Elamur Research⌕
☰
home / papers / 2009 / Frequency Matters: Pitch Accents and Information Status

Frequency Matters: Pitch Accents and Information Status

Paper · 2009 · Katrin Schweitzer, Michael Walsh, Bernd Möbius et al.

Summary

The paper examines whether a word's information status (given or new) and the frequency of a pitch accent category affect how consistently speakers produce tokens of that accent. The data are German radio news from the IMS Radio News Corpus, manually labelled with GToBI(S) accents and annotated for information status from the text alone using Riester's taxonomy. Each accent token's F0 contour is described by the six parameters of the PaIntE model, then z-scored. Within-type similarity is the cosine between pairs of token vectors, and the resulting similarity distributions are compared statistically. The analysis is framed by Exemplar Theory, which offers two competing expectations: more stored exemplars could produce more variation, or entrenchment could make frequent behaviour more uniform. The results show that rarer accent types (H∗L compared with L∗H) and rarer combinations of accent and information status have more similar tokens. Information status adds to the frequency effect. For speech technology, this means frequent accents do not map neatly onto one F0 shape.

Key points

  • After the 2.5% upper and lower outliers were removed for each PaIntE parameter, 1021 L∗H, 571 H∗L and 134 H∗ tokens remained, giving a frequency continuum from high to low.
  • The six PaIntE parameters are rise and fall gradients (a1, a2), alignment (b), rise and fall amplitudes (c1, c2) and peak height (d). For H∗L and H∗, the fitting window now includes the preaccented syllable.
  • The corpus has three speakers, and 77.92% of the analysed data comes from the male speaker. Information status was labelled independently by two annotators, with a third, more experienced annotator resolving conflicts. Only unambiguous labels were used.
  • Pairwise distribution comparisons used a conservative threshold of p < 0.017. Given and new tokens differed significantly for both L∗H and H∗L.
  • Ordered by increasing similarity, the combinations were given L∗H (114 tokens) and new L∗H (102), then new H∗L (87), then given H∗L (44). H∗ had too few tokens to interpret (10 given, 21 new).

Interesting ideas

  • I1. The paper quantifies how uniform a category's realisations are by using the spread of pairwise cosine similarities among token feature vectors, then compares this spread across categories of different frequency.
  • I2. Each label was assigned by two independent annotators, and a more experienced third party decided only where they disagreed.
  • I3. Ambiguous cases keep multiple information-status labels as underspecification, but all measurements use only the unambiguous cases.
  • I4. Information status is assigned from text alone, according to where a presupposition is resolved (discourse, environment, or encyclopaedic context).
  • I5. Feature distributions are trimmed at the 2.5% tails before comparison, and a stricter significance level is used because many pairwise tests are run.

For Mova

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

  • I1 (learn): Pairwise similarity distributions over parser feature vectors for each relation label could flag frequent labels that are internally heterogeneous and may need splitting or repair rules.
  • I2 (learn): Silver data could follow the same pattern, with witness parsers labelling independently and an LLM teacher tier called only to resolve their disagreements.
  • I3 (frames): Coreference and givenness tags could store several candidate labels when a mention is ambiguous, with evaluation restricted to single-label mentions.
  • I4 (world): Each entity in the text world could record which source licenses it, such as earlier mention, situation, or world knowledge, which would give an explainable basis for given/new decisions.
  • I5 (learn): Evaluations that compare many systems or label groups could trim extremes first and correct for multiple comparisons, reducing spurious "improvements".

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

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Arts and Humanities · 17,450 worksLanguage and Linguistics · 17,296 worksPsychology · 1,649 worksExperimental and Cognitive Psychology · 930 works

Cite

Cite as Katrin Schweitzer, Michael Walsh, Bernd Möbius, Arndt Riester, Antje Schweitzer and Hinrich Schütze. 2009. Frequency Matters: Pitch Accents and Information Status. Proceedings of the 12th Conference of the European Chapter of the ACL (EACL 2009).

The paper itself, not this page. BibTeX from the library's record of the work.

@inproceedings{schweitzer2009frequency,
  title = {{Frequency Matters: Pitch Accents and Information Status}},
  author = {Katrin Schweitzer, Michael Walsh, Bernd Möbius, Arndt Riester, Antje Schweitzer and Hinrich Schütze},
  booktitle = {Proceedings of the 12th Conference of the European Chapter of the ACL (EACL 2009)},
  year = {2009},
  url = {https://aclanthology.org/E09-1083/}
}

References

29 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Kirsten Abbot-Smith and Michael Tomasello. 2006. Exemplar-learning and schematization in a usage-based account of syntactic acquisition. The Linguistic Review, 23(3):275–290. ↗
  2. Ellen G. Bard and M. P. Aylett. 1999. The dissociation of deaccenting, givenness, and syntactic role in spontaneous speech. In Proceedings of ICPhS (San Francisco), volume 3, pages 1753–1756. find
  3. Stefan Baumann. 2006. The Intonation of Givenness – Evidence from German., volume 508 of Linguistische Arbeiten. Niemeyer, Tübingen. Ph.D. thesis, Saarland University. find
  4. Gillian Brown. 1983. Prosodic structure and the given/new distinction. In Anne Cutler and D. Robert Ladd, editors, Prosody: Models and Measurements, pages 67–77. Springer, New York. ↗
  5. Joan Bybee. 2006. From usage to grammar: The mind’s response to repetition. Language, 84:529– 551. ↗
  6. Karen Croot and Kathleen Rastle. 2004. Is there a syllabary containing stored articulatory plans for speech production in English? In Proceedings of the 10th Australian International Conference on Speech Science and Technology (Sydney), pages 376–381. find
  7. Michael A. K. Halliday. 1967. Intonation and Grammar in British English. Mouton, The Hague. ↗
  8. Keith Johnson. 1997. Speech perception without speaker normalization: An exemplar model. In K. Johnson and J. W. Mullennix, editors, Talker Variability in Speech Processing, pages 145–165. Academic Press, San Diego. find
  9. Klaus J. Kohler. 1991. Studies in german intonation. AIPUK (Univ. Kiel), 25. find
  10. Frank Kügler and Caroline Féry. 2008. Pitch accent scaling on given, new and focused constituents in german. Journal of Phonetics. ↗
  11. Erwin Marsi, Martin Reynaert, Antal van den Bosch, Walter Daelemans, and Véronique Hoste. 2003. Learning to predict pitch accents and prosodic boundaries in dutch. In Proceedings of the ACL-2003 Conference (Sapporo, Japan), pages 489–496. find
  12. Jörg Mayer. 1995. Transcribing German Intonation – The Stuttgart System. Technical report, Universität Stuttgart. http://www.ims.uni-stuttgart.de/phonetik/joerg/labman/STGTsystem.html. ↗
  13. Gregor Möhler and Alistair Conkie. 1998. Parametric modeling of intonation using vector quantization. In Third Intern. Workshop on Speech Synth (Jenolan Caves), pages 311–316. find
  14. Gregor Möhler. 1998. Describing intonation with a parametric model. In Proceedings ICSLP, volume 7, pages 2851–2854. find
  15. T. V. Pernegger. 1998. What’s wrong with Bonferroni adjustment. British Medical Journal, 316:1236– 1238. find
  16. Janet Pierrehumbert and Julia Hirschberg. 1990. The meaning of intonational contours in the interpretation of discourse. In P. R. Cohen, J. Morgan, and M. E. Pollack, editors, Intentions in Communication, pages 271–311. MIT Press, Cambridge. ↗
  17. Janet Pierrehumbert. 2001. Exemplar dynamics: Word frequency, lenition and contrast. In Joan Bybee and Paul Hopper, editors, Frequency and the Emergence of Linguistic Structure, pages 137–157. Amsterdam. find
  18. Ellen F. Prince. 1992. The ZPG Letter: Subjects, Definiteness and Information Status. In W. C. Mann and S. A. Thompson, editors, Discourse Description: Diverse Linguistic Analyses of a Fund-Raising Text, pages 295–325. Amsterdam. find
  19. Stefan Rapp. 1998. Automatisierte Erstellung von Korpora für die Prosodieforschung. Ph.D. thesis, IMS, Universität Stuttgart. AIMS 4 (1). find
  20. Arndt Riester. 2008. A Semantic Explication of Information Status and the Underspecification of the Recipients’ Knowledge. In Atle Grønn, editor, Proceedings of Sinn und Bedeutung 12, Oslo. find
  21. Antje Schweitzer and Bernd Möbius. 2004. Exemplar-based production of prosody: Evidence from segment and syllable durations. In Speech Prosody 2004 (Nara, Japan), pages 459–462. find
  22. Katrin Schweitzer, Arndt Riester, Hans Kamp, and Grzegorz Dogil. 2008. Phonological and acoustic specification of information status - a semantic and phonetic analysis. Poster at ”Experimental and Theoretical Advances in Prosody”, Cornell University. find
  23. Kim Silverman, Mary Backman, John Pitrelli, Mari Ostendorf, Colin Wightman, Patti Price, Janet Pierrehumbert, and Julia Hirschberg. 1992. Tobi: A standard for Labeling English Prosody. In Proceedings of ICSLP (Banff, Kanada), volume 2, pages 867–870, Banff, Canada. find
  24. Jacques Terken and Julia Hirschberg. 1994. Deaccentuation of words representing ‘given’ information: effects of persistence of grammatical function and surface position. Language and Speech, 37:125– 145. ↗
  25. Jan P. H. van Santen and Bernd Möbius. 2000. A quantitative model of F0 generation and alignment. In A. Botinis, editor, Intonation—Analysis, Modelling and Technology, pages 269–288. Kluwer. find
  26. Michael Walsh, Hinrich Schütze, Bernd Möbius, and Antje Schweitzer. 2007. An exemplar-theoretic account of syllable frequency effects. In Proceedings of ICPhS (Saarbrücken), pages 481–484. find
  27. Michael Walsh, Katrin Schweitzer, Bernd Möbius, and Hinrich Schütze. 2008. Examining pitch-accent variability from an exemplar-theoretic perspective. In Proceedings of Interspeech 2008 (Brisbane). find
  28. Sandra P. Whiteside and Rosemary A. Varley. 1998. Dual-route phonetic encoding: Some acoustic evidence. In Proceedings of ICSLP (Sydney), volume 7, pages 3155–3158. find
  29. George Yule. 1980. Intonation and Givenness in Spoken Discourse. Studies in Language, pages 271– 286. ↗

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/schweitzer-2009-frequency-matters-pitch-accents-information/