South African Language Research Articles

Explainable Pre-Trained Language Models for Sentiment Analysis in Low-Resourced Languages

Sentiment analysis is a crucial tool for measuring public opinion and understanding human communication across digital social media platforms. However, due to linguistic complexities and limited data or computational resources, it is under-represented in many African languages. While state-of-the-art Afrocentric pre-trained language models (PLMs) have been developed for various natural language processing (NLP) tasks, their applications in eXplainable Artificial Intelligence (XAI) remain largely unexplored. In this study, we propose a novel approach that combines Afrocentric PLMs with XAI techniques for sentiment analysis. We demonstrate the effectiveness of incorporating attention mechanisms and visualization techniques in improving the transparency, trustworthiness, and decision-making capabilities of transformer-based models when making sentiment predictions. To validate our approach, we employ the SAfriSenti corpus, a multilingual sentiment dataset for South African under-resourced languages, and perform a series of sentiment analysis experiments. These experiments enable comprehensive evaluations, comparing the performance of Afrocentric models against mainstream PLMs. Our results show that the Afro-XLMR model outperforms all other models, achieving an average F1-score of 71.04% across five tested languages, and the lowest error rate among the evaluated models. Additionally, we enhance the interpretability and explainability of the Afro-XLMR model using Local Interpretable Model-Agnostic Explanations (LIME) and Shapley Additive Explanations (SHAP). These XAI techniques ensure that sentiment predictions are not only accurate and interpretable but also understandable, fostering trust and reliability in AI-driven NLP technologies, particularly in the context of African languages.

Code-Switching as a tactical approach to enhance learner engagement with intricate concepts in oral english presentations

The South African English language curriculum states that by the time students reach Grade 7, they should have mastered certain English language skills that enable them to communicate effectively at both personal and educational levels. Therefore, a study was conducted to investigate the potential benefits of code-switching in improving students' oral participation in English classes. The study took place in a primary school in the Malamulele East circuit and involved all fifty Grade 7 students. The students' oral presentations were part of their formative assessment. Researchers used a rigorous methodology, including convenience sampling, to select participants, with the first ten girls and the first ten boys considered for the study. Data was collected during presentations, where each student had an opportunity to present. The qualitative method was used, with audio recordings and observations as data collection tools. Content analysis was employed to analyse the data collected from the Grade 7 students. Based on the students' performance in their oral presentations, code-switching seemed to help them to effectively communicate complex concepts in their primary language when there were no equivalent English words. The findings of this study showed that code-switching enhances students' oral communication skills and makes learning more enjoyable. students achieve their communication goals through code-switching, offering the audience a hopeful outlook on the potential benefits of this approach.

Revisiting the Challenges of Teaching and Learning Indigenous South African Languages

Whereas South Africa has made significant strides after the official end of colonialism and apartheid in 1994 concerning the advancement of indigenous South African languages, there are observable challenges that demonstrate endurance even in the post-colonial and post-apartheid contexts. In particular, there are enduring challenges in the teaching and learning of these languages within the basic and higher education system. With this view in mind, the aim herein is to explore and discuss four noteworthy challenges that obstruct the optimal teaching and learning of indigenous languages in the South African context. It must be noted that this article concentrates on indigenous South African languages in general, and not specific ones. The objective is to uncover and sensitize the language and education policymakers as well as affected stakeholders such as parents, learners, and students that challenges in the context of language in education exhibit determination, hence the necessity to scholarly revisit them in a bid to solicit reasonable solutions to the conundrum. The four noteworthy challenges, which are discussed in the body of this scholarly discourse, are recognized as a source of discussion. The findings and discussion underline that challenges within the language in education in the South African context are multifaceted. One of these challenges could be pinned on colonial and apartheid legacies that appraised oppressive languages to the detriment of indigenous South African languages and dialects. In fact, the Bantu Education Act of 1953 is implicated in this article as one of the principal drivers of the subjugation of these languages. In the same vein, this article implicates some post-colonial and post-apartheid perceptions towards the subjugation of these vernacular languages. The closing remarks highlight the necessity to recurrently debate challenges of indigenous languages within the educational system.

Challenges in Intellectualizing Sesotho for Use in Academic Publications

This paper explores challenges in intellectualizing Sesotho for academic works beyond literary-based theses. It also overviews resources for the intellectualization of official Indigenous languages of South Africa, focusing on national translation projects for terminology development. Furthermore, it highlights efforts by the South African Centre for Digital Language Resources in creating and collecting corpora for all official languages, including Sesotho. Challenges in Sesotho lexicography, especially with corpus-based approaches, are discussed. Additionally, the paper reflects on the authors' experiences with publishing academic articles in Sesotho in South African journals. It discusses obstacles faced, and strategies employed, including dialogues with reviewers on creating new terms and choosing between Lesothan Sesotho and South African Sesotho orthographies. Finally, the essay addresses challenges posed by limited publication avenues for indigenous languages in southern Africa and academic pressures for citations and research impact. It aims to shed light on lesser-known Indigenous languages in Southern Africa by spotlighting challenges and solutions to academic publications in Sesotho. Overall, the paper concludes that efforts at intellectualizing Sesotho and other South African indigenous languages are evident. Even so, there is far less emphasis on using Sesotho and other indigenous languages for academic discourses.

Neoliberalism in South African higher education language policy: A decolonial perspective

The global spread and hegemony of English as a medium of teaching, learning and research has come to be closely associated with neoliberalisation of higher education in the universities across the world, and has negatively impacted on epistemic issues, especially in Global South universities. No permanent solution has been found on challenges of English hegemony, particularly in African higher education. This article adds to the advancement of neoliberal discourse in higher education in South Africa by analysing neoliberal manifestations in its higher education language policy agenda. The article used desktop analysis to assess relevant literature on mainstream neoliberal ideologies and manifestations in the language policy reform post-apartheid. Drawing on relevant literature and through critical discourse analysis, it integrated evidence from the selected sources associated with mainstream neoliberal ideologies and then analysed the neoliberal manifestations in the language policy reforms pursued in South African post-apartheid context to bring out how the neoliberal actors in the higher education sector have reconfigured neoliberal principles. The decolonial perspective premised on concepts such as coloniality of being, power and knowledge was used to critique the neoliberal manifestations in South African higher education language policy directives. The paper presents the argument that universities in the country are deeply entrenched in linguistic hierarchisation which perpetuates epistemic injustices in teaching and learning. The African higher education should endeavour to underpin its language policies in indigenous knowledge systems to help harness the African content while existing in a global space where all knowledges should be regarded as equally relevant.Contribution: This article contributes to the emerging discourse of neoliberalism in higher education in the Global South and its impact on language policy formulation and implementation.

The development of synthetic child speech in three South African languages

It is well-known that children with expressive communication difficulties have the right to communicate, but they should also have the right to do so in whichever language they choose, with a voice that closely matches their age, gender, and dialect. This study aimed to develop naturalistic synthetic child speech, matching the vocal identity of three children with expressive communication difficulties, using Tacotron 2, for three under-resourced South African languages, namely South African English (SAE), Afrikaans, and isiXhosa. Due to the scarcity of child speech corpora, 2 hours of child speech data per child was collected from three 11- to 12-year-old children. Two adult models were used to “warm start” the child speech synthesis. To determine the naturalness of the synthetic voices, 124 listeners participated in a mean opinion score survey (Likert Score) and optionally gave qualitative feedback. Despite limited training data used in this study, we successfully developed a synthesized child voice of adequate quality in each language. This study highlights that with recent technological advancements, it is possible to develop synthetic child speech that matches the vocal identity of a child with expressive communication difficulties in different under-resourced languages.

An evaluation of PanSALB’s progress in monitoring and managing language rights violations in South Africa

This article evaluates the extent to which the Pan South African Language Board (PanSALB) managed to monitor and dealt with language rights violation complaints. The board’s role is to implement the language policy and to develop the official languages. The constitution of the country recognises 12 languages as official, including South African Sign Language. English continues to dominate regardless of its equal official status with other languages. This study used a qualitative method through document analysis and observations. The findings show that PanSALB has encountered various challenges in ensuring that language policy is implemented in a manner that satisfies all language groups and ensuring that language rights are protected. The study also found that few speakers of indigenous languages lodged complaints with the board compared to the speakers of other official languages. Data also revealed that out of 245 complaints lodged between 2004 and 2013, the board resolved only 60 complaints, while 185 remained unresolved. This article makes recommendations for how PanSALB could effectively fulfil its mandate and fast-track resolutions to the complaints. The article also suggests the use of awareness campaigns to encourage speakers of other languages to lodge complaints when their language rights are violated.

Perceptions about the development of Black South African languages for use in basic education in South Africa

The development of Black South African languages (BSALs) has been a subject of discussion for indigenous languages activists. This development was deemed necessary in basic education, hence our research targeted parents and guardians of the learners. The study assessed the perceptions of parents and guardians about the development and use of BSALs in basic education. Qualitative and quantitative data were collected through a questionnaire and Likert scale. That was done to investigate the awareness of parents and guardians about their freedom to choose a language of their choice for learning and their perception and attitudes about the use and development of BSALs in basic education. The survey revealed, among other things, that parents and guardians preferred English for their children to BSALs, although they accepted that these languages are also important for identity and culture maintenance. Furthermore, the notion of monolingual education was rejected. It was also discovered that they thought mother tongue education would deprive their children of socio-economic mobility. Moreover, the respondents expressed that BSALs are not developed and promoted enough by the government. These findings led the researchers to conclude that the parents and guardians value English in education for the power, status and opportunities it presents, while BSALs should be in the system to preserve culture and identity.

Updated Morphologically Annotated Corpora for 9 South African Languages

Updated Morphologically Annotated Corpora for 9 South African Languages

Kufundziswa kwetilwimi taseNingizimu Afrika ngekusebentisa bucwepheshe

African languages are behind, especially with regards to terminology in education, and entering the digital environment. This research considers the teaching of South African languages through technology and its presence in the digital space. What can we do in this regard? First, it is important that we have to educate our teachers on how to use technology in teaching South African languages. The paper explores the potential in the use of technology and occupying the digital space available. It examines how South African languages can be integrated with other subjects like coding and robotics. In other words, teaching African languages through technology can produce learners who will be empowered with 21st-century skills and be able to use the information that will be digitally relevant. The study adopted Actor Network Theory (ANT) to direct its activities. This framework gives credit to any being or factor, irrespective of its nature, human, or non-human. ANT is a theory of the progressive constitution of a network in which both human and non-human actors assume identities. The data and information that were generated through participatory action research, a desktop review, a literature review and textual analysis were analysed using critical discourse analysis (CDA). This was used to show how discourse structures reproduce, legitimize, question, or perform power and dominance relations in society. Also, CDA is an analytical research approach that analyses speech critically. A person or group of dominant tendencies review critical discourse analysis trying to explain a social reality and have a specific goal in mind. Findings suggest that collaboration with all stakeholders in digitising and teaching through technology is possible. We also look at the contribution of the Web in promoting the use and status of African languages. Participation in sharing and producing knowledge through the Web can play a key role in the economic, social and educational development of Africa. The research will be of interest for educators, especially in higher education institutions, in teaching African languages through technology. It is important to digitise all our African languages, especially Siswati. These languages can develop and then be found in the digital space in such a way that they can be used or spoken internationally. Another opportunity can be found in the enhancement of the national corpora, the use of other software and open-source platforms like Wikipedia, WordNet, and also the use of technology to teach the South African languages. African languages can indeed be satisfactorily developed in these respects. Lugcogco Tilwimi tase-Afrika betatiwa njengetilwimi letisemuva ngetintfo letiningi, lokunye kwato kungabi nemagama (terminology) kulomkhakha wetemfundvo tibuye futsi tingabikhona etindzaweni tedijithali. Inhloso yalolucwaningo kubuka kufundzisa kwetilwimi tase-Africa ngebucwepheshe besimanje kanye nekubakhona kwato etindzaweni tedijithali. Umbuto longatibuta wona kutsi, Yini lekumele yentiwe kute tilwimi tetfu tase-Afrika tifundziswe nato ngebucwepheshe besimanje? Singatisabalalisa njani etindzaweni tedijithali? Ngalemibuto lemibili kubalulekile kutsi bothishela bafundziswe kusebentisa bucwepheshe besimanje ekufundziseni tilwimi taseAfrika. Leliphepha lihlose kuveta kutsi Tilwimi tase-Afrika tingafundziswa tisebentisa bucwepheshe besimanje kantsi futsi tibuye lisatjalaliswe etindzaweni tedijithali. Lokunye lelikubukile kutsi tilwimi tase-Afrika tingafundziswa ngekuhlanganisa letinye tifundvo letifana nabo-khodingi kanye nerobothiksi. Ngalamaye emavi tifundvo letikhicita bantfwana lebatakwati kucabanga, kucatulula tinkinga nekutfola tisombululo letinemcondvo tingafundziswa etilwimini taseAfrika. Lolucwaningo lusebentise indlela yekucwaninga ye khwalithethivu lephindze isebentise indlela yekusampula lebuka injongo (purposive). Lolucwaningo lincike kakhulu ekubukisiseni lwendlela yeluhlolo lwedeksithophu(Desktop), luhlolo lwetincwadzi (document analysis). Lelucwaningo lubuke kusetjetiswa kwe-Actor Network Theory lokuyi mfundziso-njulalwati (theoretical framework) lokungilo lolukwati kusebentisana naleminye imibhalo. Kuhlatiywa kwalokutfolakele (data analysis) lutawusebentisa Indlelanchubo yekwehlukanisa ngetingcikitsi. Kufaka ligalelo: Lolucwaningo lungaba lusito kubafundzisi, ikakhulukati etikhungweni temfundvo lephakeme ekufundziseni tilwimi tase-Afrika ngekusebentisa itheknoloji kanye netilwimi tase-Afrika uma selutfolakala, selukhulunywa ngaphesheya kwetilwandle.

Repositioning the concept of Xenophobia in the African context: Why do we allow ourselves to be defined by others?

The main goal of this work is to reposition the concept of xenophobia in the African context. This is to reflect on Thabo Mbeki’s narrative on Xenophobia. Africa continues to remain a pawn due to the Eurocentric ideas that have been forced on the Africans also brought about a foreign concept of xenophobia as if Africans are not brothers in the name of brotherhood. It is incontestable that brothers are, in some cases engage in fight, but there was a laid down rules and norms of resolving such misunderstanding. With the employment of qualitative research methods, complex interdependence theory garnished with Anglo-American conspiracy theory; we argue that, Anglo-American imperialist managed to spread and enforce the usage of English through imposing the concept xenophobia in the continent while failing to capture the essence of sibling fights or fights amongst African brothers and sisters conceptualised by the Batswana speaking people as maragana teng a bana motho. Through acknowledging Mbeki’s narrative and some sampled African languages, we argue that, while there is a word for foreigner in these sampled African languages, there is no direct or indirect translation of xenophobia in African languages. We conclude that the concept xenophobia has its etymological foundations in Greek and is therefore foreign in any of the African and South African indigenous languages.

Careers in languages: Awareness by Grade 12 Tshivenḓa learners in Thembisa, Gauteng

Home language learners exhibit negative attitudes towards South African indigenous languages. These languages are perceived as languages that are not beneficial in terms of upward mobility in the workplace. Being aware of the possibilities of studying South African indigenous languages can be a good motivation for learners. A positive attitude towards careers through South African indigenous languages could help to maintain and promote these languages, Tshivenḓa in particular. The study’s objective was to explore if learners and teachers from selected high schools in Thembisa, Gauteng province, were aware of careers that can be pursued through Tshivenḓa, and thus to help determine their perceptions towards their home language as a subject at schools. The study adopted a qualitative approach and a case study design that purposively sampled 16 participants. The findings indicated that Tshivenḓa learners are not exposed to possible language careers that can be pursued through South African indigenous languages. On the other hand, the findings revealed that teachers were aware of the possible careers in these languages. This study concluded that research must be undertaken to make high school learners aware of the available language careers in South African indigenous languages.Contribution: The study closes the gap in career awareness by making learners and teachers in general, and Tshivenḓa L1 learners and teachers in particular, aware of possible careers that exist in South African indigenous language qualifications. The study highlights the need for producing more qualified language practitioners to promote previously marginalised South African indigenous languages and encourage learners to consider pursuing careers in these languages. It also encourages teachers to put more effort into the teaching of South African indigenous languages and for government departments to do more to promote South African indigenous languages by making sure that services are provided to the citizens in the languages that they understand better; as a result, this would create more job opportunities for language graduates, especially regarding the Tshivenḓa language.

What can be said?

Abstract We employ a cluster approach to explore the comparative semantic maps of several markers of modal possibility – the “potential” prefix nga‑ and expressions meaning, roughly, ‘know how to’ and ‘be able to’ – in four South African Nguni languages: isiNdebele, isiZulu, isiXhosa, Siswati. We also compare the Nguni results with results from Xitsonga, a closely related language outside of the Nguni clade. The languages exhibit cross-linguistic differences in the expansion of core meanings, some of which do not appear to follow the cross-linguistically common diachronic pattern in which goal-oriented modality precedes deontic modality. In addition, the distinction between inherent and learned ability is salient in several of these languages. Semantic maps of the markers’ functional distributions further show the diversity of modal systems that can be found even in closely related languages in significant contact with one another.

Cultural adaptation and Sepedi translation of the Activities-specific Balance Confidence scale.

The Activities-specific Balance Confidence (ABC) scale is a widely used measure to identify older adults with balance difficulties. However, its applicability in the diverse South African context is hindered by cross-cultural and linguistic differences. Limited research exists on the use of the ABC scale in native South African languages. This study aimed to translate and culturally adapt the ABC scale into Sepedi, evaluate its reliability and determine self-perceived balance confidence among elderly individuals in a rural community. The ABC scale was translated and culturally adapted into Sepedi. Two trained raters administered the Sepedi version of the ABC (ABC-S) scale to 32 individuals aged between 60 and 88 years. Test-retest reliability and inter-rater reliability were determined, with one rater re-administering the scale 2 weeks later. Ten items from the original ABC scale were modified because of cultural, semantic or contextual inappropriateness. The ABC-S scale demonstrated very good intra- and inter-rater reproducibility, with an average intraclass correlation coefficient (ICC) of 0.85 and 0.81, respectively. The self-perceived balance confidence among elderly Sepedi individuals, as evaluated by the ABC-S scale, was high, with an average score of 81.3 and a range of 58.1 to 95.9. The ABC-S scale is a reliable measurement tool to investigate balance confidence in Sepedi-speaking older adults.Contribution:The ABC-S scale is a valuable screening tool for the identification of balance difficulties in Sepedi-speaking older adults as well as research settings.

“Eish it’s getting really interesting”: borrowed interjections in South African English

Abstract This article offers a descriptive account of seven interjections, eish, yho, tjo, sho, hayi, hau, and mxm, which are adopted from different local South African languages into South African English. It investigates the frequencies, orthography, syntactic position, collocational forms and discourse-pragmatic roles of these seven interjections, through the lens of pragmatic borrowing and postcolonial corpus pragmatics. The data were retrieved from the South African segment of the Global Web-based English corpus and underwent quantitative and qualitative analysis. The findings indicate that the interjections are all emotive interjections, which mostly express negative emotions, except hayi, which is a phatic interjection that is largely used to show disapproval of some information. All the interjections favour clause-initial position except mxm, which is a loan interjection that represents the kiss-teeth or suck-teeth oral gesture that is common in some parts of Africa and the Caribbean. The article affirms that these loaned interjections accentuate the distinction of South African English from other varieties of English.

Discourse-Pragmatic Borrowing in South African English

This paper explores five borrowed discourse-pragmatic features—wena, mna/mina, yazi, phela, and ke—which are transferred from indigenous South African languages into South African English, with the objective of investigating their frequency, position, collocational patterns, and discourse-pragmatic functions. The data, which are taken from the South African component of the Global Web-based English corpus, are analysed quantitatively and qualitatively, from a postcolonial corpus pragmatic framework. The results show that the discourse-pragmatic features are generally infrequent, orthographically stable, and prefer the clause-initial position. The paper indicates that wena is used as an address term to call for the addressee’s attention, mna/mina is used to emphasise personal identity, while yazi seeks confirmation of and signals shared knowledge. Phela is an emphasis marker, while ke indicates contrast, textual coherence, and emphasis. This study underscores the contributions of indigenous South African languages to the discourse-pragmatic features of South African English.

Ideological framing of sign languages and their users in the South African press

Abstract This exploratory study provides an overview of prominent themes pertaining to portrayals of sign languages (SLs) and Deaf people in the South African press (2011–2019), as well as an analysis of a subset of articles to illustrate the discursive constructions of each of the prominent ideological framings. The findings of the paper suggest that many ways of representing South African Sign Language (SASL) and their users align with international trends. The two most prominent ideological framings are the medical/disability model and the linguistic minority model. Within the medical model, SLs are seen as inferior means of communication used by a disabled minority. Within the linguistic minority framework SLs are regarded as natural, legitimate languages deserving equal status to spoken languages. The paper also identifies an ideological framing that is not predicted by the international literature, coined here as ‘diversity tokenism’. Diversity tokenism is when SL is mentioned only to increase perceived diversity, where diversity is a commodity that holds social capital. This portrayal of SASL seems to be increasing and holds a warning: although SASL users have received official recognition and rights through the recent declaration of SASL as an official language, it might not be the end of the battle to ensure that users of SASL can live out their linguistic citizenship.

Exploring intellectualisation of South African indigenous languages for academic purposes

Language represents an individual’s identity in many respects. It is a natural ability of any average person that they use to express thoughts and ideas, investigate their traditions and experiences, and better their community and the laws that govern it. The ability to choose the official language was acknowledged in the Bill of Rights, and the Constitution acknowledges that indigenous languages are a commodity which has not been fully used. By studying the difficulties of intellectualising indigenous languages in modern South Africa, this research aims to promote the usage and enhance the prestige of indigenous languages. The challenge in basic education in South Africa is that South African indigenous languages are not prioritised. Moreover, there appears to be a disconnection between language policy and implementation.Contribution: A document analysis was undertaken to review literature on the possibility of intellectualising South African indigenous languages by considering various theories and methods of terminology development, including interborrowing within South African indigenous languages and adaptation of some English and Afrikaans words as already entrenched languages in education.

The use of technology to preserve indigenous languages of South Africa

Indigenous languages in South Africa must be preserved to ensure that they do not lose their identity and become extinct. The four indigenous languages with the fewest speakers among South Africa’s 12 official languages are: Xitsonga, Siswati, Tshivenḓa and isiNdebele. The preservation of these languages in South Africa has been a long-standing challenge because of various social and economic factors. With the advancement of technology, opportunities have arisen to preserve and promote the use of these languages. Therefore, this study explores various technological strategies that can be used to preserve the South African indigenous languages. These languages can be preserved by making them widely accessible to users through various strategies such as localisation of daily used technology, translation through crowdsourcing, digitisation and archiving. Digital learning tools such as machine translation (MT) and creating online dictionaries can also contribute to preserving these languages. Each of these strategies offers benefits on how technology could be employed effectively and facilitate the preservation of indigenous languages. This study demonstrates the significance of technology in preserving indigenous languages and promoting their use around the world.Contribution: This study fills the practical gap in the use of technology to adequately preserve minority indigenous languages of South Africa, namely, Xitsonga, Siswati, Tshivenḓa and isiNdebele. These languages do not receive much attention in terms of preservation using technology in South Africa. Therefore, the study provides practical technological strategies that need to be implemented to preserve the indigenous minority languages. The insight of this study into the use of technology to preserve South African languages fits well within the scope of Literator, which is to publish studies in linguistics and literature with a special focus on South African languages. This publication will bring solutions to how minority languages could be preserved in the context of South Africa.

South African Sign Language (SASL) Interpreter Portrayal on SABC 1 News Bulletin: What Do Viewers Think?

Television Sign Language (SL) Interpreters play a vital role in providing the Deaf community with access to information and knowledge in their primary language, Sign Language. This helps the Deaf community stay informed regarding events in their local and global environment, contributing to their development. However, a lack of research exists on SL interpreters on television. With on-screen placement of SL interpreters during news broadcasts being the primary focus in previous studies, this paper attempts to offer a unique contribution from an audience perception examination of South African Sign Language (SASL) interpreter portrayal in newscasts and its impact. The research, conducted on the SABC 1 news bulletin, employed a qualitative method and purposive sampling. Twelve participants, divided into four categories – Deaf, Hard of hearing, hearing, and SASL interpreters participated in the study through virtual interviews as data collection method. The study findings indicate the portrayal of SASL interpreters on television reflects the discrimination, inequality, and disrespect faced by the Deaf community.