1. Introduction
Shortly after 6 pm on the 26 May 2018, the result of the referendum to repeal the Eighth Amendment to the Irish Constitution was announced. The Eighth Amendment made the right to life of the unborn equal to that of the mother, thereby prohibiting abortion in virtually all circumstances. The two-thirds margin of victory for the “yes” vote was the result of decades of grassroots feminist campaigning since the initial amendment was passed in 1982, crystallising around a number of prominent cases of marginalised women being endangered by the state’s denial of abortion services – among them the cases of Ms. X, Ms. Y and Savita Halappanavar, who died in a Galway hospital in 2012 after being denied an abortion. The campaign brought about a movement of feminist activisms from the margins to the political centre stage: during the years and months preceding the referendum, Ireland saw a “mobilisation on the scale of 1970s feminism” (Connolly Reference Connolly, Browne and Calkin2020, p. 51) that brought activists into coalition with establishment politicians, medical professionals and celebrities.
This coalition also included a large population of archivists and other cultural heritage professionals, who formed a collective that would operate as a steering group for the archiving of materials related to the referendum, eventually winning funding for a major digital archiving initiative. The 36-month Archiving Reproductive Health (ARH) project began at the Digital Repository of Ireland (DRI) in 2021 and was led by researchers at Trinity College Dublin, Maynooth University and the Royal Irish Academy. ARH provided an infrastructure for the long-term preservation of material from activist organisations and developed best practice around specific archiving concerns to do with the referendum, such as the archiving of social media posts and issues of access to sensitive or personal materials that had been shared during the campaign. Collections archived by the ARH project would find a permanent home in the DRI, a Trusted Digital Repository (TDR) designed to house Ireland’s cultural heritage, social science and humanities data.
Repositories and other archival database systems are essential tools for institutions undertaking rapid-response and contemporary collecting, enabling the preservation of born-digital and digital surrogate materials, and their quick dissemination to the public. Within partnerships like ARH, they contribute infrastructures for the long-term preservation of digital materials, expert support for community archiving initiatives and facilities for open access and sharing of collection objects. Despite the increasing prominence of these large-scale, post-custodial repository systems in the archival landscape (Carbajal and Caswell Reference Carbajal and Caswell2021; Thylstrup Reference Thylstrup2019), there remains a paucity of critique surrounding the politics of their mediation of archival data. By offering an account of one such system that situates it within its collections’ social, political and institutional context, this article theorises the archival database as a historiographical infrastructure, tracing the ways in which the database’s framing of its contents promotes or occludes the emergence of certain historical narratives. Drawing attention to the database as a constituent of the production of public histories prompts reflection not only on the role of digital forms in these histories, but also on the performance of “ethical remembering” (Higgins Reference Higgins2016; Nic Dháibhéid Reference Nic Dháibhéid2023) within Ireland’s archival infrastructures (Smyth Reference Smyth2022). Moreover, in attending to how the cultural heritage database’s humanity is expressed through its infrastructures’ multiple registers of use, this article endeavours to complicate the potential relation between AI and the archive.
In this article, I argue that the digital infrastructures that underpin the ARH collection are integral to the historiographical function of the collection, which is the central national resource for the preservation of campaign and activist materials from the referendum. These infrastructures – which encompass the layout of the interface, the metadata standards and protocols governing access and reuse – are not neutral, either in their construction or their mediating effect. By promoting or occluding particular narrative possibilities and producing novel systems of relation between collection objects, database infrastructures construct histories, propagate power through a collection and shape the affective responses of the collection’s user base. Simultaneously, by rendering collection information machine-readable, they turn it into data. This transformation engenders a discursive change in how collections information is received, but it also has the material effects of making that information tractable to computation, and of introducing algorithmic logics of scale and practices of visibility to a cultural heritage domain that had previously been governed by an object-based, narrative set of logics. Yet because database infrastructures are invisible to the end user, they are often overlooked in critical appraisals of museum cataloguing, which emphasise cataloguing’s discursive effects over the operational dimension of the catalogue’s computational form (Watson Reference Watson2023).
This analysis is made more urgent by recent advances in the use of AI to perform archival tasks, such as those described by Lise Jaillant et al. (Reference Jaillant, Aske, Caputo, Jaillant, Warwick, Gooding, Aske, Layne-Worthey and Downie2025) in their discussion of prospective AI-related developments at the UK’s National Archives. They outline AI’s potential uses in selecting records for preservation, automating the process of sensitivity review and improving discovery by replacing keyword search with a finding aid based on topic models; these functions are of particular utility in relation to the large-scale government archives housed at the National Archives and similar repositories. In addition, the acceleration in development of generative AI and of large language models (LLMs) like OpenAI’s GPT and Anthropic’s Claude unlocks more complex – and less directly governable – use cases for AI in the archive; fully autonomous, agentic AI research assistants, which proactively find new information related to a researcher’s area of interest, will likely be prominent users of archival repositories in the very near future, and the extent of the use of open source archival records for model training has yet to become public knowledge. It is moreover likely that future partnerships between heritage organisations and industry will involve LLMs and their vendors; as members of the public increasingly use technologies like ChatGPT to mediate online search, it is not inconceivable that archival frontends and search interfaces based on similar technologies could proliferate within the next five years.
Given this context, I argue that multiplying the definition, and accordingly, the mediating possibilities, of the cultural heritage database infrastructure is essential to ensuring that its complexity is carried forward into the encroaching wave of AI innovation. I seek to demonstrate through this article’s analysis that the digital repository infrastructure is not merely a container for collections information as data, but is also a curatorial technology, through which the collection’s politics are articulated, its affective expressions felt, and its historiographical narrative experienced. Situating the ARH database in its contexts makes clear that its infrastructure is integral to the ways in which contested feminisms and the particular Irish dynamic of silence and shame are negotiated in the collection; these ideas are important not only for the points of view they express, but also because they raise critical questions about the ways in which AI might be approached in archival contexts going forward. This article’s portrait of the repository infrastructure as an active participant in the collection’s mediation of its constituent objects seeks to broaden the lens of the critical discussion surrounding archives and AI by expanding our shared conception of what existing database systems achieve in the context of the collection. The situated case study that follows offers an example of how a database infrastructure operates in the context of Irish feminism and in doing so moves towards a methodological intervention that repositions that infrastructure as a meaning-making force, a more-than-data object that cannot so neatly capitulate to the demands of the present wave of AI development.
2. From #Repealthe8th to #Archivingthe8th
The collections archived in the DRI originate primarily in pro-choice campaigning that took place from 2012 until the passage of the referendum in 2018. Although pro-choice campaigns in Ireland had been active since the 1980s, 2012 was a landmark year, during which the groundwork for the referendum campaigns was laid: the Abortion Rights Campaign (ARC) was founded, and people lined the streets of Dublin in protest twice, for the first annual March for Choice in September, and again in October following Savita Halappanavar’s death (Maguire and Murphy Reference Maguire and Murphy2023). The referendum was called in January 2018, following recommendations made by a Citizens’ Assembly, made up of ninety-nine randomly selected Irish voters, which sat between October 2016 and April 2017 (Side Reference Side2022). Much of the activism surrounding the referendum took place online, with a high level of campaign activity on Twitter/X and the emergence of pro-choice storytelling accounts there and on Facebook (Statham and Ringrow Reference Statham and Ringrow2022). The most prominent of these, In Her Shoes, was a Facebook page where women submitted anonymous accounts of their experiences of abortion under the Eighth Amendment and which gained over 100,000 followers between January and May of 2018 (Darcy Reference Darcy2020). Alongside abortion stories told in national outlets by prominent figures such as Róisín Ingle, Tara Flynn and Susan Cahill, In Her Shoes exemplified Irish women’s awareness “that the stories of their personal reproductive bodies are political, and that by speaking openly they are playing a key role in rewriting how abortion is conceptualized” (Mishler Reference Mishler2021). The importance of this role made the archiving of the In Her Shoes posts a central priority for the ARH project (Archiving Reproductive Health Project at the Digital Repository of Ireland 2023a).
At the same time as campaigns for the Eighth Amendment’s repeal were beginning to take hold, Ireland was also entering a significant moment in archival commemoration. The Decade of Centenaries was an umbrella term for the period between 2012 and 2023, during which the centenaries of the events that led to the establishment of the Irish Free State were marked with public commemorations. Digital archiving projects were central to how the commemorations took place, with aims ranging from the digitisation of new archives and collections to the interpretation and mediation of existing ones. The Irish State’s support for this type of work demonstrates an emphasis on the digital as a commemorative mode, as well as a desire to democratise access to public history; Mike Cronin links this to Ireland’s focus on developing its knowledge economy, itself a result of the high concentration of global tech companies with an Irish presence (Cronin Reference Cronin2017). Hannah K. Smyth notes a significant increase in funding for digital archival projects associated with the Decade, describing a “volte-face” in the funding situation of the National Archives of Ireland compared to the period before 2012 (Smyth Reference Smyth2022, p. 171).
The combination of ethical remembering and increased funds allocated to digital archiving led to an “archival turn in Irish feminism” (Smyth Reference Smyth2022, pp. 185–186). The Mná 100Footnote 1 project, which produced a timeline of Irish women’s participation in Irish political and public life over the previous 100 years, was one of the flagship initiatives of the latter half of the Decade; another major initiative, the Military Service Pensions Archive, sought to use its newly digitised records to spotlight the experiences of women active in military service (Smyth Reference Smyth2022, p. 183). As Smyth points out, these commemorative projects directly influenced the repeal campaign. The 2016 iteration of the annual March for Choice used remixed imagery from Maud Gonne’s Inghinidhe na hÉireannFootnote 2 organisation in its promotional materials, as well as the slogan “Rise and Repeal,” which invoked the 1916 Rising; posters and banners from the march are now housed at the National Museum of Ireland and appear in digital format in the DRI. This recursive movement of materials and ideas from the newly digitised archive, into a second public life, and back into the archive in a new form is the result of the collision of Ireland’s unique archival moment with a contemporary breaking of the silence and shame that had hitherto surrounded reproductive justice in the country. This moment accounts for the readiness both of Ireland’s feminist activists and of its memory institutions to assemble an archive of the Repeal movement. Moreover, the overlap between those activists and cultural heritage professionals contributed to the scale and rapidity of the associated collection, as well as the fact that it found an immediate home within an established institution, rather than existing as an independent feminist archive.
3. The Eighth Amendment’s contested legacies
3.1 Irish feminism(s) and the activist archive
Through close reading and situated analysis of ARH’s database, I track its intervention into two key tensions that animate the production and reception of a campaign archive such as this one. The first of these has to do with the character and point of view of Irish feminism. The 2018 referendum campaign, led by the broad coalition of the Together for Yes group, united disparate activist groups like ARC into a feminist voting bloc even as it crafted campaign narratives around those abortion stories deemed most palatable to so-called middle Ireland (Kennedy Reference Kennedy2022). In the wake of Repeal, feminist attention turned towards those who had been at the margins of the official campaign. Deirdre Niamh Duffy’s (Reference Duffy2020) study of abortion health activist groups tracks a post-referendum movement from an anarchic feminism, characterised by direct action and the clandestine provision of services, to a decolonising feminism, whose strategies for political resistance are focused on systemic oppressions rather than the rights of individual legal subjects, and which emphasises the continued mediation of abortion access by the border between Britain and Ireland, centring the barriers still faced by migrants and refugees to obtaining abortion care. Others, such as Chakravarty et al. (Reference Chakravarty, Feldman and Penney2020), argue that Irish feminism was too quick to rehabilitate the problematic aspects of the Together for Yes campaign, including its use of Savita Halappanavar’s image without a meaningful engagement with the racial politics of her death or of her representation in the campaign. A decolonising of Irish feminism would necessitate a “pluriversalising” of knowledge practices beyond the university, and beyond the leadership of white women (Chakravarty et al. Reference Chakravarty, Feldman and Penney2020, p.186).
These contrasting portraits of the state of Irish feminism post-2018 suggest that the differences that were already evident within Together for Yes have been amplified during the aftermath of the campaign. As the referendum passes into history, the strategic coalition that won the “yes” vote in 2018 has since dissolved into a plurality of single-issue groups, writers, and activists, from within an Irish left increasingly preoccupied by inequalities of income and opportunity, including a major housing crisis, and the growing threat of the far right. Within this context, the process of assembling an archive of the Eighth Amendment is highly political: the narrative of the campaign that emerges through the archive’s inclusions and exclusions, as well as through the infrastructural mediations described in this article, is liable to be enfolded into official retellings of the struggle for reproductive rights in Ireland. Digital cataloguing forms’ latent potential for accommodating narrative multiplicity has been theorised across multiple contexts in recent years (Duarte and Belarde-Lewis Reference Duarte and Belarde-Lewis2015; Knight Reference Knight2022); it is alongside these that my work seeks to identify moments where this theoretical possibility is realised – and moments where it is deferred.
3.2. Silences/openness
A second, related tension arises from the interplay between the affective politics of the Eighth Amendment – the shame and silence that were used as tools of control in Ireland during the 20th century (Fischer Reference Fischer2019), and the twin emotions of relief and ambivalence that characterised the aftermath of Repeal – and the open-source imperatives of a TDR. As a TDR, the DRI is a repository for academic, cultural, and social science data, working with partner organisations to ensure the long-term preservation of their materials; institutional transparency around data storage and governance is a condition of the TDR certification process (CoreTrustSeal, 2022). As such, the DRI is fully compliant with the FAIR principles, a set of data management guidelines according to which data can be made Findable, Accessible, Interoperable and Reusable (Wilkinson et al. Reference Wilkinson, Dumontier, Aalbersberg, Appleton, Axton, Baak, Blomberg, Boiten, Bonino da Silva Santos, Bourne, Bouwman, Brookes, Clark, Crosas, Dillo, Dumon, Edmunds, Evelo and Mons2016). These principles’ emphasis on removing barriers to accessing data and other digital materials has been critiqued by groups such as the Global Indigenous Data Alliance, whose CARE principles extend the FAIR principles by offering a counterpart that is oriented towards maintaining the sovereignty of Indigenous groups over their own data (Carroll et al. Reference Carroll, Garba, Figueroa-Rodríguez, Holbrook, Lovett, Materechera, Parsons, Raseroka, Rodriguez-Lonebear, Rowe, Sara, Walker, Anderson and Hudson2020). Open-source technologies like those that underpin the DRI have also been held as examples of the political right’s cyberlibertarian techno-philosophy (Golumbia Reference Golumbia2024; Winner Reference Winner1997); the left-wing activist archive, particularly one largely made up of personal testimonies that amount to a collective of resistant voices unto themselves, is an unlikely bedfellow to these technologies. Simultaneously, however, archival infrastructures provide a less precarious, more stable home for this collection than the social media platforms where much of its material originally appeared; amid increased association between platforms like Facebook and X and the second Trump administration in the US, those platforms’ already questionable ability to house and retain culturally valuable information has been compromised further.
The ambivalence of the relation between the collection and its platform can be further contextualised through attention to the ways in which silence, shame and secrecy are conceptualised both in an archival and an Irish context. The idea of silence is often invoked by feminist interventions into traditional archives, which draw on an understanding of the archive not merely as a repository of information but as the ground truth of political and cultural power, a set of discourses whose selection, prioritisation and transformation are revealing of the system of rules by which a society functions (Foucault [1972], Foucault Reference Foucault2013). Within this framework, archival silences are those discourses deprioritised by that society, whose denigration is rooted in an actual documentary absence, but unfolds into a more generalised absence from history. Kate Eichhorn argues that this archival turn’s effect on grassroots feminist cultural production is twofold: first, by providing a space for historiographical intervention and the realisation of a “genealogical politics” (Brown Reference Brown2001) that denaturalises the a priori of neoliberalism, feminist attention to archival silence is an exercise in collectively imagining alternative modes of being in the world; second, through the collaboration of feminists self-publishing work in new forms with feminists working on the inside of cultural organisations, the archival turn prompted the migration of those new feminist documents into mainstream cultural heritage institutions, rendering the archive an active participant in new feminist organising (Eichhorn Reference Eichhorn2013). The archival turn provides mechanisms both for drawing attention to silence and for resolving it. Moreover, since the mid-1990s beginning of the zine archiving Eichhorn writes about, the politics of the archival turn have infiltrated mainstream cultural heritage policy-making: I read the Decade of Centenaries’ ethical remembering as a manifestation of these politics that became central to national commemoration in Ireland.
However, I also argue that the idea of silence has an additional affective weight in Irish feminism – particularly in relation to reproductive rights. In 2021, the Irish Government published the final report of the Commission of Investigation into Mother and Baby Homes, which had investigated deaths and misconduct that occurred in the mother and baby homes that proliferated across Ireland during the 20th century. These homes were Catholic institutions where unmarried women went or were sent to deliver their babies; they also arranged the adoption of “virtually all” the children born in the homes (Commission of Investigation into Mother and Baby Homes 2021). The investigation was prompted by the efforts of amateur historian Catherine Corless, whose work on the Bon Secours Mother and Baby Home in Tuam, Co. Galway led to the excavation of what had been a septic tank containing the bodies of an estimated 796 infants and children (Corless Reference Corless2021). This discovery followed a number of similar investigations during the early 21st century – among them the 2013 McAleese Report into the Magdalene Laundries and the 2005 Ferns Report into clerical sexual abuse – that had made public the extent of the individual and institutional abuse perpetrated by the Catholic Church in collusion with the Irish State since the State’s foundation.
Although these revelations came as a shock, it is also the case that the abuses they describe were not unknown to the people of Ireland. Lindsay Earner-Byrne (Reference Earner-Byrne2023) negotiates the family’s role in the dynamics of shame and guilt that were instrumental in their continuation, arguing that although families were culpable in the perpetuation of Catholic social hierarchies in Ireland, the family was also an important site where forms of resistance might emerge. Many people in Ireland had and have family members who were institutionalised or abused; many other families hid their pregnant daughters and sisters and raised their children as their own (Wills Reference Wills2024). Moreover, the narrative of the Irish family’s dispassionate rejection of its daughters became a tool by which the Church could shirk blame for its oppression of Irish women (Earner-Byrne Reference Earner-Byrne2023). Attention to these dynamics complicates the notion of archival silence as it relates to the Irish context, reconceptualising secrecy as a potential form of agency, and the hiding of a pregnancy from surveillant neighbours and clergy as a protective, rehabilitative step. Similarly, attention to our collective shame gives way to a reckoning with Irish people’s complicity in these violent oppressions, and to the ways in which they are somehow constitutive of Irishness (Lennon Reference Lennon, Fischer and Mahon2020), which in turn demonstrates the necessity of building an archive that richly complicates silences, rather than liberating its constituents from them.
The idea of shame also forms the affective dimension of what Clara Fischer characterises as an ongoing process of postcolonial nation-building in Ireland. She describes a post-Independence national imaginary that relied on the stabilising image of the virtuous Irish woman, whose purity came to stand in for the young nation’s resistance to “secular, seemingly liberal encroachments, deemed to be moral vices (and often described as Anglican and English in nature)” (Fischer Reference Fischer2019, p. 37). The actual denial of freedom to women, through prohibitions on abortion and contraception as well as the institutionalisation of pregnant women, was a means of removing threats to Irish national identity; the later culture of travel to England for abortion, enshrined as a right in the wake of the X case,Footnote 3 was another instance of what Fischer identifies as the gendered displacement of women caused by this commitment to moral purity (Fischer Reference Fischer2019, p.39). In the case of both silence and shame, Ireland made a structural force of a personal emotion, demonstrating a uniquely affective mode of governance: what S.E. Wilmer identifies as the manifestation of a postcolonial biopolitics (Wilmer Reference Wilmer, Wilmer and Žukauskaitė2015). Shame was a means by which illegitimacy was conferred on unwed mothers and their children, an essential tool for the biopolitical control of all women.
Wilmer, citing Foucault, suggests that resistance to such control can be found in the personal refusal of shame and the assertion of one’s right to happiness; Fischer similarly identifies the 2018 referendum result as a moment of rejection of the politics of shame. During the referendum campaign, the prominence of personal storytelling through platforms like In Her Shoes seemed to accelerate this refusal and is often cited as a crucial force in the resounding victory of the “yes” campaign (Barr Reference Barr2019; Bolt and O’Leary Reference Bolt and O’Leary2024; Ralph Reference Ralph2020). This analysis, however, risks placing the burden of resisting subjugation entirely on women themselves, ignoring the conditions of possibility for the breaking of silence around abortion that had begun to be established in the wake of both Savita Halappanavar’s death and the 2000s wave of investigations and scandals that had already destabilised the Catholic Church’s hegemony in Ireland. Writing in 1995, in the aftermath of the X case, Ruth Fletcher notes the prominent silence surrounding abortion and interrogates some of the reasons behind it. Through a series of interviews with women with personal stories of abortion, she concludes that their reluctance to share those stories, despite their evident potential to change the perception of abortion in Ireland, is rooted in four phenomena: women’s desire to protect themselves and their loved ones from stigma; their feelings of ambivalence about their abortions; a concern about hurting or burdening others; and a sense of frustration with a polarised political discourse that might homogenise their experience or misinterpret complex feelings as remorse (Fletcher Reference Fletcher1995).
These concerns demonstrate the precarity of the individual storyteller of abortion; the collectivisation of personal testimony evident in parts of the DRI’s collections, most particularly in the In Her Shoes subcollection, not only protects those storytellers but also adds to the impact of their stories. The reappearance of these stories in the campaign archive draws our attention to the ways in which the archive is itself a participant in the ongoing negotiation of silences in Ireland; it is the mechanism by which testimony becomes historical record at the same time as it introduces that testimony to a new mediating infrastructure, one that comes with its own prebuilt conceptualisation of openness, access to storytelling and institutional transparency.
4. The archival stack
The ARH collections are full of lively objects. The project comprises eleven core and eight additional collections, as well as four further collections from a public collection day held in March 2023. Collections fall broadly into two categories: organisational collections, which preserve material relating to a specific campaign organisation, and research datasets. The core collections come from organisations that were key stakeholders in the project: campaign materials from Together for Yes, as well as constituent organisations ARC, the Coalition to Repeal the Eighth and Termination for Medical Reasons (TfMR), make up the majority. Alongside these sit datasets of oral history interviews and collections of photographs and administrative documents that relate to the campaign but are not necessarily attached to a specific organisation. Two social media datasets make up the rest of the core collections: these are the archive of the In Her Shoes Facebook posts, and a Twitter/X dataset containing over 2 million tweets about the referendum sent between March and May of 2018. The additional collections, and those from the public collection days, contain material from regional, historical and specialist organisations, such as feminist activist groups in Cork, Galway, and Maynooth, Lawyers for Choice, and the Women’s Information Network. These collections also broaden the focus of the project away from the direct campaign to repeal the Eighth Amendment: oral history interviews about the Irish women’s movement collected by Linda Connolly and Mary Muldowney, as well as the archived podcast “How the Yes was Won,” connect the referendum campaign to a longer history of abortion activism and feminist struggle in Ireland, while the inclusion of a collection from Alliance for Choice Derry reflects the continuation of abortion activism in Northern Ireland after the success of the referendum in the Republic.
The DRI’s open access development philosophy has directly determined its infrastructure, which is aligned with the Open Archival Information System reference model, an ISO standard conceptual framework for the storage of digital objects (Lavoie Reference Lavoie2000). Beginning from the database’s material components and working upward through the stack, the DRI consists of physical storage clusters based at Trinity College Dublin, Maynooth University and a third backup site; a set of physical virtual machine servers; and a virtualised environment that hosts the services that deliver the repository, including the user interface. The architecture’s structure was determined by a user-focused requirements gathering process and built following Agile methodologies; it adapts existing open-source technologies, like the Hydra/Samvera framework and the Fedora storage system for digital repositories. These systems were originally developed by a collective of UK academic institutions in response to a demand for a highly customisable repository infrastructure; digital repositories were beginning to proliferate as a result of storage demands for open-source e-prints in the early 2000s, and technologies like Hydra simplified their development (Awre and Green Reference Awre and Green2017). This places the DRI firmly in a genealogy of open source repositories, and more broadly, increased public access to (scientific) information within the university sector. It departs from these other, comparable systems by its specific focus on preserving data from the social sciences and humanities.
The metadata layer is the part of the DRI’s stack that is of most concern for this article. The repository is built to be metadata agnostic, supporting those ontologies and vocabularies that are most commonly used in the humanities – the Dublin Core (DC) standard for academic collections, the Encoded Archival Description archival standard, and Machine-Readable Cataloging and Metadata Object Description Schema for library collections – but remaining open to the use of new vocabularies and tools in the future (Digital Repository of Ireland 2015). Decisions for which standards to support were initially made on the bases of their popularity and their ability to facilitate search and linkage; later, the DRI adopted the FAIR principles on the recommendation of its Metadata Taskforce, and became a national aggregator to Europeana, both of which impacted the repository’s metadata provision (Digital Repository of Ireland 2020). In practice, this has led to considerable emphasis on the production of DC metadata, especially within ARH.
Following recent movements in the digital humanities and related disciplines (Bratton Reference Bratton2016; Liu Reference Liu2020; Webb Reference Webb2023), my analysis of the DRI views it as a full-stack mediating phenomenon, whose infrastructural politics are felt not only in the interface layer, but down through its code, metadata, standards and protocols and hardware. The structure of this article’s second half is loosely aligned with the layers of the stack, proceeding from a reading of the interface down towards the encoding of access to archival materials; by retaining this connection to the DRI stack, I emphasise the technological specificity of the archive’s mediations, which occur not only in its descriptive materials, but also in its infrastructures’ operational dimension.
5. Narrative possibilities on the interface
The DRI collection interface is the primary place where an ordinary user might access the collection and is also the place where the effects of the collection’s organisational logic are felt and made manifest. ARH is one of many projects hosted by the DRI, and all of these share the same basic interface, appearing to the user as a grid of thumbnails on a white background. Subcollections in the DRI are not sequestered from one another and, depending on the search criteria determined by the user, objects from entirely disparate collections appear beside, above and below each other on the collection interface. The interface, then, is a point of convergence for collection objects, where emergent collection narratives become visible, and a key site of the production of meaning and relationships between collection objects, depositing institutions and users. I start from the assumption that the assemblies generated on the interface are in some way revelatory of an organisational, or curatorial, politics: which recombinations and juxtapositions are possible, and which are prohibited? Kyle Parry describes an assembly as “any combination of expressive elements that maintains and seizes on the appearance of selection or arrangement” to generate expressive effects (Parry Reference Parry2022, p. 3); following Parry’s definition, I am interested in the effects generated specifically by the assembled qualities of the collection interface, those meanings that can only emerge from the collection by how it happens to be assembled, even when those assemblies seem to appear purely by chance. Crucial to Parry’s definition is the idea that assembly is somehow connected to protest: he looks for “terms and taxonomies for artistic, activist, and everyday media practices whose lifeblood is expressive gathering” (Parry Reference Parry2022, p.2); in the context of these collections, assembly becomes a means by which the lively objects that once channelled the collective expression of the referendum campaign continue to act with lively agency in the digital repository.
The idea of assembly echoes in the collection interface’s approach to the presentation of its materials. The DRI user interface was designed with a heavy emphasis on the ability of users to search, following a principle of human computer interaction design known as “search everywhere” (Grant et al. Reference Grant, O’Neill and Webb2015, p.52). The search bar is placed at the top-left corner of every page on the website, in a position that suggests universality, appearing physically above the specifics of the collection being searched and equivalent to the Repository’s logo and social media links in importance. Searching rewards familiarity with DC: although a good set of results can be obtained using the basic search bar, advanced search affords the opportunity to search by DC element name, enabling the user to seek exact matches by Title, Subject term, Place and so on.
Search results outputs are organised on three levels: Collections, Sub-Collections and Objects; within each of these, the user sees their results as thumbnail images within a grid-like layout. Collections and Sub-Collections correspond most often to donor organisations and contain pre-assembled sets of objects whose relation to one another is obvious and well-defined. I want to suggest that the Object search function, which outputs matches in a Collection-agnostic manner, is a key process in the assembly at work in this database. Using search to output Objects – particularly Boolean search, which allows connections and dissimilarities to multiply and overlap – recombines the DRI’s constituents in a novel fashion, creating new pools of objects from which to draw, and allowing surprising relations to appear to the user and take root between objects. With its emphasis on this type of search, the DRI seems to encourage objects breaching the borders of their collections; after a search is completed, the outputs can be analysed in exactly the same manner as official Collections and Sub-Collections, investigated as individual objects or made to relate to one another on a map or timeline.
New narrative expressions can be found through happy accidents of assembly: a search in the DRI-wide interface for my own homeplace of County Wexford surfaces local Repeal campaign materials alongside archival objects from Wexford’s maritime histories. These include artistic representations of the region’s boats – migratory settler boats like Viking longships, mailboats that traded between Rosslare and Wales during the 20th century, fishing boats that sustained local livelihoods – but also an account of the tragic deaths in 2001 of eight Turkish and Syrian refugees whose bodies were found in a container near Rosslare Harbour. Here is an arresting maritime context for the Eighth Amendment, where abortion travel from Ireland’s ports is one of many types of trans-border movement, and where the material force of the Irish Sea is a vector for the symbolic violence of Ireland’s border regime. The sea crossing out of Rosslare is represented in the DRI as a place where natural and structural forces collide with individual desires and autonomies: the sea is a route to a new life at the same time as it delivers tragedy, and the border both confers and denies the privileges of citizenry. In this assembly, abortion travel is just one of the lenses through which this aspect of the sea crossing might be experienced – and the sea crossing, simultaneously, is one lens through which to look at abortion.
Generative and thought-provoking juxtapositions like this one happen because of the DRI’s metadata, which provides disparate objects with tags or subject terms in common, tying them to a place, a theme, or a moment in time. Metadata, the user interface, and the timeline and map visualisation options native to the DRI work in parallel to produce relations between objects, create collection histories through the device of the timeline, and situate those narratives in space via the map. Importantly, the production of metadata is a communal practice within ARH: depositing organisations to the DRI are expected to provide their own metadata, and the project offered training to volunteers from those organisations and produced extensive guidance on making metadata compliant with DC and the Library of Congress vocabulary. Because of this, metadata acts as a form of careful, curatorial attention that equalises collections and dispels hierarchies and distinctions, such as the distinction in ARH between the project’s core and secondary collections. Moreover, trust is a guiding relation in a digital repository – trustworthiness, performed through infrastructural stability, is a condition of certification – and the creation of metadata is a means by which this relation of trust reveals itself. Trusting volunteers with metadata’s production acknowledges their expertise in the collections they steward; meanwhile, the extent of a dataset’s metadata provision functions as an indicator of its trustworthiness, playing a role in determining the relative centrality of one collection or another to the historical narrative being constructed within the DRI.
6. The use of metadata
As noted above, the metadata for ARH is expressed using the DC standard. First developed in 1995, the DC standard for metadata offers 15 core elements for the description of cross-domain resources. These elements – Title, Format, Subject and so on – provide a basic structure within which a digital resource can be described, and act as a means of standardising and unifying the description of objects across one or many datasets. DC is standard because it is simple, but also extensible: it is non-hierarchical and there is little semantic overlap between its elements, so it can be created by non-experts and is lightweight in terms of backend processing, yet it can accommodate a great deal of nuance in its descriptive capacities. In the case of ARH, this nuance is achieved through the use of controlled vocabularies, most notably the Library of Congress Subject Headings (LCSH). Controlled vocabularies provide stable terms for the description of common themes or objects, ensuring parity between the multiple references to the same concept that may exist in a dataset or collection; for example, data objects referencing abortion will all contain a DC “Subject” element that links to the LCSH concept of “abortion,” allowing that concept to be standardised across ARH and facilitating easier search. Controlled vocabularies are also used to standardise dates and places: of particular importance to the DRI is the Logainm spatial vocabulary, which provides a standard set of references to places within Ireland, often specific enough to describe individual streets.
The terms chosen to describe the objects in ARH reflect the project’s feminist curatorial approach. The project employed a digital archivist, Lorraine Grimes, who produced a uniform set of subject terms to be applied across the core collections (Grimes Reference Grimes2024). A majority of these are derived directly from the LCSH, with additional recourse to a controlled vocabulary called SKOS-HASSET, which makes the UK Data Archive’s Humanities and Social Science Electronic Thesaurus (HASSET) available as linked-data compliant terms. On the whole, the usage of subject terms in the database reveals a progressive, healthcare-oriented view of reproductive rights: some of the most commonly used terms across the collection are “women’s rights,” “health policy,” “social change,” “reproductive rights” and “maternal welfare.”
Each of these vocabularies represents a set of choices made about the depiction of the collection objects, both in its available terms and in its structure: the LCSH, for example, is a hierarchical tool, which situates each of its terms in relation to broader and narrower ideas. The LCSH term “abortion” was first introduced in 1986 and revised once in 2008; situating it alongside its related terms reveals a much more conservative interpretation of the idea than that allowed by ARH, one which supplants references to healthcare and human rights with repeated mention of foeticide and the suggestive broader term “birth control.” Indeed, the LCSH, a set of standards with origins in the late nineteenth century, has long been criticised for bias and exclusion in its vocabulary (Stone Reference Stone2000; Olson Reference Olson2002): revisions have been taking place since the 1960s and ongoing initiatives like The Cataloging Lab facilitate community submissions for new or revised subject headings, with particular emphasis on increasing and deepening the representation of African American, Latin American and Indigenous cultures (Fox et al. Reference Fox, Becknell, Gross, Gabbert-Montag, Kadifa and Telford2024).
Subject terms are critical because they display most clearly the interaction – and sometimes, the tension – between the intuition and expertise of a cataloguer and the limits and standards of controlled vocabularies, or between changing interpretations of archival objects and the “hierarchisation of ways of knowing and being, between inclusion and exclusion, diversity and uniformity” found in the catalogue (Sullivan and Middleton Reference Sullivan and Middleton2020, p. 65). Within the DRI, this is visible in the selective application of those vocabularies, and the creation of new categories where existing vocabularies are insufficient. Disordered metadata might reveal ill-fitting standards, yet the introduction of new metadata terms, unmoored from a recognised vocabulary, renders a collection less accessible and less interoperable. Indeed, the contrast between ARH’s feminist approach to the depiction of abortion and the regressive ideas found in the LCSH indicates a point of disjuncture between the discursive and the infrastructural, a moment where the descriptive language of metadata does not appear to be doing the same job as its computational backbone. Metadata links related ideas together and makes collections available for connection and comparison with other resources that deal with similar topics; arguably, the specific phrasing of a subject term matters less than how well it fulfils this function. However, the pursuit of an overhaul of outdated subject terms is worthwhile: better phrasing is not merely representational, but begets better connections by maintaining a standard’s ability to describe the latest subject literatures.
ARH’s documentation does not explicitly mention using critical approaches to its collection metadata, but through the project’s reliance and emphasis on community archiving, including metadata production, a counternarrative emerges wherein metadata is experienced as a felt description, whose reference to standards, where evident, is subversive of their intended use. Subject terms from the LCSH are reinterpreted by their deployment in reference to collection objects: for example, the LCSH term “personal narratives” appears very frequently in ARH, even though the LCSH documentation permits the term to exist only in combination with a specific event.Footnote 4 The LCSH lacks a subject term that refers to Irish abortion policy, though a bespoke term was created for the project that replicates the LCSH’s hierarchical structure: “Constitutional law–Ireland–The Eighth Amendment” (Archiving Reproductive Health Project at the Digital Repository of Ireland 2023b). “Personal narratives,” however, still stands alone, untethered from a specific event, acknowledging at once the centrality of personal testimony as a campaigning tool and the complexity of feeling in what is narrated.
This is compounded by the effects of one of the key critical-methodological interventions of the project: the decision to derive additional metadata subject terms directly from the In Her Shoes accounts by using sociological coding methods to isolate common topics from the posts. These methods are ordinarily used among social science researchers processing interview data, as a means of isolating frequently occurring terms and concepts in textual data. ARH researchers developed a codebook that mapped terms from the In Her Shoes dataset to LCSH subject terms, which were then inserted into the DRI metadata (Archiving Reproductive Health Project at the Digital Repository of Ireland 2023b). The terms in the codebook are emotive, and as descriptors from a collective subjectivity of those Irish women who experienced crisis pregnancy during the regime of the Eighth Amendment, they are inevitably difficult and painful, speaking to experiences of hurt and trauma that emerge not from the fact of abortion, but specifically from its prohibition. Following the example of the In Her Shoes project itself, the new metadata codebook insists on the humanity of its subjects and uses their own words to resist easy assumptions about their circumstances: terms like “contraception failure,” “emotional abuse” and “foetal abnormality” broaden the spectrum of permissible reasons to seek abortion, while “kindness from a stranger,” “in pain during travel” and “seeing other Irish women in Clinics” provide a vocabulary for the complexity of the experience of travel, which could be simultaneously deeply lonely and unexpectedly communal.
This is a profoundly unusual approach to the creation of metadata, one that seems to operationalise a conflation of computational and sociological “code” as a means of self-justification. I argue that it is, however, a successful example of a feminist data politics, intervening in a place of record to ensure the inclusion of experiences of abortion that exceed a legalistic or health-based framework, and which emerge directly from the accounts of the women involved. Here is one of Ann Cvetkovich’s “new genres of expression” (Cvetkovich Reference Cvetkovich2003, p. 7), a moment where trauma’s reconfiguration of the conventional documentary form embeds itself in archival infrastructure to create something new. The lesbian archives Cvetkovich writes about are idiosyncratic – queer – partly because they shoulder a burden of proof, addressing the “historiographical challenge” of documenting past lesbian sexualities in order to demonstrate lesbianism’s present and historical existence (Cvetkovich Reference Cvetkovich2003, p. 242). ARH takes on a similar burden in its need to preserve that which proves abortion’s normalcy; this idea represents an important discursive turn in the “Yes” campaigns, but also a central truth of modern Irish womanhood, at risk of being written out of a history of abortion that speaks only in terms of prohibition and legalisation. While Cvetkovich’s archives turn to ephemera as a means of fulfilling this evidentiary burden, it is done most powerfully in ARH through metadata, whose aggregating properties are subverted to produce a collective out of individual narratives, forming an undercommons of the catalogue visible through its infrastructures (Odumosu Reference Odumosu2020). Injecting inductive sociological codes directly into archival metadata means that collective emerges from the bottom up and can then spread beyond the In Her Shoes narratives into the rest of the campaign collections.
7. Ambivalent access
I want to deepen this analysis of the DRI’s mediation of In Her Shoes by turning away from descriptive metadata and towards a discussion of the process of translation between Facebook and the digital repository that enables the In Her Shoes narratives to appear in the project collections at all. Following discussion between representatives of ARH and the volunteers who set up and maintained the In Her Shoes Facebook page, the posts on the page were exported in bulk and provided to the DRI in JSON format in 2018, and again in 2022 when missing posts and timestamps were discovered in the original export (Archiving Reproductive Health Project at the Digital Repository of Ireland 2023a). Each story on the page was then extracted as a .txt file and metadata was created with the help of a Ruby script, before bulk ingest of both the posts and their attached metadata into the DRI. The ARH final report makes clear that the DRI’s presentation of the posts needed to be formally distinct from Facebook’s: Facebook’s UI is proprietary and cannot be reproduced elsewhere, and the complexity and impenetrability of its code make it incompatible with the FAIR principles and the DRI’s commitment to open access preservation. This is the reason for the reliance on .txt files for preserving In Her Shoes in the DRI: they divorce the posts’ content from the infrastructure that surrounds it, retaining the stories but jettisoning the platform that housed them. One unfortunate side effect of this process is the loss of the interactive quality of the original posts, which would have had “reactions” and comments attached, and which would also have been shared by readers onto their own profiles. There is also a significant privacy concern involved in the preservation of In Her Shoes, whose contributors had consented to have their stories shared on Facebook, but had not necessarily consented to their subsequent archiving by an external organisation (Archiving Reproductive Health Project at the Digital Repository of Ireland 2022); for this reason, and because of the sensitive nature of the stories’ content, access to the In Her Shoes posts via the DRI is controlled, with researchers required to log in using credentials obtained from the overseeing archivists.
Here is a curious matrix of intersecting issues: personal testimony comes into contact first with the mediating infrastructure of a particular social media platform; that testimony and the conversations it triggers become the bedrock of a campaign that not only overturns an unjust law, but transforms Irish feminism in ways that are still being negotiated; yet the preservation of personal testimony is complicated by competition between incentives towards compassion and privacy, and the mandate of the preserving organisation towards open access. Although the reason given for the decision to redact the In Her Shoes posts in the collecting statement has to do with both copyright ownership and data protection regulations, I read it as an act of care on the part of ARH researchers towards both the people who contributed stories to the project and the public who might encounter them for the first time through the DRI. Specifically, I understand the use of a digital gateway to control access to the collection as an act of care articulated through database infrastructure, where the requirement to log in in order to read the stories is a computational surrogate for the human process of bestowing trust on another individual.
This is significant not only because it represents an approach to the future use of the stories that remains accountable to their sensitive nature, but also because of the types of uses it prohibits. It is not possible, for example, to use text mining approaches to analyse the In Her Shoes files without seeking the permission of the DRI; this refusal might prompt contemplation of the kinds of discursive violence such an approach might enact on the stories they contain. Indeed, the sharing of the original posts on Facebook can be read as an act of – often necessary, often cathartic – violence, whereby the end of the regime of abortion prohibition could be brought about only by the unearthing of women’s most personal experiences; within this reading, the restoration of a measure of privacy through the restriction of access is a rehabilitative step. Although the stories’ movement from Facebook to the DRI is on the surface a movement from a proprietary, copyright-protected website to a repository that is, by definition, open source, the measures of care taken by ARH researchers, which manifest through the project’s infrastructure, prevent unfettered access and protect contributors’ interests in ways that Facebook does not.
Methods for encoding access restrictions in digital collections based on social justice imperatives have become prominent topics of digital archival scholarship since the early 2010s. Elizabeth Povinelli (Reference Povinelli2011) describes a series of Australian Indigenous projects that take advantage of the formal qualities of digital media to allow a digital archive to reproduce the social norms that would govern access to information within the Indigenous societies being archived: these replicate processes of kinship within those societies by restricting access to digital materials based on the user’s status as insider or outsider, use algorithms to subvert the assumption of total archival knowledge by limiting access for a given user to an arbitrarily-chosen sample of the whole, and use the geolocation features of modern smartphones to restrict access based on a user’s physical location. Geismar and Mohns (Reference Geismar and Mohns2011) take similar steps in their work at the Vanuatu Cultural Centre, embracing the overlap between data science and anthropology suggested by the word “relational” to build database infrastructures that reproduce the social conditions that would govern access to cultural materials in Vanuatu. Odumosu, writing about the archival remediation of a photograph of a crying child on St. Croix, also engages the possibility of repairing the harms of digital circulation by using the same technologies to impose restriction on that circulation, specifically arguing for the development of “an ethics of care in the open commons” that highlights and challenges the invasive, destructive capacity of open access collections (Odumosu Reference Odumosu2020, S299).
ARH’s mediation of In Her Shoes reflects similar practices of infrastructural, ethnographically attuned care. Outside of the context of Facebook and the referendum, the collective unburdening of shame afforded by the original sharing of the In Her Shoes stories can no longer exist; their remediation without the explicit consent of their authors into the official historical narrative via the state-adjacent, fully open access DRI risks reinscribing a politics of surveillance over women’s bodies, and a lack of ownership over their stories, that belong to the sociopolitical context of the Eighth Amendment. The decision to close access to those stories models a postcolonial feminist curatorial politics, which attends to the specific character of the labour performed by the stories in the public domain: this is an emotional labour, but also the labour of constructing a new Irish female subjectivity that would replace the matrix of shame and virtue that were central tools of the postcolonial biopolitics of the 20th century. It also echoes the forms of care that were commonplace in Ireland during the era of mother and baby homes and Magdalene laundries, whereby hiding a pregnancy from view was not only a protective act, but also a necessity of survival for many women, one which relied on the solidarity of friends and family. At the same time, however, restricting access to the stories necessitates their gatekeeping by DRI staff whose sympathies cannot be guaranteed; digital infrastructure cannot subvert this necessity, only reinforce the decisions made by those staff. Moreover, making access conditional on the doing of research engages a politics of expertise that extends the reach of the university over a movement that did not emerge from within academia, but was authentically grassroots; it locks a majority of campaigners out of engaging with a heritage that ought to be shared.
8. Savita Halappanavar and complex relations of care
The vision of care presented in the mediation of In Her Shoes is complicated by its relation to the question of whose feminism the archive privileges. I read the re-emphasis on care in ARH’s handling of In Her Shoes as an indication of a feminism expanding beyond an anarchic mode of organising to embrace principles of solidarity. However, attention to care in one part of the ARH collection does not necessarily translate across the rest of the database, which often archives campaign narratives without attending – infrastructurally or otherwise – to the problematic discourses the act of archiving perpetuates. This is most visible in those parts of the database that depict Savita Halappanavar.
Throughout the campaign, and in its immediate aftermath, many feminists worked to problematise the “Yes” campaign’s appropriative use of Savita Halappanavar’s image as the face of a largely white, settled movement (Chakravarty et al. Reference Chakravarty, Feldman and Penney2020). Ruth Fletcher echoes this criticism in her description of “the way in which a brown woman’s dead body became a site for repeal grief, when brown women’s lives were not visibly front and centre” in the movement (Fletcher Reference Fletcher2018, p. 241); more shameful, Fletcher argues, is the fact that the focus on the necropolitics of Savita’s death drew away from the analysis advanced by groups like Migrants and Ethnic Minorities for Reproductive Justice (MERJ) that demonstrated the overall poorer outcomes for migrant women in reproductive care in Ireland. MERJ member Paola Rivetti interrogates the class dimension at work in Savita’s story, contrasting her situation as a middle-class, married professional with that of Ms. Y, an asylum seeker from sub-Saharan Africa, who became pregnant by rape and was denied an abortion in 2014: “while Savita Halappanavar became highly visible as an innocent woman victim of the unjust law, Ms. Y was made both ‘hypervisible and invisible’: hypervisible to the state because of her asylum status and economic dependency, but invisible in the public discourse and sphere” (Rivetti Reference Rivetti2019, p. 184). At the same time as the preventable tragedy of her death led to mobilisation on a massive scale against the Eighth Amendment, the centrality of Savita’s image to the campaign made visible its failures of representation; for this reason, the position of many Irish feminists on the use of her image remains one of profound ambivalence.
Among the collections held by the DRI under the ARH banner is the Savita Halappanavar Memorial Collection, which was deposited by Dublin City Library and Archive (DCLA), a DRI member organisation. The collection contains almost 200 photographs of a mural showing Savita’s face, painted on the wall beside the Bernard Shaw pub on South Richmond Street in Dublin. The mural, by the artist Aches, was painted on the 24th of May 2018, the eve of the referendum, and removed a week later. During its brief tenure, it became an important site for celebration and mourning after the campaign’s end, eventually being covered in sticky notes bearing messages from voters addressed to Savita and to one another. The DCLA commissioned photographer Alastair Smeaton to preserve as many of the 1200-plus notes as possible; the collection held in the DRI is the result of his efforts.
The memorial brought together a polyphony of responses to the repeal of the Eighth Amendment; it was a tribute that centred the symbolic role of storytelling and emphasised the loss and tragedy that had resulted from the Amendment, but it also reinscribed the ambivalent visibility of the figure of Savita Halappanavar (Fletcher Reference Fletcher2018; Lentin Reference Lentin, Quilty, Kennedy and Conlon2015; Rivetti Reference Rivetti2019). MERJ member Emily Waszak writes, in a since-deleted post on Twitter/X, of her discomfort with the mural: “I love the Savita memorial for so many reasons, but it is also a visceral reminder – the physical manifestation of white tears covering over an image of a brown migrant woman and her pain” (@waszaaaaak 30 May 2018). The DRI’s representation of the photographs seems to reinforce this idea: because of their focus on capturing the writing on the note cards, most of them contain only a partial image of Savita, or no image of her at all. Most of the cards themselves are addressed directly to Savita, and most of them begin with apologies or the words “never again,” rendered on officially branded Together for Yes stationery. Because so much of the violence done to Savita by the campaign centred on a single image of her face, one must wonder at the politics of creating a memorial to her that not only uncritically reuses that image, but also repeatedly obscures it with the messages of those who are complicit in that violence. Further, there is a problem of aggregation at work in the DRI’s collections, which means that the photographs are never singular, but are always available for recombination as part of a collection, or a tag, or a timeline; because they are tagged with Savita’s name within the DRI’s metadata, photographs of other people’s words dominate any search for her in the interface, actively obscuring other documents in the repository – still scant in number – that provide deeper context to her life and death.
What care is visible in ARH’s handling of Savita’s image and legacy? In contrast to the In Her Shoes stories, there are no limitations to access here, and no self-care protocol directed at the researchers who will use the collection. Moreover, the metadata subject headings attached to the Memorial collection – which refer to reproductive rights and social change, but not to migration or racial justice – indicate a lack of attention to these latter ideas in the database’s curatorial subjectivity. Taken together, the treatment of In Her Shoes and the images of Savita reveal the limits of the database’s ability to structure care: as an aggregative, expansive technology it prioritises maximums of storage, access and use over the nuances of representation. The hollow references to Savita Halappanavar are a significant gap in the ARH collection; the database emphasises gaps like this one rather than repairing them. Similarly, while the database is used as a means of enacting care via the restriction of access in relation to In Her Shoes, its extension of care simultaneously prevents the ordinary use of the collection. The mechanisms by which care is enacted are the same mechanisms that would sequester the collection out of sight entirely; the distinction in their use has much more to do with the politics of the project’s curators than the specifics of the database infrastructure.
9. Conclusion
The work of creating equitable conditions of access to abortion in Ireland did not finish in 2018. The legislation that replaced the Eighth Amendment permits abortion up to twelve weeks, or later in exceptional cases, but retains a right of conscientious objection for doctors as well as a mandatory three-day waiting period between appointments. Fiona de Londras (Reference de Londras2020) argues that this exposes pregnant people to significant dignitary harms, removes decisional insecurity, and fails to render them full rights-holders in a constitutional sense. Abortion rights in Northern Ireland have also come about more slowly than in the Republic, or indeed the rest of the United Kingdom; abortion was decriminalised in the North in 2019 and provision began in 2020, but collapses in funding and a lack of provision for protest exclusion zones around clinics means that barriers to access persist (Morgan et al. Reference Morgan, McLaughlin, Hunter, Kavanagh, Kirk, Bloomer and Campbell2022). Issues of structural discrimination against Ireland’s migrant and Traveller populations, as well as socioeconomic limitations, also still circumscribe access in the Republic (Side Reference Side2020). Despite some fracturing in Ireland’s feminist movements in the immediate aftermath of the referendum, organisations like the Abortion Support Network continue their work, and have begun campaigns in solidarity with countries like Poland and Malta, where abortion remains illegal. The fight is manifestly not over.
The ARH database offers, among other things, a set of blueprints for how struggles for reproductive rights might be won. It holds documentary evidence from official campaigning bodies and peripheral activist movements that are of real use to feminists building similar movements in other countries or contexts. It also makes up a primary historical source through which future Irish publics will interpret the referendum campaign; within the context of the Decade of Centenaries, the import of such sources to Irish identity construction becomes clearly visible.
My contention in this article has been that the infrastructure of the collections database is an important constituent in how the stories contained in the ARH collections are told. Its mediation of the collection objects produces new systems of meaning and relation, building narratives in which the technological is unavoidably imbricated. Moreover, I have shown that database infrastructures are themselves contingent, a site where the computational is interwoven with tensions that are contextually specific to the project of Irish feminist activism. By offering this richly complicated picture of an archival database at a particular historical moment, this article ought additionally to complicate the relation between the archive and AI, destabilising the assumption of infrastructural neutrality that undergirds AI’s archival imaginaries and raising critical questions about what the introduction of AI to the archival milieu can and will do. I have argued that, particularly at the present moment of rapid expansion in AI development, while AI’s application to the archive is still uncertain, a broadening in our conception of the mediating effect of present archival database technologies, and an understanding of the ways in which those infrastructures are integral to a particular collection’s activist politics and historiographical function, are necessary critical steps in preparing for what is to come.
Acknowledgements
I would firstly like to thank all those at the DRI who provided me with access to materials and data during the preparation of this article. I am grateful to Clare Lanigan and Brenda Malone for speaking with me about the process of collecting materials after the referendum at their respective institutions. I would also like to extend my gratitude to Caroline Bassett and Clair Wills, whose generous feedback was instrumental in shaping this paper, and to the peer reviewers for their insightful comments.
Funding statement
This work was undertaken during my PhD, which was funded by a Centenary Studentship awarded by the Faculty of English at the University of Cambridge.
Competing interests
None
Orla Delaney is a critical digital humanities researcher and technical practitioner specialising in the infrastructures that shape our understanding of cultural heritage collections. Her research investigates how digital forms like code, databases, and metadata standards influence the production of meaning in collections, across contexts including UK-based museums pursuing decolonisation and activist archives in Ireland. She completed a PhD in English in 2025 at the University of Cambridge, where she is currently a Teaching Associate, and has previously been a research fellow at the National Gallery in London, UK.
Open access
