored locally, it is messing with my head a bit.… It felt different to paper correspondence, as the second one is art, the process of creation of a certain artefact.And now I have an artefact that exists [on my computer] … [it feels] like the advent of war set back time by about a century.
'Paper correspondence' is an interesting metaphor as it shows the conflict between affordances that constitute mediated memories.Leonor says that paper letters are artefacts and writing them is (to some extent) a conscious creative process.In contrast, digital media are places of more spontaneous and synchronous communication, presentation of Self to others, and of flowsflows of words, memories, performances, and interactions.My interpretation is that such 'messing with one head' that Leonor mentions happens because in some contexts there is a need to curate, shape, and define the format of memories that are derived from platforms, while there is no tradition, well-known references, or practices to do this.
In the previously mentioned research, Petrelli and Whittaker (2010) notice that digital (but not networked) evocative objects are much more isolated and not integrated into social relationships or active remembering.Stories of people I talked to expand on this idea and contrast online materials to digital materials stored locally.Online materials can be very engaged in creative and social remembering, but when they are downloaded, this causes challenges.
Another thing that Petrelli and Whittaker mention is that people get lost in data sorting and filtering, and this was an issue in the observed cases.The idea of 'saving everything' from online profiles did not work well, as full profiles contained too much data, and remembering would require sorting, filtering, and organising in ways that go beyond the platform's interface.All my interlocutors wanted to structure their materials but were slightly frustrated or pushed away by technical difficulties, amounts of data, or a lack of meanings or visions of what they were doing.

Curated and potential evocative objects
In the previous part, I have described different confusions that people faced while trying to preserve their digital mementoes during social media censorship.The main challenges were linked to the fact that many digital traces were not perceived as memory artefacts before, but users were trying to conceptualise them as such and to treat them in ways to which these artefacts were not purposed.This brings us to a need to reassess the notion of 'evocative object'.
Heersmink describes evocative objects as material artefacts with which people evoke memories and thoughts about the past.My interlocutors' craving to download the data illustrates a need to keep evocative objects close and safe when they are in danger.Tackling these artefacts inspires new ideas and inspirations, but not all evocative objects are the same.Some are easier to remember with, and some are expected to serve this role, but suddenly fail.How can we further understand this difference?
While introducing the notion of 'mediated memories', José van Dijck writes, 'Memory work involves the production of objects …with a double purpose: to document and communicate what happened ' (2007, p. 5).While van Dijck's book stresses the overlap between these two purposes, my case study shows that sometimes they should be separated, as different sentiments and expectations are connected to them, and the overlap causes confusion.Let us remember again the story of Leonor and 100 000 digital messages that never were meant to become an artefact.These messages were supposed to serve the purposes of communication in the first place.When they were downloaded from the social network and into the laptop, Leonor was frustrated about what to do with them and how to read them.They were never meant to become similar to paper correspondence that is treasured and kept in a wooden box.
On the other hand, we could remember the story of Albertine and their VK albums that organised photos by topics and time periods, but were only available for the author, access to them was restricted for other people.Albertine organised and curated them carefully, taking places, years, and events as categories for classification.Comparing stories of Albertine and Leonor, we can clearly see that people treat online artefacts differently in relation to practices of remembering.To conceptually grasp this difference, I suggest two notions: potential and curated evocative objects.
Curated evocative objects are singular artefacts, offline or online, digital or analogue, which were purposefully stored by a person to engage with them and evoke memories.Curated evocative objects in the context of memory are easier to imagine, as they have a long tradition.In the context of analogue media, I think about my mother, who keeps a dozen of family photo albums in her apartment.Albums as her evocative objects are well-structured, they provide a particular order of pictures inside of them and often have notes.These are the objects that were designed to remember, their affordances are well known and well navigated by the family, and they are regularly used in remembering practices.Similarly, Basin had a collection of notes, texts, and pictures that they were collecting online to keep different digital traces in one place.They were experimenting with different platforms and aimed to build one online 'home' for all memories that were important to preserve.Basin selected materials that mattered the most and tried to find the best way to store them for future uses.
Another example is curation of an Instagram profile as a personal collection of memories that captures important moments of life, moods, aesthetic preferences, and feeling of Self through the years, something that both de Norpoi and Albertine have been doing for a long time.The evocative power of objects is known to these people, and this makes users treat materialsaccordingly, keeping in mind possible affordances of these collections and taking care of a future Self that might want to read them.
Potential evocative objects, in contrast, were never meant to be mementoes, but were produced along with other practices.Accidental digital traces, parts of an active selfpresentation and communication: message histories, old posts, log of profile picture changes.Their power to evoke memories is spontaneous or sometimes just potential, not actively used.They are autobiographically meaningful, but this is only realised when these objects are put at risk.Then, people hustle to archive, preserve, and copy everything important.
The potential of being evocative makes these materials valuable; but as they were never meant to serve remembering purposes before, people do not understand how to curate them and engage them into remembering practices (beyond spontaneous interactions).This creates frustrations: what to do with these materials, once they are in a 'safer place'?How do they make sense beyond online platforms and social networks?When people download data and create a folder for it, it does not make data an evocative object right away.In order for it to become one, it needs some structuring, filtering, or other organisation that would create clear ways for a person to use it, and, in turn, to practice remembering.While no curation is present, this 'archive' exists on a hard drive like a dusty shoe box with random letters and pictures.It stores something but is not inviting for an interaction.
Jens Brockmeier pictures autobiographical remembering 'as a form and practice of human agency ' (2015, p. 119).The stories about people's attempts to convert profiles and their data into evocative objects lack this agency.The format of preservation is dictated by the platform and does not leave enough room for creativity.The further question here is, to what extent are these issues of a cultural or a technological matter, i.e., to what extent should we criticise the platforms for monopolisation of memories?If so, the solution might include collaborations between memory researchers, users, and developers to work on easier archiving and curating methods.Some of my interlocutors were hoping to find software and delegate this conceptual work to another technology.They themselves experienced a lack of ideas about how to frame de-networked digital artefacts.
While new software can be a solution, I suppose that potential evocative memory objects can also be stuck in this grey zone because of a lack of available curation practices.Alternatively, to address this issue, we could reject technosolutionism and seek solutions in beyond-platform cultural exploration (in art, cultural work, creative coding) of different possible ways and techniques of extracting memories from their sticky digital networks.
This binary opposition is purposefully slightly exaggerated.The solutions might appear on different levels: platform governance, design of affordances, vernacular creativity and communal creation of meanings, practices, and traditions.

Adaptability of online autotopographies
There were other reactions for changes beyond archiving: I.e., slow migration to other platforms and services.In this last chapter, I will discuss some reactions that were focused on an ongoing self-remembering.Then I will address the question of how these stories can help to elaborate on the notion of 'autotopography'.
Limited access to usual platforms was a disruption not only because people could lose all the created artefacts, but also because the tools and places for their regular selfwriting were not working.It was difficult to continue posting on Instagram or writing on Facebook.One example is the story of Andrée, a person who was professionally working with web archives when changes in the social media landscape started.Even when facing technical troubles, Andrée stayed relatively optimistic, expecting that there will be either new technical solutions or cognitive adjustments, I would be sorry to lose the materials I have accumulated or the lists of friends.But there is an understanding that if these social networks are gone, there will be others.There is even some fatigue from these social networks, so I *want* to feel that it would be even better if they were gone.It's a pity, of course, but I will adapt.
Another illustration is the story of Léa.During March 2022, Léa was under home arrest; the Memorial Human Rights Centre officially recognised Léa as a political prisoner.As Léa was temporarily released later, they fled the country.For Léa online self-writing was very important.After being arrested in 2021, police confiscated all their digital devices and Léa, who used a password manager on their laptop, lost access to all social networks.Léa was prohibited from using the internet; but avoided these restrictions in different creative ways (by writing anonymously or from a third person perspective).Commenting on the effect of these disruptions, Léa compared their story to stories of people who lived through a deficit of crucial resources, As I've been losing my accounts … many times, I've developed a strange, semi-panic feeling … I think it can be compared to the way elderly people who have experienced hunger … how they always try to stockpile food … I have a similar panicky feeling right now that [makes me] treat all my online digital spaces as something that could be lost from me forever.So, I constantly want to keep myself safe.
What does it mean, 'to keep oneself safe' in this context?Léa suffered from the loss of both their archives within social networks, from the loss of their 'digital spaces' and tools for shaping their representation and self-narrative.Instead of turning to archiving as a solution, Léa intensified the networking and social aspects of online environments.Léa tried to save their 'places' by sharing administrative access with other people, and thus, sharing the ability and responsibility of keeping the materials safe: Now when I create a Telegram channel, I also add another person as an administrator, so that if [my access to] that channel is gone, it can stay somewhere.And it's a very strange feeling of unreliability.Before, it seemed [for me] like everything that was published on the internet would be kept by the internet forever.
So, the very act of producing digital traces in a blog, in a social network, in a digital letter was sometimes more important than an additional preservation.Creative solutions beyond archiving often targeted an ability to continue using mementoes actively.André considered usage of data analytical tools to play with all the saved amounts of data.De Norpoi wanted to write a novel based on the saved message histories.Robert printed out digital photos to decorate the room.*** Therefore, what are digital autotopographies, if we look at them from this perspective of memories put at risk? Autotopography refers to an organised group of evocative objects, and structures of objects-to-remember-with that people create around themselves.
In existing research, the idea of autotopography implies some extent of stability, especially in studies based on the analysis of mementoes in people's homes (Petrelli et al. 2008;Petrelli and Whittaker, 2010).Cases that I describe above show that we could also understand autotopography as processual and interpretative.The disruption of online memory practices did not destroy the usual autotopographies of my interlocutors but made people adjust their autotopographies to the situation.
When some usual parts of autotopography were not reliable or comfortable anymore, people started shifting their practices to other spaces, objects, and practices.In this manner autotopographies were 'fixing' themselves fast, never allowing people to end up without means for the autobiographical process at all.This observation helps to move from an understanding of autotopography as static to processual and flexible.
Thus, internet environments are a dynamic part of an embedded autobiographical process, where different materials are being redefined, reclassified, filtered, and interpreted in all types of other practical ways.But what ensures such adaptability?I suppose that this can be explained if we distinguish between two understandings of autotopographies: through a broader approach and a focused, narrowed approach.

Narrow understanding of autotopography
We can narrow our understanding of autotopography to observe only organised groups of curated evocative objects.Those can include photo collections, diaries, databases, archives of messages, et cetera.We can assume that these types of autotopographies are usually stabilised and shaped by particular traditions of preservation (they may be inherited from the previous generations, mass culture, friends, material culture, et cetera).Structured photo albums in VK, carefully selected pictures of one's child in Google Drive or a library of notes -people use digital devices and online environments to carefully curate parts of their autotopographies that are most important for them.Some of my interlocutors already had curated autotopographies, and in the risk situations, they knew very well what they wanted to preserve.For example, Albertine used Instagram as a photo album that captures different moments of life, and they already have regularly been archiving these pictures onto local devices.So, in the context of censorship, they held onto these practices.

Broader understanding of autotopography
At the same time, we can approach autotopography in a broad sense, as a network of potential and curated evocative objects that can be enacted in different constellations and practices, based on the context and challenges the person is facing.
Social media are environments where the creation of autotopography can happen even without people's concern.People leave digital traces.Messages and photos are produced and shared in extensive flows.This creates a field of many options from which people can choose when they need to hold onto something that evokes their past.The opportunities might not be realised unless usual memory practices are put in danger.When memory practices become disrupted, or access to curated evocative objects is endangered, a person could turn to previously passive parts of autotopography to ensure the process of self-remembering.

Conclusion
This paper describes and illuminates what was happening with tech-savvy educated people between 20 and 40 years old in Russia, while usual digital tools and environments for the autobiographical process were changing under the social media censorship in the Spring of 2022.With eight in-depth interviews, I explored these changes while they were happening.These conversations illuminated the ways people navigated the challenges practically and made sense of them.Facing censorship of platforms, surveillance, and an inability to pay for services, people started adjusting.As part of these adjustments, they deleted profiles, migrated to other platforms, censored themselves (deleted old posts and changed the style of posting), and often created archives of autobiographically meaningful materials.
The first finding is that people who tried to archive their mementoes (saving their poetry, photos from different life periods, pictures of their kids) heavily relied on technological affordances.Objects were de-networked and taken out of the platform interface, therefore, they changed meanings.This made users face a problem of not knowing what to do with de-networked memories.I interpret this problem as either a lack of technological instruments that would connect both online and offline practices of selfremembering, or a gap in culturally available tools.
Second finding concerns our understanding of platform-mediated memories.Trying to adjust to changes, people enacted digital traces that were never meant to be mementos before.To understand this phenomenon, I suggested two notions: potential and curated evocative objects.
Third finding addresses the importance of a dynamic, ongoing autobiographical process.Taking this observation to the level of theoretical discussions, I further developed the notion of autotopography, already usefully but briefly introduced by other researchers.I argue that autotopography is a handy conceptual tool to address the hybrid materiality of the autobiographical process without losing sensitivity to the vivid, ongoing remembering of Self.
This research has its limitations.First, I worked with a small sample within a particular social group, which does not give a full overview of social practices and reactions.For example, it is easy to imagine people whose relationships with online media were completely different; if people lacked digital literacy, they could have lost access to Instagram and Facebook completely.People who were not bothered by surveillance could transfer from Facebook and Instagram to VK.Because of this, my insights should not be applied to describe the Russian population in general.Still, the offered sampling showed a variety of possible risks, reactions, challenges, and practices that show how online memory tools and places can be at risk, but also what role social media can play in autobiographical remembering.
Second, I concentrated on a case of malfunctioning.Further research might be needed to calibrate these results.For example, to what extent can we address autotopography as a highly dynamic and creative process, if people use social networks under stable, peaceful circumstances.To further understand this, a follow-up, quantitative or comparative research might be needed.
Anya Shchetvina (she/they) is a PhD student and research fellow at the research group 'Kleine Formen' at Humboldt University.They work on an intersection of Internet Studies and Cultural Studies.Their research interests include internet imaginaries and metaphors, internet manifestos, web history, networked memories, autobiographical process, and Self.
Generally, my Instagram tactics have not yet changed.I need to use a VPN, yes, but now I have the Instagram archive, and if anything happens, most of my life until March has been saved there.