Elamur Research⌕
☰
home / papers / 2024 / Topics in the Study of the Pragmatic Functions of Phonetic Reduction in Dialog

Topics in the Study of the Pragmatic Functions of Phonetic Reduction in Dialog

Paper · 2024 · Nigel G. Ward and Carlos A. Ortega

Summary

Speakers in conversation often articulate less carefully, but this phonetic reduction has rarely been studied in dialog, either acoustically or for what it signals. This technical report collects material left out of the authors' journal article, whose main claim is that reduction tends to accompany positive assessments in American English. The authors hand-labeled perceived reduction on a graded scale in stretches of English and Spanish conversation, and a second annotator relabeled the same regions to measure agreement. They then checked which prosodic features track the labels and built a simple baseline predictor from acoustic/prosodic features. Unlike in read speech, reduction in dialog goes with high pitch, wide pitch range and intensity. The baseline correlates with human judgments at only 0.24 for English and 0.17 for Spanish, so prosody alone barely predicts reduction. The report also catalogs further candidate pragmatic functions, analyzes why annotators disagreed, and reports a reenactment experiment on prominence. It releases what the authors call the first public dataset annotated for perceived reduction, which matters for evaluating future detectors.

Key points

  • The annotated data covers parts of 5 English (EN) and 6 Spanish (ES) conversations. The original first-pass labels were lost and had to be redone.
  • Annotator agreement was moderate for clear enunciation, weak for normal and mild reduction, and very poor for strong reduction, which was also the rarest label.
  • The baseline predictor reaches correlations of 0.24 (English) and 0.17 (Spanish) with perceived reduction. The original goal of adding reduction to the Midlevel Prosodic Features Toolkit was not achieved.
  • In the actor reenactment study, actors tagged about 28% of words as prominent, but their tags were inconsistent. Prominence showed no significant interaction with positive-vs-neutral reduction (chi-square, p > 0.25).
  • In Spanish, negative assessments (e.g., "no me gusta") showed a tendency toward reduction, often together with a falling pitch and slowing.

Interesting ideas

  • I1. When two annotators differed by more than one scale step, the authors reviewed every such case and sorted the causes into categories, such as reliance on context and prosodic confounds like emphasis, lengthening and creaky voice.
  • I2. Human inter-annotator agreement is treated as the reference point for judging an automatic detector, so a low model score is read against how hard the task is for people.
  • I3. Once a listener knows which word was meant, it seems obvious in the signal, so context leaks into supposedly context-free judgments.
  • I4. The prominence hypothesis failed partly because the concept could not be operationalized, and the authors recommend a pilot study to check that a label is usable before testing hypotheses with it.
  • I5. Each pragmatic function seems to carry its own bundle of cues, of which reduction is only one, rather than relying on a single marker.

For Mova

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

  • I1 (learn): Cases where the witness parsers strongly disagree could be sorted into named failure causes in the same way, giving targeted input for self-induced repair rules instead of a single disagreement count.
  • I2 (learn): Evaluation of uncertain labels such as irony or implicature could report human or LLM-teacher agreement alongside model scores, so weak results are not mistaken for bugs.
  • I3 (answer): Negative controls could strip or shuffle surrounding context to test whether a reader truly extracts an answer from the local text rather than from what it already knows.
  • I4 (frames): Before adding a new pragmatic category to prag, a small pilot could check whether annotators or teachers can apply it consistently.
  • I5 (frames): Speech-act and assessment detection could combine several weak, explainable cues per function instead of keying on one feature, keeping each cue visible in the explanation.

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

 Institutions

🎓The University of Texas at El PasoEl Paso · United States4 works · 17 citations

 Fields, topics and concepts

📚 Shelves 6The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 worksPsychology · 1,552 worksExperimental and Cognitive Psychology · 870 works

🧭 Topics 2Research subjects and languages the work is about.English · 2,263 worksDialogue systems · 624 works

Cite

Cite as Nigel G. Ward and Carlos A. Ortega. 2024. Topics in the Study of the Pragmatic Functions of Phonetic Reduction in Dialog. arXiv:2405.01376.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.48550/arxiv.2405.01376 · arXiv 2405.01376

@misc{ward2024topics,
  title = {{Topics in the Study of the Pragmatic Functions of Phonetic Reduction in Dialog}},
  author = {Nigel G. Ward and Carlos A. Ortega},
  year = {2024},
  doi = {10.48550/arxiv.2405.01376},
  eprint = {2405.01376},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.48550/arxiv.2405.01376}
}

References

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

  1. Aguilar, L., Blecua, B., Machuca, M. & Mann, R. (1993), Phonetic reduction processes in spontaneous speech, in ‘Third European conference on speech communication and technology (Eurospeech)’, pp. 433–436. find
  2. Beechey, T., Buchholz, J. M. & Keidser, G. (2018), ‘Measuring communication difficulty through effortful speech production during conversation’, Speech Communication 100, 18–29. ↗
  3. Bradlow, A. R. & Bent, T. (2002), ‘The clear speech effect for non-native listeners’, The Journal of the Acoustical Society of America 112(1), 272–284. ↗
  4. Burdin, R. S. & Clopper, C. G. (2015), Phonetic reduction, vowel duration, and prosodic structure, in ‘ICPhS’. find
  5. Cammenga, K. S. (2018), The Effect of Vocal Fry on Speech Intelligibility, PhD thesis, Michigan State University. ↗
  6. Chen, L., Jiang, C., Gu, Y., Liu, Y. & Yuan, J. (2022), Automatically detecting reduced-formed English pronunciations by using deep learning, in ‘Proceedings of the 17th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2022)’, pp. 22–26.
  7. Cohen Priva, U. & Strand, E. (2023), ‘Schwa’s duration and acoustic position in American English’, Journal of Phonetics 96, 101198. find
  8. Ernestus, M., Baayen, H. & Schreuder, R. (2002), ‘The recognition of reduced word forms’, Brain and language 81(1-3), 162–173. ↗
  9. Ernestus, M. & Warner, N. (2011), ‘An introduction to reduced pronunciation variants’, Journal of Phonetics 39(SI), 253–260. ↗
  10. Gahl, S., Yao, Y. & Johnson, K. (2012), ‘Why reduce? phonological neighborhood density and phonetic reduction in spontaneous speech’, Journal of Memory and Language 66(4), 789–806. ↗
  11. Gustafson, J., Székely, E., Alexandersson, S. & Beskow, J. (2023), Casual chatter or speaking up? Adjusting articulatory effort in generation of speech and animation for conversational characters, in ‘Workshop on Socially Interactive Human-Like Virtual Agents’. find
  12. Jiao, Y., Berisha, V., Liss, J., Hsu, S.-C., Levy, E. & McAuliffe, M. (2016), ‘Articulation entropy: An unsupervised measure of articulatory precision’, IEEE Signal Processing Letters 24(4), 485–489. ↗
  13. Jurafsky, D., Bell, A., Fosler-Lussier, E., Girand, C. & Raymond, W. D. (1998), Reduction of English function words in Switchboard, in ‘ICSLP’. find
  14. Kahn, J. M. & Arnold, J. E. (2015), ‘Articulatory and lexical repetition effects on durational reduction: Speaker experience vs. common ground’, Language, Cognition and Neuroscience 30(1-2), 103–119. find
  15. Khaghaninezhad, M. S. & Jafarzadeh, G. (2014), ‘Investigating the effect of reduced forms instruction on EFL learners’ listening and speaking abilities.’, English Language Teaching 7, 159–171. find
  16. Kissine, M., Geelhand, P., Philippart De Foy, M., Harmegnies, B. & Deliens, G. (2021), ‘Phonetic inflexibility in autistic adults’, Autism Research 14(6), 1186–1196. ↗
  17. Koreman, J. (2006), ‘Perceived speech rate: The effects of articulation rate and speaking style in spontaneous speech’, The Journal of the Acoustical Society of America 119(1), 582–596. ↗
  18. Lee, S., Bresch, E., Adams, J., Kazemzadeh, A. & Narayanan, S. (2006), A study of emotional speech articulation using a fast magnetic resonance imaging technique, in ‘Ninth International Conference on Spoken Language Processing’. find
  19. Lu, Y. & Cooke, M. (2009), ‘The contribution of changes in F0 and spectral tilt to increased intelligibility of speech produced in noise’, Speech Communication 51(12), 1253–1262. ↗
  20. Lubold, N., Borrie, S. A., Barrett, T. S., Willi, M. M. & Berisha, V. (2019), Do conversational partners entrain on articulatory precision?, in ‘Interspeech’, pp. 1931–1935. find
  21. Ludusan, B., Wagner, P. & Wlodarczak, M. (2021), Cue interaction in the perception of prosodic prominence: The role of voice quality, in ‘Interspeech’, pp. 1006–1010. find
  22. Machač, P. & Fried, M. (2023), Utterance comprehension in spontaneous speech: phonetic reductions and syntactic context. manuscript, Charles University, Prague. find
  23. Malisz, Z., Brandt, E., Möbius, B., Oh, Y. M. & Andreeva, B. (2018), ‘Dimensions of segmental variability: Interaction of prosody and surprisal in six languages’, Frontiers in Communication 3, 25. ↗
  24. Niebuhr, O. (2017), Clear speech — mere speech? how segmental and prosodic speech reduction shape the impression that speakers create on listeners, in ‘Interspeech’, Vol. 18, pp. 894–898. find
  25. Niebuhr, O., Gors, K. & Graupe, E. (2013), Speech reduction, intensity, and F0 shape are cues to turn-taking, in ‘SigDial’, pp. 261–269. find
  26. Niebuhr, O. & Kohler, K. J. (2011), ‘Perception of phonetic detail in the identification of highly reduced words’, Journal of Phonetics 39(3), 319–329. ↗
  27. Niebuhr, O. & Michaud, A. (2015), Speech data acquisition: The underestimated challenge, in ‘Kieler Arbeiten in Linguistik und Phonetik’, Vol. 3, Christian Albrects Universitat zu Kiel, ISFAS, pp. 1–42. find
  28. Picart, B., Drugman, T. & Dutoit, T. (2014), ‘Analysis and HMM-based synthesis of hypo and hyperarticulated speech’, Computer Speech & Language 28(2), 687–707. ↗
  29. Schubotz, L. M., Oostdijk, N. H. & Ernestus, M. T. (2015), Y’know vs. you know: What phonetic reduction can tell us about pragmatic function, in S. Lestrade, P. de Swart & L. Hogeweg, eds, ‘Addenda. Artikelen voor Ad Foolen’, Nijmegen: Radboud Universiteit Nijmegen, pp. 361 – 380. find
  30. Shriberg, E. (2001), ‘To ‘errr’ is human: Ecology and acoustics of speech disfluencies’, Journal of the International Phonetic Association 31, 153–169. find
  31. Tu, M., Grabek, A., Liss, J. & Berisha, V. (2018), Investigating the role of L1 in automatic pronunciation evaluation of L2 speech, in ‘Interspeech’. find
  32. Turnbull, R. (2015), Patterns of individual differences in reduction: Implications for listener-oriented theories, in ‘ICPhS’. find
  33. Turnbull, R. (2017), ‘The role of predictability in intonational variability’, Language and Speech 60, 123–153. ↗
  34. Uther, M., Knoll, M. A. & Burnham, D. (2007), ‘Do you speak E-NG-LI-SH? A comparison of foreigner-and infant-directed speech’, Speech Communication 49(1), 2–7. find
  35. van Son, R. J. & Pols, L. C. (1999), ‘An acoustic description of consonant reduction’, Speech Communication 28(2), 125–140. ↗
  36. Wagner, P., Origlia, A., Avesani, C., Christodoulides, G., Cutugno, F., D’Imperio, M., Escudero Mancebo, D., Gili Fivela, B., Lacheret, A., Ludusan, B., Moniz, H. & Ni Chasaide, A. (2015), Different parts of the same elephant: A roadmap to disentangle and connect different perspectives on prosodic prominence, in ‘Proceedings of the 18th International Congress of Phonetic Sciences’. find
  37. Ward, N. G. (2019), Prosodic Patterns in English Conversation, Cambridge University Press. find
  38. Ward, N. G. (2023), Midlevel https://github.com/nigelgward/midlevel. prosodic features toolkit (2016-2023). ↗
  39. Ward, N. G., Gomez, R. O., Ortega, C. A. & Bugarini, G. (2024), ‘Some pragmatic functions of phonetic reduction’. in preparation. ↗
  40. Ward, N. G., Kirkland, A., Włodarczak, M. & Székely, E. (2022), Two pragmatic functions of breathy voice in American English conversation, in ‘11th International Conference on Speech Prosody’, pp. 82–86. find
  41. Ward, N. G. & Ortega, C. A. (2024), Preliminaries to a study of the pragmatic functions of reduced articulatory precision in dialog, Technical Report UTEP-CS-24-XX, University of Texas at El Paso. find
  42. Zellers, M. (2017), ‘Prosodic variation and segmental reduction and their roles in cuing turn transition in Swedish’, Language and Speech pp. 454–478. ↗
  43. Zellers, M., Schuppler, B. & Clayards, M. (2018), Introduction, or: Why rethink reduction?, in F. Cangemi, M. Clayards, O. Niebuhr, B. Schuppler & M. Zellers, eds, ‘Rethinking reduction’, de Gruyter, pp. 1–24. *Figure 4: Feature correlations with reduction for English* *Figure 5: Feature correlations with reduction for Spanish* find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/ward-2024-topics-study-pragmatic-functions-phonetic/