Ní Saoirse go Saoirse na mBan: Gender and the Irish language in the linguistic landscape of Ireland's 2018 abortion referendum

Abstract

In a 2018 referendum, the Irish public voted to lift the Irish state's near-total constitutional ban on abortion, bucking a recent global trend towards restrictions on reproductive rights. While abortion rights have long been a major concern of Irish feminists, appeals to national identity have often been viewed with suspicion by the women's rights movement in Ireland due to the historic role of national identity construction in perpetuating gender-based inequalities. This article explores the way(s) in which discourses of Irish identity and gender were mediated by the use of Irish in the linguistic landscape (LL) at the time of the vote. Proposing a modified version of Du Bois’ (2007) stance triangle, I argue that signs use Irish as both a means of stancetaking and as an object of stance in itself, thus effectively taking a stance on both the referendum and on Irish national identity, indexed by the language. (Stancetaking, Irish (language), national identity, gender, abortion, Eighth Amendment)

INTRODUCTION

In 2018, the Irish electorate passed a referendum proposal to repeal the Eighth Amendment of the Constitution, thereby lifting Ireland's near-total constitutional ban on abortion. The result also bucked a recent global trend of increased opposition to and restrictions on abortion, which has been accompanied by a concomitant surge in nationalism. While abortion rights have long been a major concern of the women's rights movement in Ireland, given the (historic and ongoing) treatment of women in Ireland and the male-dominated project of Irish national identity construction (L. Smyth 2005; Fischer 2019), appeals to national identity have often been viewed with suspicion, if not outright hostility, by those advocating for women's rights in Ireland (Meaney 1993). Moreover, the fact that more than 170,000 women and pregnant people have travelled to Britain for abortions over the last four decades brings the geopolitical relevance of abortion into focus, particularly as Irish national identity has largely been defined in opposition to coloniser Britain (L. Smyth 2005).¹ It might therefore come as a surprise that the push for legalisation of abortion would be framed in nationalist terms through the use of Irish language slogans, such as Ní Saoirse go Saoirse na mBan ‘There is no freedom until women are free’ or even Tiocfaidh ár Mná ‘our women will come’, an adaptation of the Irish Republican slogan Tiocfaidh ár Lá ‘our day will come’.

This article explores how discourses of Irish identity and gender were mediated by the use of Irish in the linguistic landscape (LL) during the referendum campaign. LL research ‘attempts to understand the motives, uses, ideologies, language varieties and contestations of multiple forms of “languages” as they are displayed in public spaces’ (Gorter 2019). While Ireland has been a notable site of LL research, with a particular focus on the Irish language, gender and sexuality have not been addressed in LL research specifically in this context. In order to theorise the relationship between national identity and gender, I propose an approach based on stancetaking—in its simplest terms, ‘taking up a position with respect to the form or the content of one's utterance’ (Jaffe 2009:3)—which seeks to integrate the two dimensions of nation and gender into one model. This article therefore makes an empirical argument regarding the use of Irish in the context of the referendum as well as a theoretical argument regarding stancetaking.

First, I present brief overviews of both the issue of abortion in Ireland and the Irish language, before outlining the data and methods underpinning the study. I then analyse interviewees’ comments, which is used to guide a distributional analysis of the LL data, before moving on to a qualitative analysis of the data. Using a modified version of Du Bois’ (2007) stance triangle, I argue that signs use Irish as a way to create common ground with readers: a shared (positive) stance towards Irish mediates the alignment between sign and reader on the basis of stance towards abortion; in other words, solidarity on the grounds of shared national identity or appreciation for Irish (even if this appreciation is not matched by high levels of competence in the language) becomes the basis for agreeing on the relationship between national identity and gender. I also claim that, even if Irish use in the referendum context may have served to re-define the relationship between the Irish language and gender by disrupting language ideologies which previously framed Irish as incompatible with gender equality, it fails to challenge the national(ist) project itself.

THE EIGHTH AMENDMENT AND ABORTION IN IRELAND

Abortion has been criminalised in Ireland since the Offences Against the Person Act (1861). Following a 1983 referendum motivated by fears of abortion being introduced either through legislation or court ruling, an amendment acknowledging the equal right to life of ‘the unborn’ and ‘the mother’ was added to the Constitution of Ireland Article (Article 40.3.3: Constitution of Ireland 1937; amended 1983). In practice, Article 40.3.3°, or the ‘Eighth Amendment’, made for an almost total ban on abortion in the state. More than 170,000 women and pregnant people travelled to England between 1980 and 2018 to obtain a termination,² with 2,879 doing so in 2018 alone (Department of Health & Social Care 2019). Thousands more ordered safe but illegal abortion pills online each year (Aiken, Digol, Trussell, & Gomperts 2017). In 2012, seventeen-week pregnant Savita Halappanavar died from septicaemia at University Hospital Galway after being denied a termination; her death can be viewed as a turning point in the push for legalisation of abortion in Ireland, as she became a symbol which galvanised the abortion rights movement. In response to Halappanavar's death, legislation was passed to allow for terminations to be carried out in very limited circumstances (Protection of Life During Pregnancy Act 2013), yet the constitutional question remained unresolved and the movement seeking repeal of the Eighth Amendment gained momentum.

Following several years of growing political pressure, in September 2017 Taoiseach (Prime Minister) Leo Varadkar announced that a referendum on the issue was planned for May or June the following year (RTÉ 2017). The date of the vote was set as the 25 May 2018, with early March to late May 2018 seeing the most intense period of campaigning. The final result of the referendum was a decisive 66.4% in favour of repealing the Eighth Amendment (‘Yes’), 33.6% against (‘No’) (Referendum Commission 2018). Abortion services became available in Ireland from January 2019, yet significant numbers of women and pregnant people continue to travel to Britain for abortions due to ongoing issues with accessing these services (Hickey 2019; O'Regan 2019).

The issue of women travelling abroad for abortions—and specifically to Britain—brings into focus the relationship between abortion, gender, and nation. The issue of abortion has loomed large in the Irish national consciousness, to the extent that Lentin (2013:230) has claimed that ‘abortion is central to the construction of contemporary Irish identities’. Following the War of Independence (1919–1921), the early years of the twenty-six-county Irish state were characterised by the construction of an Irish identity distinct from Ireland's colonial oppressor, Britain, with a particular emphasis on traditional social values and Catholicism (as distinct from Protestant Britain). This identity was rooted in superior purity and virtue, ‘essentially a sexual purity enacted and problematized through women's bodies’ (Fischer 2016:822). The relationship between women and the Irish state has historically been a fraught one, including incarceration and abuse in both Magdalene Laundries and Mother and Baby Homes (McAleese 2013; Commission of Investigation into Mother and Baby Homes 2021), as well as controversies surrounding women's healthcare, particularly the 2018 Cervical Check scandal (Towey 2018).

Yet abortion—particularly opposition to abortion—is more commonly linked to religion than national identity (L. Smyth 2005). A focus on national identity and its relevance to abortion rights in this article is motivated by several factors. Only 12% of voters claimed religious views to have influenced their vote in the referendum, and in fact 68% of ‘Yes’ voters identified as Catholic (RTÉ & Behaviour & Attitudes Exit Poll 2018). Kozlowska, Béland, & Lecours (2016) have argued that Catholicism is important to abortion restrictions insofar as it defines national identity; in jurisdictions where Catholicism is not crucial to national identity (e.g. Spain, Quebec), abortion access has been less restricted than in those where it is a pillar of national identity (e.g. Ireland, Poland), suggesting that national identity may be the key differentiating factor. At the time of the referendum, the recent political upheavals in the United Kingdom (Brexit) and the United States (Trump) created a keen awareness of a global trend towards a more nationalist and conservative politics, including opposition to abortion. Given the geopolitical relevance of the issue of abortion, a focus on national identity rather than religious identity can help situate the Eighth Amendment referendum campaign in a broader, global political context.

A critical evaluation of how state-building or national-identity construction is implicated in the historical (mis)treatment of Irish women might be expected to align with attitudes towards abortion, or the Eighth Amendment. Muldowney (2013) notes the characterisation of abortion as foreign or anti-Irish during the 1983 referendum campaign, including ‘repeated suggestions that Irish women were being led astray by sinister forces from outside the country’ (Muldowney 2013:43). Such a view has in recent years been explicitly promoted by anti-repeal campaigners, who have claimed that ‘[i]f the 8th amendment—the pro-life clause in our Constitution—goes, there will be nothing to be proud of anymore’ (Quinn 2017:8). In contrast, Ailbhe Smyth, abortion rights campaigner and co-director of the Together for Yes 2018 referendum campaign, has previously written of her anger at Irish society for its treatment of women, stating that she ‘hold[s] no candle for Irish “exceptionalism” ’ (A. Smyth 2015:118). Arguably, it becomes less tenable to subscribe to a positive view of ‘Irishness’ if that same ‘Irishness’ has failed women, or has even been built on the back of the mistreatment of women. The debate surrounding reproductive rights is thus closely connected to different conceptions of what Ireland is, or should be, as well as the historic role of both the Irish state and national identity construction in restricting access to abortion.

AN GHAEILGE ‘THE IRISH LANGUAGE’

Any discussion of Irish national identity cannot fail to mention the Irish language. Irish is the first official language of the Republic of Ireland, with superior legal status to English (Article 8: Constitution of Ireland 1937). Yet only 1,774,437 (30.3%) of the Republic of Ireland's 4,761,865 residents declare themselves Irish speakers, with just 73,803 (1.7%) speaking the language on a daily basis outside the education system (Central Statistics Office 2016). Despite relatively low levels of usage on a day-to-day basis, Irish remains an important marker of national (and ethnic) identity, as well as political nationalism, particularly in the north of Ireland (Mac Ionnrachtaigh 2013). O'Rourke (2005:276) reports that 62% of respondents to a survey on the relationship between language and identity believed that ‘Ireland would not be Ireland without the Irish language’, although only 32% believed that ‘language is the most important part of the Irish identity’. This encapsulates the double-edged sword of Irish: occupying an ideologically powerful yet practically precarious position due to the low number of daily speakers. From Ó Tuathaigh's (2017:66) perspective, the Irish government views Irish as ‘as a significant (but not essential) marker of a distinctive Irish nationality’. This is the ideology of the cúpla focal or ‘a few words (will do)’: Irish use in public bodies in Ireland reflects a more general ideology that a limited (and perhaps symbolic or superficial) sprinkling of Irish use is sufficient, with no need to meaningfully invest in Irish language provision or widespread Irish use (Walsh 2012; Brennan & O'Rourke 2019).³ Writing in the late 1980s, Lee (1989:673) even goes so far as to claim that ‘[p]olicy for about two decades has clearly been to let the language die by stealth’.

Previous sociolinguistic research has looked at Irish through the prism of ideologies of ‘authenticity’ and ‘anonymity’ (O'Rourke & Walsh 2015; Atkinson & Kelly-Holmes 2016; O'Rourke & Brennan 2019). This research has shown that Irish is often framed as the authentic or natural language proper to the fulfilment of a traditional vision of Irish national identity, whereas English is somewhat bleached of meaning, the ‘language from nowhere’. Pujolar (2018) notes that a post-nationalist orientation to language as a commodity often co-exists with more traditional appeals to the nation-state frame, so that ‘the creation of new products and services is often done by recasting ideas and values attached to languages that derive from this nation-state paradigm’ (Pujolar 2018:501). Irish still holds a particularly important place as a signifier of Irish national identity, with the potential to elicit strong emotional responses from speakers, including shame, pride, and a sense of (national) duty (Walsh 2019b). Tovey, Hannan, & Abramson (1989) note that the language

became associated with a package of cultural and ideological elements which had to be swallowed whole: Irish music, dance, republicanism, particularistic versions of history, conservative Catholicism and general anti-Britishness. (Tovey et al. 1989:20, cited in Brennan & O'Rourke 2019:128)

Irish was therefore closely tied to the ‘hegemonic construction of Irishness as familial, Catholic, traditional, and heterosexual’ (L. Smyth 2005:47). Walsh (2019a) has examined how gay speakers of Irish negotiate the intersection of their national and sexual identities through Irish, which is particularly relevant given that ‘[q]ueer people were among those erased from the dominant version of Irishness’ (Walsh 2019a:57). A majority of those interviewed by Walsh found it difficult to reconcile their Irish-speaking and queer identities, although Walsh (2019a:78) argues that even if gay new speakers ‘do not overtly contest the historical discourses of conservatism associated with Irish, by their very existence they queer the assumption (which lingers both in queer circles and in wider society) that non-heteronormative sexuality is not compatible with speaking Irish’.

Irish feminism, too, has historically tended to be suspicious of Irish (Meaney 1993; Nolan 2007), as Meaney (1993:241) explains:

The question of Irish identity and the question of feminine identity often… have mutually exclusive answers. Moreover the political exclusion implicit in this valorisation of the Irish language is undeniable and runs the risk of a return to the same old insular Irishness.

However, views have arguably evolved since Meaney's comments almost three decades ago. It is unfair to homogenise the Irish language movement or to label all Irish speakers as conforming to a conservative ideology or social values. Many Irish speakers—particularly younger Irish speakers today—are socially and politically progressive or even radical. This is perhaps unsurprising given the radical tradition of some Irish-language activism, particularly among Northern Republicans (Mac Ionnrachtaigh 2013), as well as the language's role in the Gaelic revival and resistance to colonial linguistic and cultural hegemony in the late nineteenth and early twentieth centuries, although this impetus was arguably abandoned in the Irish state post-independence (Lee 1989; Mac Ionnrachtaigh 2013). Assertions of Irish as ‘proxy Catholicism’ (Ó Croidheáin 2006), or as part of a post-Independence ideology which ‘promoted Irish sports, Catholicism, rural society and economy, and, of course, the Irish language’ in an effort ‘to create a homogeneous citizenry’ (Watson 2002:745), are often based on the political significance invested in Irish in the early years of the Irish state, which wielded language and culture in line with its conservative, regressive, and even counter-revolutionary aims (Whelan 2004; Mac Ionnrachtaigh 2013). However, as Walsh (2019a:56) points out, ‘[a] comprehensive study of the indexing of the Irish language as conservative has yet to be conducted but the association emerges not only in attitudinal surveys but also in histories and media discourse’.

In recent years, the language's associations with certain parts of this ideological constellation have been challenged—for instance, the widely held public perception that Irish is intimately linked to Catholicism (Ó Séaghdha 2019). There have also been attempts to change traditional understandings of Irish. Walsh (2019a) and Meaney (1993) have argued that Cathal Ó Searcaigh's and Nuala Ní Dhómhnaill's Irish poetry attempted to resignify Irish by challenging the heteronormative and patriarchal associations of the language, respectively. This strategy seeks to make the Irish language, and arguably Irish national identity more broadly, seem compatible with and more inclusive of those who have been excluded from the hegemonic version of Irish identity, defined in traditional, Catholic, patriarchal, and heteronormative terms. The fact that Irish language campaign groups were set up to campaign for Yes votes in both the 2015 (marriage equality) and 2018 referenda would suggest that Irish is now, at least to some extent, compatible with a progressive social agenda.

Given its concern with space/place, LL research has been addressing geopolitical issues since the inception of the field (e.g. Landry & Bourhis 1997). Since Milani's (2014:221) critique that LL research has ‘largely ignored—erased even—the gendered and sexualized nature of public space’, gender and sexuality have started to receive more concerted attention in the field, notably the 2018 Special Issue of Linguistic Landscape focusing on gender and sexuality. This included studies addressing the intersection of discourses of motherhood and gentrification in Brooklyn, New York (Trinch & Snajdr 2018), and the commodification of women's breasts on Columbian websites advertising cosmetic surgery (Correa & Shohamy 2018). In particular, though, Milani, Levon, Gafter, & Or (2018) provided an account of two queer groups’ divergent interactions with Tel Aviv Pride, negotiating a path between gender and sexual identities, on the one hand, and national identity, on the other. I draw on these studies to combine a focus on the body, reproduction, and gender with an intersectional perspective which examines these issues in interaction with national identity.

Building on this research, as well as on previous LL work on Irish which has addressed language ideological debates over Irish place names and the commodification of Irish for both tourists and locals (Kallen 2009; Moriarty 2012, 2014, 2015; Thistlethwaite & Sebba 2015), I aim to examine the changing associations of Irish in the context of the Eighth Amendment referendum. Specifically, I am interested in what insights Irish may allow into the intersection of national identity and gender at the time of the referendum campaign, as well as how it might be possible to theorise the use of Irish in this context alongside stances towards abortion.

DATA AND METHODS

The Notes to Savita dataset is made up of 419 LL items which appeared on a wall in Dublin following the referendum vote as an ‘in memoriam’ to Savita Halappanavar (Holland 2018). The day before the vote (24 May 2018), a mural by street artist Aches depicting Halappanavar had appeared on Richmond Street South, Dublin. On the day of the vote itself (25 May 2018), members of the public started leaving flowers and sticking hand-written notes on the wall next to the mural, either addressed directly to Halappanavar or commenting in more general terms on the referendum, including the result (Halpin 2018). The majority of these handwritten notes are written on postcard-sized cards produced by the Together for Yes campaign group (see Figure 1). Images of these notes were collected and made available online by news outlet thejournal.ie.⁴

Figure 1. Notes to Savita card; Richmond Street South, Dublin.

Interviews were also carried out with members of campaign groups to benefit from the perspectives of sign-makers and those involved most actively in the referendum campaign (see Table 1). Interviewees were contacted either through campaign organisations or via recommendations from previous interviewees. Interviews were conducted in English, lasted roughly thirty minutes on average and were semi-structured. While interviewees were encouraged to expand on topics of particular interest to them, each interview covered a set list of topics (prepared in advance to cover issues of theoretical relevance) including: the use of Irish; the use of dialectal or regional linguistic features; the purpose of signage; the scope for individuals to express themselves through signage; design elements of signage; offensive or graphic signage; the use of humour (and its appropriateness); and the relative importance of different discourse frames (e.g. gender, the nation, religion, human rights). The purpose of these interviews was to uncover what was relevant for campaigners (i.e. those producing and engaging with signage), so as to use meaningful and relevant categories during analysis of the LL data collected. Interviews were transcribed using ELAN and coded in NVivo. A thematic analysis of the interview data was carried out to identify recurring themes which came up when interviewees were discussing the Irish language in the context of the campaign.

Table 1. Interview details.

The (relative) uniformity of their location, size, material, mode of inscription, and time of emplacement in the Notes to Savita dataset minimises the differences between LL items, thus allowing each note to be treated as equivalent to another. The dataset can therefore be operationalised both quantitatively and qualitatively, as advocated by Blackwood (2015). An initial distributional analysis of language choice aimed at contextualising the use of Irish in more general terms was followed up by a more in-depth qualitative analysis of stancetaking and language ideology. First, I present a thematic analysis of interviewees’ comments regarding the motivations behind and associations of Irish-language signage in the context of the referendum. Following this, I present the results of a distributional analysis carried out to investigate both the proportion of Irish in the data and the syntactic limitations on code-mixing, in order to understand whether Irish was used in more limited contexts (compared to English) and to index more specific meanings, as interviewees had suggested. These first two stages of analysis seek to demonstrate that Irish use in the referendum campaign's LL was highly marked, and therefore that the use of Irish represents an ideological rather than simply a communicative choice. On the basis of this understanding of how and why Irish was used in the data, I propose a qualitative analysis of the data which seeks to integrate language ideology into an approach based on stancetaking.

CAMPAIGNERS’ VIEWS ON THE IRISH LANGUAGE AND THE 2018 REFERENDUM

The interview analysis presented here serves as a first step toward understanding how and why sign-makers might have used Irish in the referendum. First of all, interviewees spoke about their impressions of the proportion and distribution of Irish signage throughout the referendum campaign's LL. While Ciara and Lauren pointed out that Together for Yes produced and displayed posters in Irish, this meant that, as Lauren phrased it, “you would see a few posters with, like, Tá [‘Yes’], dotted around the place”. This impression of the relative scarcity of Irish in the referendum's LL was commonplace among interviewees. In fact, Catherine noted that she “never saw any… completely Irish signs” and that “people would use the odd word [of Irish], but… unless you were a Gaeilgeoir [‘Irish speaker’], I don't think you'd do a complete sign in Irish”. These comments point to the prevalence of code-mixing. Andrew noted “flecks of Irish in things” and “little bits thrown in here and there”, while Lauren commented that “I really loved seeing the little bits… of Irish scattered through the English campaign”. This involved both the use of oft-repeated Irish words or stock phrases, as well as code-mixing, whereby the ‘odd word’ of Irish might be included in an otherwise English slogan. Lauren and Andrew both commented on the centrality of the word Tá ‘Yes’ to the Yes campaign's messaging, with Lauren noting that Tá badges were more popular than Yes badges. Kayla, a No campaigner, also noted the popularity of badges in Irish.

Interviewees commented on the frequent use of slogans, either exclusively in Irish, such as Ní Saoirse go Saoirse na mBan ‘There is no freedom until women are free’, Mná na hÉireann ‘Women of Ireland’, and Sábháil an tOchtú ‘Save the Eighth’, or combining English and Irish, for example, Stand in Awe of all Mná ‘Stand in Awe of all Women’. Lauren also brought up the concept of Tágetherness, a portmanteau of Tá ‘Yes’ and Togetherness expressing the feeling of solidarity among ‘Yes’ voters, which underscores the pervasiveness of code-mixing.⁵ These accounts of how Irish was used during the referendum campaign highlight not only that the proportion of Irish signage was limited, but when present its use manifested through the repetition of a number of stock phrases or key words (e.g. Tá). These impressions align with previous LL research which has found that Irish signage is frequently qualified by an accompanying English ‘translation’, rather than sufficing on its own (Thistlethwaite & Sebba 2015).

A second major theme in interviews was the purpose of Irish-language signage from a sign-maker's point of view. Both Andrew and Kayla linked patterns of Irish use to the position of Irish in Ireland. Andrew stated that “it was a part of the texture of it, as opposed to being ‘we need to make sure we're targeting someone who only speaks Irish’, because that's not the experience of people in Ireland”. Kayla's campaign group “decided to go with English” in order “to grab peoples’ attention”, “to reach as vast a group of people [as possible]”, and “to have more broad appeal”. On the other hand, Lauren, an Irish speaker and part of an Irish-language campaign group in 2018, argued that

it really means something to people that you—basically, that you've taken the time… to… bother translating and producing something in Irish, when you could just say, ‘Oh well, English will reach everybody.’ But Irish will mean something different and… resonate in a different way to Irish speakers.

Although Andrew and Kayla express different attitudes towards the use of Irish compared to Lauren, among interviewees there was a common orientation to English as a ‘functional’ language comprehensible to all, while Irish is invested with greater symbolic meaning. These associations correspond to Woolard's (2016) distinction between ideologies of ‘anonymous’ and ‘authentic’ languages.

If Irish is invested with greater symbolic meaning, this begs the question of which specific meaning(s). While certain associations with the Irish language may have been changed as a result of the referendum—for example, its (previous) proximity to the Church or (newfound) compatibility with pro-repeal stances—interviews evidenced its enduring association with Irish identity. Interviewees created links between the language and national identity which were specific to the context of the referendum campaign. ‘No’ campaigner Kayla noted that her group's use of Irish “was successful because it did try and… stir up that kind of patriotic, nationalistic, ‘this is our language’ [feeling]”, before going on to draw a parallel between the Eighth Amendment and the Irish language on the basis that both were things “we should be proud of”. ‘Yes’ campaigner Lauren drew a parallel between Irish and the campaign more generally, reflecting on the use of Irish during the referendum: “it almost refreshed—to me—in peoples’ memories why Irish is important, … and why it's part of our identity. … The whole thing was we were talking about our identity, who we are as a people”.

Although Karen and Lauren campaigned on opposing sides, both saw an affinity between the Irish language and Irish identity which was assumed to be natural or beyond question. As the legitimate or authentic language of the national community, significant authority was invested in the Irish language. In Elizabeth's view, the fact that the term beo gan bhreith ‘unborn’ (lit. ‘alive without birth’) was used in the wording of the Eighth Amendment itself “destroys a bit of [the ‘Yes’ side's] argument” that the fetus is not a fully-fledged, living, person because “your language, your tradition… says otherwise”. By contrast, Lauren explained that her Irish-language campaign group had to develop a dictionary for terms relating to abortion in Irish “because some of those words in the Irish language… come from a much older place, and were… laced with… opinion”, adding “that's what that means, and that's awful, and that's not what it actually is”. Interestingly, this points to the possibility of changing the language, rather than changing language ideology: the authority of the language is kept intact, instead of questioning or challenging the link between Irish and national identity, upon which its authority rests. Asked whether the referendum had changed the way people viewed Irish, Lauren answered: “my belief would be, strongly, yes, that I think it did—not that it was something brand new, that people suddenly said, ‘Oh, what's this language?’ But just that it… made people stop and reflect and think about Irish more”. By contrast, Lucy motivated her reticence to use Irish during the campaign by arguing that the Irish language is “synonymous with nationalism” and “in terms of identifying what it is to be Irish, it has always been about reproducing our language”. Lucy was the only interviewee to question this assumed relationship between language and identity, arguing that “the political moment has come where we need to separate… the state, a sense of Irishness and the Catholic Church from each other”.

DISTRIBUTION OF IRISH IN THE NOTES TO SAVITA DATASET

Interviewees’ comments suggested that while Irish was generally less visible than English in the referendum campaign's LL, it was nonetheless used in more limited ways to index specific meanings (often associated with national identity). According to interviewees’ expectations, monolingual English messages should be more numerous than Irish messages in the data (because sign-makers are generally more competent in English and thus are more likely to write a message in English—or, at least, entirely in English). From a quantitative standpoint, there was indeed minimal use of Irish in the Notes to Savita data (see Table 2).

Table 2. Notes to Savita messages and words by language.

Of the 419 Notes to Savita, 388 (92.60%) contain English only, making it by far the most common language in the dataset; only twenty-one signs (5.01%) use any Irish, and only six (1.43%) are written exclusively in Irish. If sign-makers are generally more competent in English than Irish, and thus are capable of creating significantly longer messages in English than in Irish, English messages should be not only more numerous, but also longer than Irish messages. Although Irish messages are on average shorter than English messages (11.00 vs. 13.27 words per message, respectively), this is not a statistically significant difference (p > 0.05). Additionally, bilingual messages are in fact longer than both monolingual English and Irish messages, pointing to a lack of any simple correlation between language choice and message length. However, the average number of Irish words in each bilingual message (2.47) is far lower than that of English words (12.33) to a statistically significant level (p < 0.05).⁶ Therefore, when a sign-maker writes in both English and Irish, Irish is used sparingly relative to English, which makes up the majority of the message, suggesting that Irish is used to ‘add flavour’ (i.e. language ideological use) rather than as the primary means of communicating a message.

This is supported by an analysis of code-mixing and code-switching in this dataset.⁷ It is necessary to distinguish between code-mixing, ‘cases where lexical items and grammatical features from two languages appear in one sentence’, and code-switching, ‘the rapid succession of several languages in a single speech event’ (Muysken 2000:1).⁸ Where there is code-mixing in bilingual messages, it can be categorised as ‘insertion’, whereby lexical items or constituents from one language are inserted into a structure from another (Muysken 2000). Figure 2 is an example of this; it reads: Stand in awe of all mná. I'm sorry it took so long. The only Irish word in the message is mná ‘women’. Irish noun phrases are often slotted into otherwise English syntactic structures (rather than vice-versa), with verbs or adjectives rarely used. Stock phrases and slogans are also repeated, whether in the form of code-mixing in bilingual messages (e.g. Stand in awe of all mná [‘women’] or We repealed for you, & for all mná na hÉireann [‘women of Ireland’]) or more fluent Irish use, for example, monolingual Irish messages consisting entirely of common slogans (e.g. Ní Saoirse go Saoirse na mBan ‘There is no freedom until women are free’). Although repeated slogans are also found in monolingual English messages (e.g. NEVER AGAIN), they are more common in the Irish and bilingual messages than in the English messages, where there is a much greater variety of language on display.

Figure 2. Notes to Savita card; Richmond Street South, Dublin.

In summary, the distribution of Irish in the Notes to Savita dataset aligns with interviewees’ perceptions that Irish was rarely used, and even when it was, it was rare to find signs written exclusively in Irish. Typically, there might be the ‘odd word’ of Irish in an otherwise English message. Therefore, Irish use during the referendum campaign was highly marked. This suggests that the choice of Irish was not (primarily) motivated by a desire to communicate with an Irish-reading public, but, in Andrew's words, was used as “more of an indicator, a signifier”, indexing national identity.

A STANCETAKING APPROACH TO GENDER AND LANGUAGE IDEOLOGY

On the basis of the preceding analyses, we can proceed to a more in-depth qualitative analysis of the role played by the Irish language in stancetaking. Figure 3 can be considered typical of the Notes to Savita dataset because it is written entirely in English, and thus offers a useful counterpoint before moving on to an analysis of Irish messages. Figure 3 includes several common discourses which appear in many notes, irrespective of language: We are sorry is a common sentiment expressing the Irish public's collective guilt or sorrow for what happened to Savita Halappanavar, as is the promise of Never again, that this will not happen again (because of the referendum result). More relevant to this analysis, though, is the declaration of a New Ireland as a result of the vote. This explicitly frames the referendum result as part of the history of the evolution of the national project, a point that is analysed in more detail below. Because English is unmarked in this context, references to national identity must be made explicitly (e.g. reference to a New Ireland). When Irish is used, however, it comes with added ideological ‘baggage’: unlike the ‘transparent’ functionality of English, the use of Irish can itself be interpreted as a commentary on national identity.

Figure 3. Notes to Savita card; Richmond Street South, Dublin.

I argue that an approach based on stancetaking can provide an analysis of Irish which allows for insights into the evolving relationship between Irish national identity and attitudes to abortion (and gender more broadly). I base my own approach to stance here on Du Bois’ (2007) stance triangle: a tri-partite model whereby social actors both position themselves and align with one another on the basis of their evaluation of some object of stance (see Figure 4). Du Bois’ stance triangle is modelled on a typical face-to-face conversational exchange between two interlocutors (one speaker and one listener). In applying this to LL research, it is necessary to note that ‘sign’ replaces ‘speaker’ as the active subject taking a stance, whereas the ‘reader’ replaces the ‘listener/addressee’ as the one being directly addressed by the sign (Lou 2016). This may seem on the face of it to lack the sense in which stance is ‘achieved dialogically’ (Du Bois 2007:163), that is, an active negotiation on the part of two subjects, given that a reader is often passively addressed by a sign. However, readers play an active role in constructing the meaning of each sign for themselves, as pointed out by Kress (2010:36): ‘without interpretation there is no communication; yet it is the characteristics, the shape, of the prompt, which constitute the ground on which the interpretation happens’. A sign is thus a prompt which is interpreted by a reader and conditions—but by no means entirely determines—the interpretation. A reader can also take an even more ‘hands-on’ role, for example, by defacing a sign whose stance they disagree with. The stance object in the referendum context is a particular LL item's stance on the referendum question (i.e. whether to repeal the Eighth Amendment or not); this is taken to indirectly reflect stance on the issue of abortion more generally.

Figure 4. Du Bois’ 2007 stance triangle model.

When looking at stance through the prism of language choice, the language used by a given sign plays a role in the stance act. Kiesling (2009) argues that sociolinguistic variables can be used to do stancetaking work, for instance, the use of non-standard (ING) by American frat boys to take stances predicated on ‘casual power (that is, power exercised in a seemingly effortless manner) and cool solidarity’ (Kiesling 2009:181). Likewise, Morgan (2017) shows how interviewees use dialectal features from different varieties of Albanian to take stances on a particular variety and how it compares to other varieties. These choices in the evaluative vector of stancetaking take language as both the means to express a stance and the object of the stance itself. However, Du Bois’ (2007) model, predicated on a single stance object, does not clearly account for the way in which language ideologies can be integrated into stancetaking. My argument here is that stance towards national identity mediates stance towards the referendum question (or abortion more generally). By this, I do not mean that stance towards national identity determines stance towards abortion; instead, I am interested in how national identity is leveraged by a sign to create alignment between sign and reader, as well as the role which Irish plays in this negotiation.

As seen in the preceding sections of analysis, Irish is a highly marked choice in the Notes to Savita dataset, suggesting that it is not a neutral communicative tool but a vehicle for (language) ideology. Irish is used to build solidarity between sign and reader on the basis of (supposed) shared national identity, or at least a shared appreciation of Ireland's national language. In Du Bois’ (2007) terms, then, a sign attempts to align with a sign-reader on the basis of its implicit (positive) evaluation of the Irish language, positioning itself as valuing the national identity indexed by the choice of Irish. On the basis of this alignment, a sign can then attempt to align with the reader with regard to another stance object: attitude towards the referendum question, or attitude towards abortion. We can think of stance towards national identity as mediating stance towards abortion: language choice, which indexes national identity in the case of Irish, acts as a ‘gate’ which stands between the reader and attitude towards abortion. If the reader does not align with the sign on the basis of stance towards the Irish language, this may ‘block’ or prevent alignment on the basis of attitude towards abortion. Language choice acts as both a ‘gate’ and a stance object in and of itself, meaning that it cannot be accounted for within Du Bois’ (2007) process of ‘evaluation’.

A new ‘stance triangle’ model which can account for the processes described above is therefore necessary. Figure 5 reimagines Du Bois’ (2007) stance triangle in such a way. This model represents the relationship between two stance objects, Irish identity (indexed by the Irish language) and stance on abortion (indexed by stance on the referendum question), as well as the relationship between them. The first, Irish identity, is the immediate object of stancetaking by a given sign; once there is alignment between sign and reader on the basis of stance towards the first stance object (Irish identity), this provides a basis for alignment with regard to the second stance object. The sign goes through a two-step meaning-making process, depending in the first instance on whether it is written in Irish (or not), and then in terms of stance on abortion. In practice, however, these two steps may take place almost simultaneously, considering the Irish text is both a stance-object and also the means for expressing a stance on abortion. This requires there to be at least some degree of compatibility between the two stance objects: if the Irish language is deemed (by a given reader) to be incompatible with a pro-repeal stance, for example, then the act of stancetaking may seem incongruous. In the case of Irish specifically, given strong public support for the Irish language (Darmody & Daly 2015; McMorrow 2019), it is unlikely that readers would have negative attitudes towards Irish, which would subsequently cause them to not align with a sign on this basis (rather than on the basis of a sign's position on abortion). However, the model must allow for such a possibility, especially given the arguments I make subsequently about the compatibility between stance towards Irish national identity and attitude towards abortion.

Figure 5. Adapted stance triangle model.

The Notes to Savita dataset consists of messages posted after the vote. While they are not taking a stance in the same way as campaign posters in the lead-up to the referendum, given that they are celebrating the referendum result rather than attempting to convince voters, they are taking stances in expectation that these messages will be read and will meet with either the approval or disapproval of the reader. For example, the only Irish in Figure 6 is the popular slogan NÍ SAOIRSE GO SAOIRSE NA mBAN ‘THERE IS NO FREEDOM UNTIL WOMEN ARE FREE’. The inclusion of this phrase invokes a specific understanding of the relationship between national identity and gender in Ireland: that freedom, understood as Irish national freedom, is not complete without women's freedom within that nationalist project. To express this through Irish constitutes an ideological statement that nationalist ideology is not only compatible with, but necessarily also includes, a concern for gender equality. This understanding of the relationship between national identity and gender requires a positive evaluation of both Irish identity and gender equality on the part of the reader for there to be alignment between LL item and reader. Given that Irish political nationalism has been a male-dominated project, to the extent that Banerjee (2012) has framed it as ‘muscular nationalism’ and argued that women have been marginalised within the history of Irish Republican activism, this is not a trivial stancetaking move. This stance is complicated, however, by the inclusion of I'm SO SORRY FOR WHAT OUR COUNTRY DID TO YOU at the beginning of the message, which suggests a critique of Ireland—or, at least, Ireland as it was before the 2018 referendum—which undermines the nationalist sentiment of the Irish phrase at the end of the message. These two parts of Figure 6's message are not contradictory, but conceptualise Irish national identity in a specific way: if NÍ SAOIRSE GO SAOIRSE NA mBAN acknowledges the theoretical compatibility of a positive stance towards Irish national identity and gender equality, I'm SO SORRY FOR WHAT OUR COUNTRY DID TO YOU accepts that in practice these have not been successfully integrated, as Ireland's history in relation to abortion demonstrates. This should be understood against the backdrop of the larger theoretical debate on (Irish) nationalism. While some authors argue that Irish nationalism can be emancipatory, albeit on the condition of economic, cultural, and linguistic justice (Mac Ionnrachtaigh 2013), feminist authors have tended to argue otherwise: for example, Lisa Smyth's (2005:36) argument that ‘the articulation of a “national” feminism amounts to an inability to criticize the “naturally” patriarchal character of the (naturally) familial-nation’ amounts to a flat rejection of even the theoretical compatibility of gender equality and nationalism.

Figure 6. Notes to Savita card; Richmond Street South, Dublin.

This stance is also found in Figures 7 and 8, which proclaim Éire nua ‘[a] new Ireland’.⁹ If the use of Irish here is marked, and thus constitutes an index of Irish national identity rather than an attempt to communicate with an Irish reader, then the stance being adopted here is one which considers gender equality to be possible in ‘a new Ireland’, suggesting that there is an ‘old Ireland’ which preceded the 2018 referendum but which has now been transformed by virtue of the incorporation of women's reproductive rights into the dominant version of Irishness (as demonstrated by the overwhelming ‘Yes’ vote). Like Figure 6, Figures 7 and 8 implicitly acknowledge the failings of the ‘old Ireland’ but look forward to a ‘new Ireland’ in which positive stances towards both national identity and gender equality are possible and compatible. It is intriguing that an appeal to a ‘new Ireland’ should be made through Irish, considering the language's associations with tradition and historicity (O'Rourke & Brennan 2019). This neatly encapsulates the contrasting dynamics of stasis and change in the relationship between gender and nation in the Notes to Savita: while these messages in some sense signal the maturation of a new vision of Irishness in terms of the relationship between national identity and gender, at the same time they demonstrate the enduring relevance of national identity as a frame of reference for sign-makers.

Figure 7. Notes to Savita card; Richmond Street South, Dublin.

Figure 8. Notes to Savita card; Richmond Street South, Dublin.

Relating this analysis to previous LL research highlights the dynamics at play in the Notes to Savita. Milani and colleagues’ (2018) distinction between strategies of affirmation and transformation is applicable here, too: the use of Irish in conjunction with discourses of gender equality might be seen as a strategy of affirmation which seeks to incorporate gender equality into the dominant national identity project, rather than a transformational move questioning the compatibility between national projects and gender equality more fundamentally. Trinch & Snajdr's (2018) account of gentrification in the New York LL is also relevant: while white, middle-class women were publicly defying patriarchal power through their impact on the LL, which made traditionally ‘private’ issues public, they were driving gentrification at the same time. In a similar dynamic, the incorporation of gender equality into a new vision of Irishness falls back on the enduring relevance and discursive authority of national identity rather than discarding it entirely, or decoupling Irish from it. While this analysis shares with these studies an interest in the relative valorisation of intersecting planes of identity, the specific context is also important: for example, this analysis deals with a minoritised language, rather than majority languages (in their respective contexts) such as Hebrew or English, which cannot be detached from issues of colonisation and subsequent national identity construction specific to Ireland. This analysis therefore dovetails with but can also provide an alternative perspective to this existing body of research.

CONCLUSION

By modifying Du Bois’ (2007) stance model and positing two stance-objects rather than one, it is possible to account for both the communicative and ideological work being done by the choice of a marked (and ideologically potent) variety such as Irish: language choice is not merely the neutral vehicle for the communication of a stance, but can also itself do stance work. With signs using the Irish language to celebrate the passing of the vote in its immediate aftermath, what might once have been seen as a language associated with a traditional, Catholic and patriarchal vision of Irishness now appears compatible with a progressive endorsement of gender equality. These stances resignify the Irish language and recruit national identity to the cause of repealing the Eighth Amendment; some of the Notes to Savita even explicitly acknowledge that women have not always enjoyed an equal stake in the national project.

In one sense, this is a radical rearticulation of Irish national identity, particularly bearing in mind the rapid change in social attitudes—within the space of a single generation—since 1983. From this perspective, Irish can be seen as a language re-appropriated from its use in the conservative, heteronormative project of nation-building at the outset of the state to a symbol of the more progressive twenty-first century vision of Irishness. However, this arguably does not change the terms of the debate by questioning the centrality of national identity on a more fundamental level: the link between the Irish language and Irish national identity remains intact; the difference being that now the dominant version of Irish national identity is compatible with a concern for gender equality. Whether or not this constitutes a rupture in ‘business as usual’ for Irish national identity is beyond the scope of this analysis. However, the Notes for Savita demonstrate the dynamics of stasis and change in the ideological framing of the Irish language today.