Reconstructive Metadata: A Self-Portrait of the Feminist Reader

by: Leela O'Connor , November 10, 2025

Memordata is a hand-sized box filled with index cards, styled after an old catalogue drawer. Written over four months in 2022 in response to my daily life, these cards form a self-portrait of the feminist reader. Each holds information about books I have read, specifically the vital ways they inform how I walk through the world. The result is metadata that is subversive because it is idiosyncratic; it seeks to reflect the intersection between political theory and personal reality. Overall, Memordata uses a reconstructive form to invite the viewer to honour their individual needs as a feminist reader, allowing them to ‘choose their own adventure’ when moving from one card to another.

Traditional Metadata

‘Metadata’ is simply data which describes other data. The described data could be any resource, such as a book, a file, an audio recording, or an artefact. Metadata then includes key descriptors of this resource, such as when it was made, its creator, its subject matter, and the materials used. Imagine a library—how we describe the books’ genres affects where they are placed, with biographies in one aisle and fiction in another. This is how metadata literally informs the physical spaces we move through. It is also analogous to the significant role metadata plays in our digital lives.

Upon starting my Master’s degree in Librarianship, I quickly learned that metadata has a reputation. Some students dread learning metadata processes because they require a high level of precision and repetition. To ensure your metadata is consistent with other people’s, you often must follow the rules of a pre-existing standard. Down to the punctuation, all formatting must be perfectly followed. You may even determine which words to use from a pre-existing set of ‘vocabularies’; the authorised word is then written not just in letters but in an identifying number code.

VRA Core is an example of a metadata standard. Below is their guide to recording the materials used to create an artwork. (Visual Resource Association 2007: 18).

Moving Beyond

Despite the formulaic nature of metadata, what struck me was the task’s simultaneous open-endedness. Ironically, creating metadata is a task that requires us to engage with deeply philosophical ideas. For example, when I describe the contents of a resource, I wonder what it means for a text to be ‘about’ a subject? When I try to make a source easily discoverable, I consider whether people can search for entirely new information, or ultimately only for what they already know exists.

One particular point was jarring for me: the supposed simplicity of cause and effect in the task I was asked to undertake. The best practice is to first identify ‘the user’s need’ and then make your metadata relevant to it. For example, if you determine your resource is going to be sought out by art history scholars, you include precise descriptive words for visual elements. My issue with this is that, in a feminist worldview, ‘the user’s need’ may not neatly pre-exist a text. Instead, texts are filled with radical theories that inform our worldview, our social reality, and our sense of what we may need. Audre Lorde said it best in the opening line of Poetry is Not a Luxury: ‘the quality of light by which we scrutinise our lives has direct bearing upon the product which we live’ (1984: 36). For Lorde, poems (referring broadly to authentic expressions), ‘formulate our implications of ourselves, what we feel within and dare make real’ (1984: 39).

There can be a time before you are given the language to seek what you need. I know this because I was once an isolated feminist reader. Ironically, this was when I was undertaking a formal humanities education at university in my late teens. I was learning about many aspects of society, but my readings were heavily dominated by male perspectives, whether in ‘objective’ political analysis or philosophical propositions. I was appropriately reverent, earnestly engaging with these texts. Sara Ahmed describes her similar experience at the same university decades earlier and recalls the implicit command to align oneself with a single male theorist (2017: 15). When I did find texts that offered me a more dynamic, agitating account of the structures I had witnessed my whole life—those that provided instructions for resistance—I devoured these lifelines alone in empty tutorial rooms after class.

Once, my male lecturer asked me if I felt that being born female had affected my life. I gave him a tepid account, filled with the subtlest of claims, because I didn’t want to undermine myself by seeming uncritical in my argument. I had not yet been emboldened by the solidarity and clarity that further feminist reading would provide me, enabling me to engage with long-established fields of critical discussion. Ahmed argues that diverse authors are often not cited in academia even when their publications are readily available (2013). For example, she recalls attending a panel on reproductive justice where most papers solely cited male philosophers, despite feminists having ‘rather extensively’ written on the subject (2013). She argues that the act of citation is a ‘reproductive technology’, a method through which disciplines reinforce their power hierarchies and screen which bodies are welcome into their space. I believe that, in our reality in which violence against women has been normalised, it is akin to gaslighting for university curricula to deprive students of the diversity of scholarship. I had a clear, desperate ‘user’s need’ for texts I had never been exposed to.

Even as I consciously sought out more by creating a feminist reading group with friends to explore texts excluded from our courses, I knew the scope of our reading list was still impacted. I later joined the Reading Feminism book club that introduced me to its members’ deep pool of gender studies knowledge. The book club quickly structured my months with the practice of reading the chosen pieces of theory and the ritual of sharing them. As a group, we unravelled how the content of these books intersected with each of our lives—informing our understanding of education, family, work, experiences of oppression, and friendship. Uncovering each link felt revelatory; reading feminist theory was an act of witnessing one another. Experiencing this dynamic interchange between text and experience, I wondered: how could metadata tell this story?

Making the Artwork

The goal was to create ‘metadata’ that was grounded in context—a directory for the isolated feminist reader I once was. This context was my everyday experiences, specifically those experiences where theory is made relevant by the personal reality that reflects (and informs) it. For these connections to emerge organically, I needed time. I went to my local newsagent and bought a pack of index cards, carrying them around with me for four months. Over these months, I wrote down a note each time I had a gendered experience. For the purposes of this artwork, that meant moments when I paused and experienced a clarity about how overarching structures informed my position. Such instances occurred in both new and familiar places: through conversation with close friends and family, individual reflection, or more distinct catalysing experiences like being groped by a customer at my retail job. I then linked each of these moments to texts I had read that had informed my understanding of what I experienced. I drew on authors who deconstruct political structures in their respective ways, offering everything from personal accounts, acknowledgement, mourning, celebration, and resistance. Identifying each of these influential texts alludes to Ahmed’s practice of citation as feminist memory (2017: 15).

The result is metadata about books deeply contextualised in the user’s needs, including my own and potentially those of others. Young people might not already have these particular books on their ‘to be read’ lists, but they have likely had similar life experiences, for which I offer these books as a guide.

Below are six examples of my index cards:

For each of the texts that are mentioned in the ‘experience’ cards, I created a dedicated index card that summarises the reading. These descriptions are not typical blurbs like you would find on the back of books. Instead, I combed through the notes in the margins of my books to identify which themes and techniques had emerged as most significant during my reading, and during my book club’s discussions.

A list of texts mentioned in Memordata.

Resistance in Data Practice

Choosing to play with metadata is part of a contemporary shift in data practices. Increasingly, there is recognition that data is never innocuous. Ellen Van Neerven frames data practice as the process of narrative-making about the data’s subject, a process made more powerful by its typically concealed nature (2019).

Some resistance is apparent in formal industry changes. For example, there are critiques of the established controlled vocabularies providing outdated and offensive metadata terms. In response, professionals have formed new standards, such as AUSTLANG, which presents codes for specific Indigenous Australian languages, as opposed to the generic term ‘Australian languages’ (National and State Libraries Australasia n.d.).

There is also a debate about the extent to which the description of texts should be ‘interpretative’.Gailey notes how the industry priority is to keep markup succinct and accurate so the item can be searched easily by people using search engines, which she argues does not account for ‘a reader’s extended engagement with the text’ (2011: 133). Instead, engagement is supported by recording subjective context to the text, such as allusions to key historic figures, connections to other texts, and likely metaphors (Gailey 2011: 133). Navigating this is critical for people working with culturally significant collections. For example, the Orlando Project, an archive of women’s writing, adopts a feminist methodology by intentionally including semantic tags related to subjectively interpreted elements such as motifs and narrative voices that enable users ‘to make meaningful connections between writers, culture, and history’ (Schilperoort 2015: 16). They have an ‘intertextuality tag’ that they apply to suggest to the user connections between different texts (The Orlando Project n.d.). One intertextuality tag reads: ‘Critic Rees-Jones sees in the title of Carol Ann Duffy’s Fifth Last Song: Twenty-one Love Poems a reference to Adrienne Rich’s ‘Twenty-One Love Poems’ in A Dream of a Common Language, 1978.’ (The Orlando Project n.d.). The subcategory of this particular intertextuality tag is ‘allusion acknowledged,’ while other possible subcategories are to mark instances when one text has quoted, parodied, imitated, or adapted another text (The Orlando Project n.d.). It is with similar practices that Gailey proposes a future that recognises the value of ‘nonliteral and nonstandard [metadata] content’ to which people could opt in or out of when searching (2011: 140).

Interpretative metadata was indeed shown to be impactful by the staff at the Germaine Greer Archive. Previously, a group of archivists wrote over a million words of metadata that describe the archive’s contents (Weber & Buchanan 2019). Staff ran a sentiment analysis program over some of this metadata, which tracks the frequency of specific words used and, based on the connotations of these words, summarises the attitude of that piece of text (Weber & Buchanan 2019). The researchers found that their ‘richly textual’ metadata itself varied as positive, negative, or ambivalent, which corresponded to the tone of the respective resources the metadata described (Weber & Buchanan 2019). Thus, they conclude that, via their archivists’ affective experiences, their metadata successfully ‘reveals rather than obscures, the collection’s emotive landscape’ (Weber & Buchanan 2019).

I have also been deeply moved by artistic reclamations of data. Living in Australia, I have read literature by Indigenous Australian authors that embeds violent historical government documentation within their own storytelling. Among other books, my readings included Natalie Harkin’s Archival-Poetics, Kim Scott’s Benang, and Elfie Shiosaki’s Homecoming. Harkin reckons with state records about her grandmother by approaching them as ‘threads to weave’ into her poetry (2019: 32), while Scott’s novel adopts governmental terms that were used to categorise humans by race but italicises them as parody (Scott 1999: 85). The ‘authority’ of these dehumanising records is fundamentally undermined when they are relocated within biographic writing that asserts the personhood of the subjects.

Across all these varied responses to data, the subversive act is simply adding new and sufficient context to the original data. I aspire to this mechanism in Memordata.

Reconstructive Forms as an Invitation

My chosen method of filing these index cards enables Memordata to function as ‘reconstructive metadata.’ The cards are all secured to the box with a wooden dowel that runs through them. This form is modelled after a catalogue drawer. Traditional libraries had rows of catalogue drawers with cards one could browse by subject to identify an author and text and then be directed to the correct shelf location. A single drawer made from regular office materials like pen, paper, and cardboard, Memordata guides the user to investigate further. With twenty-eight cards in total, each card refers to another card, or several other cards, for related information (e.g. ‘see 8’ or ‘referenced in C, J’). Both text cards and personal experience cards can lead to either card type. Therefore, depending on their choices, users can follow many different pathways.

Enabling various possible pathways is a sought-after effect in the literary trend known as ‘reconstructive literature.’ Reconstructive literature requires the reader to bring together parts of a fragmented narrative (Funk 2012: 42). For example, book chapters might be printed in non-chronological order, so the reader must piece together the chronology. This is a way to live amongst disruption: you do not deny inherent chaos but instead celebrate your ability to choose to piece together a personal and meaningful representation. In reconstructive literature, authenticity exists purely and sufficiently within the very attempt to represent one’s experience (Funk 2015: 3). I made further choices to enhance the reconstructive form of this metadata. I decided against filing the book summary cards in chronological order of publication. Instead, I value two concurrent chronologies: I provide both the year published and the year I read the book on each index card, filing it according to the latter. Ultimately, the definitive order is open to each person who browses it. They can construct their path between different experiences and books according to their own feminist interpretation, to their own unique ‘user need.’

Data as Self-Portrait

Memordata is not a normative vision of a feminist metadata methodology. Instead, I utilise metadata as a tool for capturing a self-portrait of a feminist reader existing within networks.

Data self-portraits are a rapidly emerging form of portraiture, likely in response to our high-surveillance environment in which data is big business (Sampaio et al. 2019: 110). Donath identifies data portraits as distinct from data visualisation (2014: 2). While data portraits are still a form of visualisation, the distinction is made because portraits are not concerned with scientific accuracy, but instead fully embrace subjectivity (Donath 2014: 2). The artist must make intentional choices about what ‘salient and emblematic data’ captures the character of their subject (Donath 2014: 3).

If the data portrait is a distinct phenomenon from simple data visualisation, I would argue that the metadata self-portrait is a particular type of data portrait. Many human activities will generate some primary data as a byproduct. In contrast, the creation of metadata is usually the intentional act of categorisation; it is explicitly and inherently a descriptive task. Therefore, I find it especially primed with narrative potential. This nuanced difference between my metadata portrait and most data portraits is reflected in the form; in contrast to Memordata, Sampaio et al. find that the majority of data portraits are abstract visualisations of data that is passively collected (2019: 117 & 119). For example, TimeMachine by CADA (2012) tracks participants’ time spent in different locations, and the abstract digital images then feature different colours according to the findings.

However, regardless of form, the properties of data allow data portraits to share key functional similarities. Data offers an expansive temporality to a portrait. Data recorded about a subject over a period of time can reveal patterns, providing insight into the person’s past but also emboldening the viewer to make predictions about their future. For example, for her 2014 work The Outage: Her Story, Erica Scourti hired internet security experts to form a profile of her online life that would later be converted into a ghostwritten memoir, with the team acquiring data such as her email texts and web searches (Newman 2020: 22). Newman argues that Scourti is being ‘algorithmically written,’ with the data functioning not simply as proof of the subject’s presence, but as a projection of their character which includes their probable future behaviours (2020: 22). Temporality is also central to the effectiveness of Katie Lewis’s 201 Days (2009), as suggested by the title. Lewis’s final work appears as a dense winding of red strings, but the spread and density precisely correspond to where and how deeply Lewis felt numbness in her body each day (Gaddy 2011). Counterintuitively, each day of data accumulation actually tells a story of deterioration; recording data about her intangible chronic pain translates it beyond her body, allowing the viewer to bear witness. Data in these works functions as a dynamic mapping of the self. This is ideal for my self-portrait because I aim to capture the movement of self within a network of text. I trusted that recording the index cards over a period of four months would reveal and generate meaningful patterns. The expansive network potential of the metadata, including the connections drawn between cards by both the viewers and me, can mimetically model how text operates within my life.

Conclusion

Creating metadata about books according to the user’s needs is the task that led me to deeply reflect on what it means to be a feminist reader. I am all too familiar with the user’s need for texts, having been an isolated feminist reader in my youth. I believe all metadata is storytelling, which is true even when it is standardised and spare, creating a story that is not yet whole. In Memordata, I strive to use metadata as a tool of self-portraiture to testify about the profound ways feminist theory informs my work, art, and relationships. Generated over four months of daily life, my metadata captures the expansive, emancipatory relationship between text and self, where one does not neatly pre-exist the other. This complex relationship is celebrated by a reconstructive form that invites viewers to develop their own authentic navigation between cards and honour their individual needs as a feminist reader.