Navigation – Plan du site

AccueilNuméros9(2)Comptes rendusDavid Roberts & Stephen L. Walter...

Comptes rendus

David Roberts & Stephen L. Walter (eds.), Tone orthography and literacy: The voice of evidence in ten Niger-Congo languages

Helen Eaton
Référence(s) :

David Roberts & Stephen L. Walter (eds.), Tone orthography and literacy: The voice of evidence in ten Niger-Congo languages (Studies in Written Language and Literacy 18), Amsterdam/Philadelphia, John Benjamins, 2021, xxii + 375 p.

Texte intégral

1The introduction to Tone orthography and literacy: The voice of evidence in ten Niger-Congo languages, edited by David Roberts and Stephen L. Walter, states that the aim of the volume is to report on experiments designed to investigate the effects of full tone marking on reading and writing in ten African languages, all from the Niger-Congo phylum. The need for this research is evident in the book’s review of previous work in this area. In more than a hundred years of orthography development in African languages, “not a shred of empirical evidence” (Roberts, Merz & Reeder, 3) has been put forward to confirm that full tone marking in the orthography (understood here as marking one fewer tones than the number of phonemic tones in the language) improves reading and writing performance.

2Two previous tone orthography experiments (Bird 1999; Bernard et al. 2002), carried out on Grassfields languages of Cameroon, present empirical evidence on the efficacy of full tone orthographies, but this evidence shows that far from improving reading and writing fluency, these orthographies hinder it. Despite this, full tone marking remains common in newly developed orthographies for African languages. There is clearly a need to replicate the experimental research done by Bird and Bernard et al. and also to expand on it, by investigating tone languages other than those in the Grassfields family and by increasing the limited sample size of these previous experiments.

3Tone orthography and literacy reports on experiments which meet both these challenges impressively. Ten languages (representing various types of tone language) were investigated: Elip, Mmala, Yangben (Cameroon); Tem, Ife (Togo); Mbelime, Nateni, Idaasha (Benin); Yoruba (Nigeria); and Eastern Dan (Côte d’Ivoire). 308 speakers of these languages participated in the research, which was led by principal investigator David Roberts and carried out with the help of a number of field linguists and local facilitators.

4In their introductory chapter, Roberts, Merz & Reeder set out the background for the research and state their research question as “To what extent does full tone marking contribute to oral reading fluency, comprehension, and writing accuracy, and does that contribution vary from language to language?” (4). Their focus is on comparing full tone marking with zero tone marking and excludes the various partial tone marking options that might in fact be more effective than either of these first two methods. They then go on to discuss the ethno-literacy context of experimental research in Africa, explaining clearly how the non-WEIRD (Western, Educated, Industrialized, Rich, Democratic; Henrich et al. 2010a and 2010b) setting presents various challenges to the experiment design. They also give an overview of the languages in which the research was done (referred to throughout as the “focal languages”), covering their location, genealogical affiliation and vitality status. Although the ten languages were selected simply by virtue of being the ones accessible to the researchers who volunteered to take part in the project, they represent a range of language families (Mande, Gur, Ede and Bantu A62) within the Niger-Congo phylum and, as is seen in later chapters, great variety in ethno-literacy backgrounds. The introductory chapter continues with a helpful review of literature concerning tone pedagogy in primers and transition guides, covering such theoretical concerns as the concept of a consistent word image (e.g. Bird 1999) and whether to teach tone patterns rather than individual tones (Snider 2018).

5Part I takes the ten focal languages in turn (with the exception of the three Bantu languages, which are treated together) and systematically gives the reader information on the languages themselves (focusing mainly on phonological information), their orthographies and their ethno-literacy backgrounds. Part II begins with a chapter explaining the experiment design in detail (Chapter 10), which broadly follows that of Bird 1999, but with modifications. The same four texts from Bird’s experiment were translated and used for the oral reading task, with an additional text in English (for Nigeria) and French (elsewhere), in order to incorporate L2 data into the experiment design. A further modification of Bird’s original experiment concerns the way in which comprehension was measured. Bird (1999: 103) recognised this area as a weakness in his design and consequently comprehension questions were added to address this. The chapter gives detailed information on how the different types of performance data (oral reading speed, accuracy, comprehension, and writing accuracy) were elicited and scored. The final part of Chapter 10 walks the reader through the multilevel statistical methodology used for the experiment.

6Part II then continues with five chapters which present the results of the experiments as an analysis of oral reading speed (Chapter 11), general accuracy (Chapter 12), tonal accuracy (Chapter 13), comprehension (Chapter 14) and tone writing (Chapter 15). After the detail of these chapters, a final chapter summarises the results and implications for each of the focal languages and draws attention to potential areas for future research. In their conclusion, Roberts, Reeder & Walter state that their experiments do not provide “across-the-board evidence” that tone marking reduces fluency (279), but full tone marking may have some benefit for readers with less education or less L1 literacy experience. The experiments suggest that variation in performance has little to do with how tone is marked. The linguistic and orthographic profiles of a language appear to be weaker predictors of reading and writing performance than the ethno-literacy profile of the language community and the social profile of the individual. This can be seen most clearly when the data from Yoruba and Ife are compared (Roberts, Reeder & Walter, 294) and has been given more extensive treatment in a separate article (Roberts et al. 2022).

7The final chapter also contains a section on tone pedagogy, drawing together recommendations for primers and transition guides, based on the findings of the experiments. The following section recognises weaknesses in the experiment design and makes suggestions for improvements in future research. Then a final section steps back from the specifics of tone orthography and offers some thoughts on the literacy situation in Africa in general, pointing out that, “On average, the participants, all adults, orally read their L2 with a speed equivalent to what is expected of an American seven-year old, and their L1, whether tone is marked or not, with a speed considerably slower than what is expected of an American six-year old” (Roberts, Reeder & Walter, 303). Deciding whether and how to write tone in an orthography is not a simple matter and should be made on the basis of linguistic analysis and experimental data in tandem. The authors end with an exhortation for others to take on the challenge of classroom research — despite the difficulties it presents — in order to provide evidence which can direct future literacy efforts.

8The series of experiments which form the basis of Tone orthography and literacy have clearly made an important contribution to an under-researched area and gone some way to bridge the gap between phonological analysis and the classroom when it comes to orthography development. The focal languages, though not specifically chosen to be such, are a fascinating mix of different types of tone language and different stages of language development. For several languages, orthography development is a recent endeavour and orthography decisions are still very much in flux. This makes for a messier but also more interesting situation, with the book providing a snapshot of where the orthographies stood at a particular point in time.

9The book as a whole is readable, systematic and clear. The chapters on different languages follow the same template as regards their sections and subsections, but this does not seem repetitive as the languages and their contexts are different. The long and fascinating history of the Yoruba orthography, for example, contrasts with the very recent (since 2005) development of orthographies for Elip, Mmala and Yangben. All the language chapters (2‑9, comprising Part I) are detailed and well researched, but there are still some areas where further detail would have been welcome. For example, in relation to Yoruba, Harley & Reeder comment that, “It is impossible to determine from looking at the primer if teachers are encouraging students to mark tones on vowels as they write or not” (126). More information on actual practice would have been useful.

10For some, the value of Tone orthography and literacy will be as a source of information about tone languages and tone orthographies. It is possible to dip in and out of the book if only certain languages or types of tone systems are of interest (the presence of a language index and a topical index make this easy to do). For others, the value of the book will be as a textbook that sets out in detail how to replicate the kind of experiment described. The authors are very thorough in their description of the experiment design, even describing step-by-step how the test texts were translated (Roberts & Walter, 190) and explaining how spreadsheets for scoring were set up. A link is provided at the end of Appendix 1 to a downloadable spreadsheet containing the Level 1 demographic variables and performance variables gathered during the experiments.

11The statistical methodology is also rigorously explained, with an explanation of the basic notions provided for readers who are unfamiliar with the approach (Roberts & Walter, 198‑199) and helpful guidance on how to interpret the results as the analysis proceeds (e.g. “Most of the detail in Table 4 is technical and of interest only to those who make use of multilevel modelling, but a summary will be helpful for the inquisitive reader.” Roberts & Walter, 216). Each of the five analysis chapters includes a subsection called “Interpretation and discussion” to which the statistics-averse can turn if the detail of the preceding subsections is too much. It is in these subsections that we find specific insights from the analysis, such as the suggestion that “good readers are minimally affected by whether or not tone is fully marked while struggling readers are much more likely to rely on them, or at least to be more perturbed by their unexpected absence” (Roberts & Walter, 232‑233). In some cases the authors note how the observations made from the data may have multiple explanations, such as the following comment in relation to the high rate of tonal errors in oral reading for Eastern Dan: “Either marking tone with punctuation is less effective than its proponents claim, or tone pedagogy needs improving, or this language community is an aberrant case for cultural, historical, or educational reasons” (Roberts & Walter, 246). In this particular case, a follow-up experiment was done which addressed this issue (Roberts et al. 2019), so readers can turn to this publication to find out more.

12The transparency of all the chapter authors regarding the significance of their findings is commendable. For example, the evidence in support of full tone marking in three languages (and marginal support in two more) is described as “uneven and patchy” (Roberts, Reeder & Walter, 279). The authors’ critical assessment of the shortcomings of their experiment design and their recognition of the factors they were not able to control are also honest and helpful. Their detailed recommendations for researchers wishing to replicate this design (Roberts, Reeder & Walter, 298‑303) are particularly thorough and useful.

13It is easy to see how this book will be of practical benefit to those wishing to develop tone orthographies for African languages, and in particular to those wanting to conduct robust experiments on the efficacy of these orthographies. The “voice of evidence”, as referred to in the book’s subtitle, is clearly heard throughout the book and I join the authors in hoping that this book will inspire and inform other researchers to contribute their own evidence in time.

Haut de page


Bernard, H. Russell, George N. Mbeh & W. Penn Handwerker. 2002. Does marking tone make tone languages easier to read? Human Organization 61(4). 339‑349.

Bird, Steven. 1999. When marking tone reduces fluency: An orthography experiment in Cameroon. Language and Speech 42(1). 83‑115.

Henrich, Joseph, Steven J. Heine & Ara Norenzayan. 2010a. Most people are not WEIRD. Nature 466. 29.

Henrich, Joseph, Steven J. Heine & Ara Norenzayan. 2010b. The weirdest people in the world? Behavioral and Brain Sciences. 33(2‑3). 61‑83.

Roberts, David, Dana Basnight-Brown & Valentin Vydrin. 2019. Marking tone with punctuation: Orthography experimentation and reform in Eastern Dan (Côte d’Ivoire). In Yannis Haralambous (ed.), Graphemics in the 21st Century. Brest, June 13-15, 2018. Proceedings (Grapholinguistics and its Applications 1), 293‑327. Brest: Fluxus Editions.

Roberts, David, Matthew Harley & Stephen L. Walter. 2022. The contribution of full tone marking to oral reading fluency and comprehension in Yoruba and Ife. Written Language & Literacy 25(2). 253‑282.

Snider, Keith L. 2018. Tone analysis for field linguists. Dallas, TX: SIL International.

Haut de page

Pour citer cet article

Référence électronique

Helen Eaton, « David Roberts & Stephen L. Walter (eds.), Tone orthography and literacy: The voice of evidence in ten Niger-Congo languages »Linguistique et langues africaines [En ligne], 9(2) | 2023, mis en ligne le 20 décembre 2023, consulté le 18 juin 2024. URL : ; DOI :

Haut de page


Helen Eaton

SIL International

Haut de page

Droits d’auteur


Le texte seul est utilisable sous licence CC BY-SA 4.0. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.

Haut de page
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search