Elamur Research⌕
☰
home / papers / 2022 / An Investigation of Indian Native Language Phonemic Influences on L2 English Pronunciations

An Investigation of Indian Native Language Phonemic Influences on L2 English Pronunciations

Paper · 2022 · Shelly Jain, Priyanshi Pal, Anil Vuppala et al.

Summary

Speech technology often fails on accented English. India is a hard case: it has many native languages but few linguistic descriptions of how they shape spoken English. The authors examine how the phoneme inventories and phonotactic constraints of 18 Indian languages (Indo-Aryan, Dravidian, Tibeto-Burman) carry over into Indian English (IE).

They build on two sources:

  • Indic TIMIT, with recordings of 80 speakers split into five regional groups.
  • An earlier study (Pal et al., 2022) that compared pronunciation rules from the literature with rules mined from phonetic transcriptions, both measured against Received Pronunciation.

Each candidate rule is then checked against native-language (L1) phonology. Rules that L1 behaviour explains are kept and labelled universal or regional. Rules with no L1 explanation, or that contradict it, are dropped.

The result is a profile of IE pronunciation with pan-Indian and region-specific traits. Examples include tense vowels replacing lax ones and dental or retroflex stops replacing alveolar /t/, /d/. Such a profile supports Indian-accented ASR and TTS, custom lexicons, accent conversion and mispronunciation detection.

Key points

  • Data: Indic TIMIT (80 speakers, five groups from Northeast/East to Lower South). Two linguists transcribed 2,342 recordings. Pal et al. (2022) extended the annotation to 15,974 recordings, with intra-rater Cohen's Kappa of 0.827.
  • Prior rules fall into three categories: in both literature and data, only in the data, or only in the literature.
  • Across the 18 languages there are 70 unique phonemes. Only 9 appear in all 18 languages, and just 2 of those are vowels. Percentile tiers split phonemes into three bands: present in 16+ languages, in 10+ languages, and in fewer than 10.
  • Only about a third of the languages distinguish vowel length. This explains why the data show short vowels where long ones were predicted, so all vowel-substitution rules are treated as universal.
  • Several rules were rejected: palatalisation (/j/ insertion), gemination and schwa insertion, plus some literature-only regional claims such as Telugu /S/→/s/ and Gujarati/Marathi /f/→/pʰ/. These rules were either unsupported by the data or lacked an L1 justification.

Interesting ideas

  • I1. Accept a rule only when two independent sources agree, here data-driven mining and an explanation from L1 phonology; candidates without a plausible cause are dropped rather than kept as statistical noise.
  • I2. Sort every rule by its provenance (both sources, data only, literature only) and investigate each bucket separately instead of merging them.
  • I3. Rank features by how many related varieties share them, then use percentile thresholds to split them into universal, frequent and rare tiers.
  • I4. Separate universal behaviour (good for handling unseen varieties) from regional behaviour (good for adapting to a known variety) within one profile.
  • I5. Explain a surprising mismatch (short vs long vowels) by a missing distinction in the source system, then merge the categories rather than count it as an error.

For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

  • I1 (learn): Self-induced repair rules could be admitted only if silver data supports them and a linguistic rationale (e.g. a UD guideline or VerbNet class) explains them, with unexplained rules logged as rejected.
  • I2 (learn): Disagreements between witness parsers and documented annotation conventions could be bucketed the same way, so "data-only" patterns are flagged for review rather than silently trained on.
  • I3 (tree): When converting between UD dialects or treebanks, constructions shared by most treebanks could become default rules, with rare ones kept as treebank-specific exceptions.
  • I4 (tokens): Lexicon and morphology rules for English variants (spelling, regional forms) could be split into a shared core plus variety-specific overlays, keeping the tagger explainable about which layer fired.
  • I5 (learn): In evaluation, errors between distinctions the gold standard does not reliably make could be merged or reported separately, so scores do not punish unresolvable contrasts.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Arts and Humanities · 9,296 worksLanguage and Linguistics · 9,206 worksPsychology · 798 worksExperimental and Cognitive Psychology · 463 works

🧭 Topics 1Research subjects and languages the work is about.English · 1,373 works

Cite

Cite as Shelly Jain, Priyanshi Pal, Anil Vuppala, Prasanta Ghosh and Chiranjeevi Yarra. 2022. An Investigation of Indian Native Language Phonemic Influences on L2 English Pronunciations. arXiv:2212.09284.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.48550/arxiv.2212.09284 · arXiv 2212.09284

@misc{jain2022investigation,
  title = {{An Investigation of Indian Native Language Phonemic Influences on L2 English Pronunciations}},
  author = {Shelly Jain, Priyanshi Pal, Anil Vuppala, Prasanta Ghosh and Chiranjeevi Yarra},
  year = {2022},
  doi = {10.48550/arxiv.2212.09284},
  eprint = {2212.09284},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.48550/arxiv.2212.09284}
}

References

38 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Arun Baby, Pranav Jawale, Saranya Vinnaitherthan, Sumukh Badam, Nagaraj Adiga, and Sharath Adavanne. 2021. Non-native English lexicon creation for bilingual speech synthesis. ArXiv, abs/2106.10870. ↗
  2. Ram Krishna Bansal. 1990. The pronunciation of English in India. Studies in the pronunciation of English: A commemorative volume in honour of AC Gimson, pages 219–230. find
  3. Ram Krishna Bansal and John Brian Harrison. 1994. Spoken English: A Manual of Speech and Phonetics. Sangam. find
  4. Peri Bhaskararao and Arpita Ray. 2017. Telugu. Journal of the International Phonetic Association, 47(2):231–241. find
  5. Jacob Cohen. 1960. A coefficient of agreement for nominal scales. Educational and psychological measurement, 20(1):37–46. ↗
  6. Matthew S Dryer and Martin Haspelmath. 2013. The World Atlas of Language Structures online. find
  7. Sreeram Ganji, Kunal Dhawan, and Rohit Sinha. 2019. IITG-HingCoS corpus: A Hinglish code-switching database for automatic speech recognition. Speech Communication, 110:76–89. ↗
  8. Silke Hamann. 2003. The phonetics and phonology of retroflexes. Netherlands Graduate School of Linguistics. find
  9. Xian Huang, Xin Jin, Qike Li, and Keliang Zhang. 2020. On construction of the ASR-oriented Indian English pronunciation dictionary. In Proceedings of The 12th Language Resources and Evaluation Conference (LREC), pages 6593–6598. find
  10. India. 1956. States Reorganization Act. find
  11. Danesh Jain and George Cardona. 2007. The Indo-Aryan Languages. Routledge. ↗
  12. Shelly Jain, Aditya Yadavalli, Ganesh S Mirishkar, Chiranjeevi Yarra, and Anil Kumar Vuppala. 2021. IE-CPS Lexicon: An automatic speech recognition oriented Indian-English pronunciation dictionary. In Proceedings of the 18th International Conference on Natural Language Processing (ICON). find
  13. Elinor Keane. 2004. Tamil. Journal of the International Phonetic Association, 34(1):111–116. find
  14. Rajesh Khatiwada. 2009. Nepali. Journal of the International Phonetic Association, 39(3):373–380. find
  15. Sunhee Kim, Kyuwhan Lee, and Minhwa Chung. 2011. A corpus-based study of English pronunciation variations. In Proceedings of the 12th Annual Conference of the International Speech Communication Association (ISCA). find
  16. SP Kishore, Rohit Kumar, and Rajeev Sangal. 2002. A data driven synthesis approach for Indian languages using syllable as basic unit. In Proceedings of Intl. Conf. on NLP (ICON), pages 311–316. find
  17. Bhadriraju Krishnamurti. 2003. The Dravidian languages. Cambridge University Press. ↗
  18. Rohit Kumar, Rashmi Gangadharaiah, Sharath Rao, Kishore Prahallad, Carolyn Penstein Rosé, and Alan W Black. 2007. Building a better Indian English voice using "more data". In SSW, pages 90–94. Citeseer. find
  19. Shakuntala Mahanta. 2008. Directionality and locality in vowel harmony: With special reference to vowel harmony in Assamese. Netherlands Graduate School of Linguistics. find
  20. PJ Mistry. 1996. Gujarati writing. pages 391–394. find
  21. Dipankar Moral. 1997. North-East India as a linguistic area. Monkhmer Studies, pages 43–54. ↗
  22. Raymond WM Ng and Keikichi Hirose. 2012. Syllable: A self-contained unit to model pronunciation variation. In International Conference on Acoustics, Speech, and Signal Processing (ICASSP), pages 4457–4460. IEEE. find
  23. Priyanshi Pal, Shelly Jain, Anil Vuppala, Chiranjeevi Yarra, and Prasanta Ghosh. 2022. Study of Indian English pronunciation variabilities relative to Received Pronunciation.
  24. B Ramani, S Lilly Christina, G Anushiya Rachel, V Sherlin Solomi, Mahesh Kumar Nandwana, Anusha Prakash, S Aswin Shanmugam, Raghava Krishnan, S Kishore Prahalad, K Samudravijaya, et al. 2013. A common attribute based unified HTS framework for speech synthesis in Indian languages. In Eighth ISCA Workshop on Speech Synthesis. find
  25. Pingali Sailaja. 2012. Indian English: Features and Sociolinguistic aspects. Language and Linguistics Compass, 6(6):359–370. ↗
  26. Harold F Schiffman and F Harold. 1999. A reference grammar of spoken Tamil. Cambridge University Press. ↗
  27. Aniruddha Sen. 2003. Pronunciation rules for Indian English text-to-speech system. In Workshop on Spoken Language Processing. find
  28. Christopher Shackle. 2003. Panjabi. The Indo-Aryan Languages, pages 580–621. find
  29. Michael C Shapiro. 2003. Hindi. The Indo-Aryan Languages, pages 250–285. find
  30. Robert Weide. 1998. The Carnegie-Mellon University Pronouncing Dictionary for American English. Pittsburgh, PA: Carnegie Mellon University. find
  31. John Christopher Wells. 1982. Accents of English: Beyond the British Isles, volume 3. Cambridge University Press. find
  32. Robert W Wilkinson. 1974. Tense/lax vowel harmony in Telugu: The influence of derived contrast on rule application. Linguistic Inquiry, 5(2):251–270. find
  33. Shakuntala Mahanta. 2012. Assamese. Journal of the International Phonetic Association, 42(2):217–224. find
  34. Ramawatar Yadav. 2011. A reference grammar of Maithili. De Gruyter Mouton. find
  35. Colin P Masica. 1993. The Indo-Aryan Languages. Cambridge University Press. find
  36. Chiranjeevi Yarra, Ritu Aggarwal, Avni Rajpal, and Prasanta Kumar Ghosh. 2019. Indic TIMIT and Indic English lexicon: A speech database of Indian speakers using TIMIT stimuli and a lexicon from their mispronunciations. In Proceedings of the 22nd Conference of the Oriental chapter of International Committee for the Co-ordination and Standardisation of Speech Databases and Assessment Techniques (O-COCOSDA), pages 1–6. IEEE. find
  37. Rajend Mesthrie. 2008. Introduction: Varieties of English in Africa and South and Southeast Asia. In Africa, South and Southeast Asia, pages 23–34. De Gruyter Mouton. find
  38. Mrinal Miri. 2003. Linguistic Situation in North-East India. Concept Publishing Company. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/jain-2022-investigation-indian-native-language-phonemic/