Abstract
We report on a study that aimed to improve an existing tone label prediction algorithm for Sesotho, an official language of South Africa. Tone is an important prosodic feature of Sesotho, since speakers use tone to distinguish meaning. In order to implement tone in a Text-to-Speech system for Sesotho, a tone modeling algorithm must receive as input the tone labels of the syllables of each word. Then it can predict the appropriate intonation of the word. Since Sesotho does not mark tone labels in orthography, the labels have to be predicted according to the tonal rules of the language. The existing tone label prediction algorithm has two drawbacks, namely it implements three tonal rules only and is restricted to the clitic phrase domain. In our study, we developed an algorithm that implements four additional tonal rules and addresses all parts of speech. The results show that the latter algorithm significantly improves the existing one by increasing the number of matched tone labels.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Acute accents indicate a high tone. Low tones are unmarked.
Please note the following: \(\varpi \) indicates the phonological word domain. Numbers indicate the grammatical person, e.g., the number 1 indicates the first person. Underlying high tones are underlined. Moreover, between the Sesotho sentence and its English translation is a gloss, i.e., a brief notation, that states the grammatical category of each word in the sentence. Glosses are as follows: AGR = agreement marker, SM = subject marker, NEG = negative form, PRP = preposition, T/A = tense and aspect marker, ADV = adverb, INF = Infinitive, CONJ = conjunct, OM= object marker, FUT = future tense, NP = noun, PC = possessive concord, PRN = pronoun and SG = singular.
In Khoali (1991), the English translation is given as “I like to go a lot”.
In Khoali (1991), the English translation is given as “Thabo likes mutton and wants them to eat it only.”
The Human Language Technology group at the Meraka Institute of the Council of Scientific and Industrial Research has a corpus of Sesotho sentences taken from poetry, folk tales, and the South African Government Gazette.
A trisyllabic noun stem is a stem with three syllables, for example di.po.tso “questions”. The full stops in the word represent syllable boundaries.
A degree of freedom is the number of values that are free to vary.
References
Boslaugh, S., & Watters, P. A. (2008). Statistics in a nutshell (1st ed.). USA: O’Reilly Media Inc.
Brennan, R. L., & Prediger, D. J. (1981). Coefficient kappa: Some uses, misuses, and alternatives. Educational and Psychological Measurements, 41(1), 687–699.
Cole, D. T., & Mokaila, D. M. (1962). A course in Tswana (1st ed.). Washington, DC: Georgetown University.
Creissels, D., Chebane, A. M., & Nkhwa, H. M. (1997). Tonal morphology of the Setswana verb (1st ed.). Munich: Lincom Europe.
Demuth, K. (1995). Problems in the acquisition of tonal systems. In J. Archibald (Ed.), The acquisition of non-linear phonology (pp. 111–134). Hillsdale, New York: Lawrence Erlbaum Associates.
Doke, C. M., & Mofokeng, S. M. (1957). Textbook of Southern Sotho grammar (1st ed.). London: Longmans Green and Co.
Dtunji, A. D., Wonga, S., & Beaumonta, A. J. (2007). A fuzzy decision tree-based duration model for standard Yoruba text-to-speech synthesis. Computer Speech and Language, 21, 325–349.
Du Plessis, J. A., Gildenhuys, J. G., & Moiloa, J. J. (1974). Tweetalige woordeboek Afrikaans-Suid-Sotho/Bukantswe ya maleme-pedi Sesotho-Seafrikanse (1st ed.). Cape Town: Via Afrika Beperk.
Faasz, G., Heid, U., Taljard, E., & Prinsloo, D. J. (2009). Part-of-speech tagging of Northern Sotho: Disambiguating polysemous function words. In EACL 2009 Workshop on Language Technologies for African Languages, Athens, Greece, pp. 38–44.
Fleiss, J. L. (1971). Measuring nominal scale agreement among many raters. Psychological Bulletin, 5, 378–382.
Gandour, J. T., & Harshman, R. A. (1978). Crosslanguage differences in tone perception: A multidimensional scaling investigation. Language and Speech, 21, 1–33.
Halle, P. A., Chang, Y., & Best, C. T. (2004). Identification and discrimination of Mandarin Chinese tones by Mandarin Chinese vs. French listeners. Journal of Phonetics, 32, 395–421.
Khoali, B. (1991). A Sesotho tonal grammar. Ph.D. Thesis, University of Illinois.
Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical data. Biometrics, 33, 159–174.
Lee, T., Kochanski, G., Shih, C., & Li, Y. (2002). Modeling tones in continuous Cantonese speech. In Proceedings of the 7th International Conference on Spoken Language Processing, Colorado, USA, pp. 2401–2404.
Lewis, M. P. (2009). Ethnologue: Languages of the world (16th ed.). Dallas, Texas: SIL International.
Li, Y., Lee, T., & Qian, Y. (2004). Analysis and modeling of F0 contours for Cantonese text-to-speech. ACM Transactions on Asian Language Information Processing, 3(3), 169–180.
Lombard, D. P. (1976). Aspekte van toon in Noord-Sotho. Ph.D. Thesis, University of South Africa.
Louw, J. A., Davel, M., & Barnard, E. (2005). A general-purpose isiZulu speech synthesiser. South African Journal of African Languages, 25(2), 92–100.
Mabille, A., Dieterlen, H., & Paroz, R. A. (1961). Southern Sotho-English dictionary. Morija: Morija Sesutu book depot, Lesotho, 8th. rev. and enlarged ed edition.
Nespor, M., & Vogel, I. (1986). Prosodic phonology (1st ed.). Dordrecht: Foris Publications.
Ni, J., Sakai, S., Shimizu, T., & Nakamura, S. (2008). Prosody modeling from tone to intonation in Chinese using a functional F0 model. In Proceedings of the 2nd International Symposium on Universal Communication, pp. 397–404, Osaka, Japan.
Odejobi, O. A., Beaumonta, A. J., & Wonga, S. (2006). Intonation contour realisation for standard Yoruba text-to-speech synthesis: A fuzzy computational approach. Computer Speech and Language, 20, 563–588.
Odejobi, O. A., Wonga, S., & Beaumonta, A. J. (2008). A modular holistic approach to prosody modelling for standard Yoruba speech synthesis. Computer Speech and Language, 22, 39–68.
Raborife, M. (2009). The implementation of Sesotho tonal rules in a text-to-speech system. Honours Research Report: School of Computer Science, University of the Witwatersrand.
Raborife, M. (2011). Tone labelling algorithm for Sesotho. Masters Dissertation, School of Computer Science, University of the Witwatersrand.
Raborife, M., Zerbian, S., & Ewert, S. (2011). Developing a corpus to verify the performance of a tone labelling algorithm. In Proceedings of the 22nd annual symposium of the pattern recognition association of South Africa, Vanderbijlpark, South Africa, pp. 126–131.
Randolph, J. J. (2005). Free-marginal multirater kappa: an alternative to Fleiss’ fixed-marginal multirater kappa. In Joensuu University Learning and Instruction Symposium, Joensuu, Finland.
Randolph, J. J. (2008). Online kappa calculator. Retrieved 4 February 2011, from http://justus.randolph.name/kappa.
Schadeberg, T. C. (1981). Tone in South African Bantu dictionaries. Journal of African Languages and Linguistics, 3, 175–180.
Syrdal, A. K., & McGory, J. (2000). Inter-transcriber reliability of ToBI prosodic labeling. In Proceedings of the 6th International Conference on Spoken Language Processing, Beijing, China.
Van Niekerk, D. R., & Barnard, E. (2010). An intonation model for TTS in Sepedi. In INTERSPEECH, Makuhari, Japan.
Wang, R., Liu, Q., Tang, D. (1996). A new Chinese text-to-speech system with high naturalness. In Proceedings of the 4th international conference on spoken language processing, pp. 1441–1444, Philadelphia, USA.
Zerbian, S. (2006). High tone spread in the Sotho verb. In Proceedings of the 35th annual conference on African linguistics, pp. 147–157, Somerville.
Zerbian, S. (2009). Segmental and suprasegmental properties of monosyllables in Sotho/Tswana. In International conference “Monosyllables—from phonology to typology”. University of Bremen, Germany, pp. 111–115.
Zerbian, S., & Barnard, E. (2008). Phonetics of intonation in South African Bantu languages. Southern African Linguistics and Applied Language Studies, 26(2), 235–254.
Zerbian, S., & Barnard, E. (2010). Word-level prosody in Sotho-Tswana. In Proceedings of Speech Prosody, Chicago, USA.
Acknowledgments
We would like to thank the reviewers for their constructive comments.
Author information
Authors and Affiliations
Corresponding author
Additional information
This material is based upon work supported financially by the National Research Foundation (NRF) of South Africa. Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and therefore the NRF does not accept liability in regard thereto.
Rights and permissions
About this article
Cite this article
Raborife, M., Ewert, S. & Zerbian, S. Improving a tone labeling algorithm for Sesotho. Lang Resources & Evaluation 49, 19–50 (2015). https://doi.org/10.1007/s10579-014-9281-4
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10579-014-9281-4