Crowd sourcing Welsh language data in the Covid era

Welsh language place names risk being forgotten as people choose to use English versions of street names, properties, towns and villages and even rename properties with English names. Many official datasets for historic buildings in Wales do not record the original Welsh names. This presentation will look at how we have adapted our projects to work online during the Pandemic, and how we are crowdsourcing Welsh names of Wales' built heritage and adding them to Wikidata. The talk will focus on the activities of our recent Wicipics project which saw the public contribute remotely, creating Welsh language data and sharing their openly licenced images of historic sites in their area. This session will also look at how we might use this data to enrich our historic record. For example, by combining with OpenStreetMap to develop a Welsh language map interface and by sharing our crowdsourced data with 3rd party websites and other Welsh heritage organisations.

Picking the crucial language technologies and how Wikipedia can help: the Welsh experience

When you're making plans to get people using your language as much and as often as possible, there's a list of things related to Wikipedia which can really help. I'll share our experience with the Welsh language. Supporting the Welsh-language Wikipedia community forms Work Package 15 of 27 in the Welsh Government's Welsh Language Technology Action Plan https://gov.wales/sites/default/files/publications/2018-12/welsh-language-technology-and-digital-media-action-plan.pdf. We like supporting Welsh language Wikipedia editing workshops, video workshops and other channels that encourage people to create and publish Welsh-language video, audio, graphic and text content because we're on a mission to try to help double daily use of Welsh by 2050. I'll share developments we're funding in speech, translation and conversational AI. The partners we're giving grants to publish what they develop under open licence. So we can share what we've funded with many companies. We think Microsoft might have used some to make their new synthetic voices in Welsh. We're excited by the potential Wikidata offers. We'll look at its potential in populating Welsh maps this year. We've already used Wikipedia search data as a way of prioritising the training of a Welsh virtual assistant. Welsh may not be spending as much as Icelandic and Estonian do on language technologies, but we'd like to share what we're learning as a smaller language about the important areas to focus on and how Wikipedia can help.

The Well-being of Students in a Welsh Secondary School Before and After a COVID-19 Lockdown

Background: There has been considerable research on the well-being of secondary school students, most of which focuses on health-related behaviour or mental health issues. The well-being process model provides a framework that examines predictors of positive and negative well-being outcomes. The model has been validated in many studies of workers and university students. The present study examined the model's applicability to secondary school students whose education is conducted through the medium of the Welsh language. COVID-19 has disrupted education and well-being, and the present study presents profiles of well-being before and after the first COVID-19 lockdown in Wales. Aims: The first aim was to examine the applicability of the well-being process model to secondary school students. A second aim was to study students where teaching was in the Welsh language. Finally, the research examined the impact of the COVID-19 lockdown and identified predictors of current well-being after lockdown and the long term implications of COVID-19. Methodology: The research was carried out with the informed consent of the volunteers and approved by the School of Psychology, Cardiff University ethics committee. The participants were 214 students (111 males), and they represented each year group. An online survey was carried out, and the pre-COVID-19 associations between the well-being process predictor variables and outcomes were examined using regression analyses. Predictors of current and long-term well-being post-lockdown were also examined. Results: The data relating to the pre-COVID 19 periods confirmed that positive well-being was predicted by high scores for psychological capital and social support. Daytime sleepiness was negatively associated with positive well-being. Stress at school was predicted by high student stressors, negative coping, social support and low psychological capital scores. Post-lockdown well-being was predicted by psychological capital and negatively associated with academic stress and fear of infection, and the stress of isolation. The longer-term negative impact of COVID-19 was predicted by problem-focused coping, fear of infection, and social isolation. Conclusion: The results confirmed the applicability of the well-being process model to Welsh secondary school students. Lockdown during COVID-19 affected well-being, with the risk of infection and the stress of isolation and academic stress being the major negative influences.

A Tale of Two Bridges: The Poetry and Politics of Infrastructure in Nineteenth-Century Wales

Drawing on Brian Larkin’s concept of ‘infrastructural poetics’, this article considers and compares a selection of English- and Welsh-language poems, by writers including Eliza Mary Hamilton, Frederick Faber, Richard Llwyd, and Eben Fardd, about two nineteenth-century infrastructures that transformed North Wales and Great Britain’s relationship to Ireland: the Menai Suspension Bridge (1826), and the Britannia Tubular Bridge (1850). I argue that these non-canonical poems complement perspectives derived from parliamentary records, official reports, technical planning documents, scientific manuals, and journalism, enhancing our understanding of the nineteenth-century infrastructural imagination. Specifically, building on the association of infrastructural development and modernity, I explore how the poems under discussion participate in nineteenth-century negotiations about Wales’s place and future in the United Kingdom, and how these negotiations evolved between 1819 and 1852. I show that, although Wales was the site of impressive engineering feats and accelerating industrial extraction, English-language poems present the Menai Bridge in picturesque terms, drawing on popular images of the Celtic fringe that evoke timeless, ideal beauty. Anglophone verse about Britannia Bridge, by contrast, focuses explicitly on the infrastructure’s technological modernity but claims it as an English landmark. Both strategies, I suggest, effect an erasure of Wales – as a distinct cultural and political entity – from a future conceived as Anglo-British. Poems written in Welsh, and the work of Welsh writers in English, complicate this picture, not because they reject British nationalism and imperialism, but because they seek to embed a modern Welsh nation more centrally within those political and ideological frameworks.

Hiraethi Jan Morris

Within transgender studies, Jan Morris casts a problematic shadow, with Aren Aizura identifying how “Morris's entire literary and historical oeuvre . . . [is] a tacit articulation of a British colonial ideology.” Yet this position appears to be based on Morris's works between the 1950s and 1970s, up to and including her memoir Conundrum, and represents arguably only the first of three periods in Morris's writing. This essay argues that two subsequent periods diversify our understanding of Morris as a complex, transcultural figure: her broadly leftist, anticolonial writing on Wales and the Welsh language (1980s–90s), and then in the twenty-first century when Morris increasingly appears to question the colonial, nationalist, and cisheteropatriarchal ideologies that have shaped her previous writing. This essay concludes that Morris's body of work provides valuable evidence as to the complex interplay of Welsh, British, and European conceptions of gender that characterize her attitude and writing on transgender identity.

Creating Welsh Language Word Embeddings

Word embeddings are representations of words in a vector space that models semantic relationships between words by means of distance and direction. In this study, we adapted two existing methods, word2vec and fastText, to automatically learn Welsh word embeddings taking into account syntactic and morphological idiosyncrasies of this language. These methods exploit the principles of distributional semantics and, therefore, require a large corpus to be trained on. However, Welsh is a minoritised language, hence significantly less Welsh language data are publicly available in comparison to English. Consequently, assembling a sufficiently large text corpus is not a straightforward endeavour. Nonetheless, we compiled a corpus of 92,963,671 words from 11 sources, which represents the largest corpus of Welsh. The relative complexity of Welsh punctuation made the tokenisation of this corpus relatively challenging as punctuation could not be used for boundary detection. We considered several tokenisation methods including one designed specifically for Welsh. To account for rich inflection, we used a method for learning word embeddings that is based on subwords and, therefore, can more effectively relate different surface forms during the training phase. We conducted both qualitative and quantitative evaluation of the resulting word embeddings, which outperformed previously described word embeddings in Welsh as part of larger study including 157 languages. Our study was the first to focus specifically on Welsh word embeddings.

English–Welsh Cross-Lingual Embeddings

Cross-lingual embeddings are vector space representations where word translations tend to be co-located. These representations enable learning transfer across languages, thus bridging the gap between data-rich languages such as English and others. In this paper, we present and evaluate a suite of cross-lingual embeddings for the English–Welsh language pair. To train the bilingual embeddings, a Welsh corpus of approximately 145 M words was combined with an English Wikipedia corpus. We used a bilingual dictionary to frame the problem of learning bilingual mappings as a supervised machine learning task, where a word vector space is first learned independently on a monolingual corpus, after which a linear alignment strategy is applied to map the monolingual embeddings to a common bilingual vector space. Two approaches were used to learn monolingual embeddings, including word2vec and fastText. Three cross-language alignment strategies were explored, including cosine similarity, inverted softmax and cross-domain similarity local scaling (CSLS). We evaluated different combinations of these approaches using two tasks, bilingual dictionary induction, and cross-lingual sentiment analysis. The best results were achieved using monolingual fastText embeddings and the CSLS metric. We also demonstrated that by including a few automatically translated training documents, the performance of a cross-lingual text classifier for Welsh can increase by approximately 20 percent points.

“It sounds like elves talking” – Polish migrants in Aberystwyth (Wales) and their impressions of the Welsh language

The purpose of this paper was to gain a better understanding of the perceptions of the Welsh language held by the Polish adult migrants in Aberystwyth, Wales. Using qualitative research methods, we collected data from participants concerning their perceptions of the sound and spelling system of Welsh. Data obtained showed that adult Poles in Aberystwyth perceive the phonetics and phonotactics of Welsh to be markedly different from that of their native Polish. The participants believed Welsh to have small number of vowels and large number of consonantal clusters. By comparing consonantal and vowel inventories we were able to demonstrate that Welsh has a more complex vowel inventory than Polish. The consonantal inventories of both languages show great similarities and should not pose major problems to Polish learners of Welsh, who are also speakers of English. As for the phonotactics, Polish possesses a far more complex inventory of consonantal clusters than Welsh. We show that claims of the study’s participants that Welsh pronunciation is markedly different from Polish is not based on the linguistic grounds. Instead, such claims must be rooted in the social and ideological perceptions of the Welsh language on the part of the participants in the study.

12. The nations and the union

This chapter explores the UK as a state which is made up of a number of diverse parts. These parts are Wales, Scotland, England, and the territory of Northern Ireland. Each part has its own characteristics which show through in the political and constitutional makeup of the UK as a whole. The chapter describes these different components. It discusses the various differences between them and looks at issues related to maintaining coherence. Using theoretical models, it analyses the nature of the UK as the state, the nation state, and the multinational state. It looks at the concepts of consociationalism, the unitary state, the union state, and federation. It provides a number of practical examples which demonstrate how these ideas operate in the real world. It also considers the Welsh language, territorial variation in the party system, the ‘English Votes for English Laws’ procedure in the UK House of Commons; and the ‘Barnett’ formular for the allocation of funding in the UK.

Community Music Wales: Training musicians for the future

This article focuses on the development of the charity Community Music Wales (CMW), which grew from a small collective of musicians in Cardiff in the 1980s to a national organization. Although comparisons can be drawn to other UK-based organizations such as Making Music UK, CMW is unique in its broad range of activities. The article outlines key milestones in the development of CMW throughout its 28 years of operation, including the introduction of its first music mentoring scheme, its community record label and its Welsh language label – Ciwdod. The article also highlights the development of community work that engages with key themes such as mental health and the environment. By engaging with company archives, the article considers quantitative data, such as the numbers of musicians who have attended training. Most importantly, it outlines how tutor training has upskilled the workforce and supported the creative economy of Wales.