A former high school teacher's quest to improve the poor reading ability of learners in low-resource languages has resulted in a groundbreaking PhD. Centered on the development of a means to measure text readability while developing digital language resources for Sesotho, this research is the first of its kind and will also serve as a model for other African languages.

Dr Johannes Sibeko, a lecturer in digital humanities at Nelson Mandela University (NMU) where he coordinates the Digital Humanities Hub, was recently conferred a PhD in languages and literature by the North-West University (NWU).

He conducted his innovative research under the supervision of Prof Menno van Zaanen, professor of digital humanities at the South African Centre for Digital Language Resources (SADiLaR).

Titled “Measuring text readability in Sesotho” Dr Sibeko’s study addresses the reading challenge within the South African context of low-resource languages. It does this by proposing methods for selecting texts suitable for learners' reading at the levels of specific grades. “Text readability indicates how difficult or easy a text is to read. For my research, I surveyed readability measures (formulas that measure how easy or difficult a text is to read) for high-resource languages and adapted them for Sesotho, using data from examination papers, and translations such as the Bible.

“I also developed four basic language resources for Sesotho including two syllabification systems (to chop words into their syllables) which are the first of their kind for Southern African languages,” he explains.

“Such resources are essential for the creation of the readability measures, with text features from Sesotho examination texts used to compute readability values and generate specific models.”

Laying the groundwork for machine learning applications

According to Dr Sibeko, this work resulted in nine readability measures for Sesotho, providing an objective ranking of Sesotho text readability.

“The adaptation of traditional English readability formulas addresses a gap in South African indigenous languages, with the models laying the groundwork for machine learning applications in Sesotho readability assessment and beyond to other low-resource languages. The findings also address the reading ability gaps by addressing the readability levels of the texts thereby ensuring that texts are correctly aligned to the target readers.”

From humanities to digital humanities

“Johannes has done very well during his PhD,” says Prof Van Zaanen. “He was not really computational when he started out, but quite a bit of work in his PhD is now computational. In that sense, he is a wonderful example of someone from the field of humanities who has moved into the field of digital humanities.”

Despite his success, Dr Sibeko remains humble.

Asked how he feels about his new status, he says it still feels unreal. “Being capped as a PhD graduate at the NWU was incredible. My PhD journey was long and lonely at times, but here I am, and the end is quite sweet.”

His advice for other prospective PhD students is to do their thesis by publication. “It gives you a handful of achievements, and keeps you motivated, while you work towards the bigger picture. There is no feeling that compares to seeing one of your submissions in publication state.”

More about Dr Sibeko

Dr Sibeko’s educational journey started with a BA degree in language practice with Sesotho, followed by an MA in applied language studies and a postgraduate certificate in education. His research interests initially focused on language policy and linguistic landscapes, later shifting to language teaching and assessment upon completing his MA. He then decided to transition to digital humanities and the development of resources for low-resource languages.

He has received several accolades for his research, both locally and abroad. He received the Best Paper Award at the Digital Humanities Association of Southern Africa (DHASA) conference in 2021, and he was the only student presenter from Africa invited to the 2022 CLARIN (Common Language Resources and Technology Infrastructure) Annual Conference. His PhD research paper was included in the Post Conference Proceedings publication of this conference. He was also named the Faculty of Humanities Emerging Researcher of the Year at NMU in2022.

Dr Johannes Sibeko was recently conferred a PhD in languages and literature by the NWU.