Elamur Research⌕
☰
home / papers / 2026 / Pop Lyrics through Time: Challenges in Corpus-Based Modeling of Linguistic and Emotional Dynamics in German Pop Lyrics

Pop Lyrics through Time: Challenges in Corpus-Based Modeling of Linguistic and Emotional Dynamics in German Pop Lyrics

Paper · 2026 · Roman Schneider

Summary

German song lyrics blend spoken-style and written-style traits, which makes them a difficult domain for standard NLP tools. This paper uses Songkorpus, a TEI-encoded, multi-layer annotated monitoring corpus of German lyrics, to trace how structural and emotional properties of songs have shifted since 1970. The author builds reproducible pipelines that compute four song-level features: lexical diversity, a subjectivity score based on first- and second-person pronouns, modal particle density, and a sentiment intensity value normalized by text length.

Modal particles and sentiment are especially error-prone, so both are validated. For modal particles, a hand-annotated gold standard is used to train and test sequence taggers, and their scores are compared with human agreement. Lexicon-based sentiment is checked against judgments from a human annotation study. Correlation and regression analyses then relate the features to each other and track them over time.

Lexical diversity and sentiment intensity decline, while personal pronouns rise slightly. The work offers a methodological template for culture-specific, non-standard corpora, and its evaluation materials are shared.

Key points

  • Resource: Songkorpus (Release 7.0, songkorpus.de) contains more than 15,000 lyrics and about 5 million tokens covering over six decades. It is TEI-encoded and annotated with lemmata, STTS POS tags, named entities, neologisms, constituents, verse structure and, in some cases, rhyme. It is updated annually, including a growing "Deutschrap" archive.
  • Scope: The analysis covers songs from 1970 onward. Only segments marked as German through TEI language attributes are used, which filters out code-switched English and other passages.
  • Features: MATTR and MTLD measure lexical diversity. The other features are pronoun-based subjectivity, modal particle density and length-normalized sentiment intensity.
  • Validation (RQ1):
    • Modal particles are tagged by sequence models trained on a manually curated gold standard; the resource list cites CRF++. Model performance is related to inter-annotator agreement.
    • Sentiment scores from a lexicon-based resource are compared against a human rating experiment.
  • Findings (RQ2–RQ3): Over time, lexical diversity and sentiment intensity decrease, and first- and second-person pronoun use increases slightly. The author frames these results as exploratory rather than definitive evidence of cultural change.

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

🔬Leibniz Institute for the German LanguageMannheim · Germany5 works · 22 citations

 Fields, topics and concepts

📚 Shelves 7The 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 worksSocial Sciences · 1,772 worksGeneral Social Sciences · 227 worksLiterature and Literary Theory · 519 works

🧭 Topics 1Research subjects and languages the work is about.Sentiment & emotion · 972 works

💡 Concepts 1Key ideas and methods explained on our concept pages.Type–token ratio · 14 works

Cite

Cite as Roman Schneider. 2026. Pop Lyrics through Time: Challenges in Corpus-Based Modeling of Linguistic and Emotional Dynamics in German Pop Lyrics. Proceedings of the 12th Workshop on Challenges in the Management of Large Corpora.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.63317/4dh6sfnuqcnr

@inproceedings{schneider2026pop,
  title = {{Pop Lyrics through Time: Challenges in Corpus-Based Modeling of Linguistic and Emotional Dynamics in German Pop Lyrics}},
  author = {Roman Schneider},
  booktitle = {Proceedings of the 12th Workshop on Challenges in the Management of Large Corpora},
  year = {2026},
  doi = {10.63317/4dh6sfnuqcnr},
  url = {https://doi.org/10.63317/4dh6sfnuqcnr}
}

References

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

  1. Abercrombie, G., Rieser, V., and Hovy, D. (2023). Natural Language Processing with Intra-Annotator Agreement. ArXiv, Preprint 2301.10684. https://arxiv.org/pdf/2301.10684 ↗
  2. Bello, A., Ng, S.-C., and Leung, M.-F. (2023). A BERT Framework to Sentiment Analysis of Tweets. Sensors, 23(1), 506. https://doi.org/10.3390/s23010506 ↗
  3. Bestgen, Y. (2024). Measuring Lexical Diversity in Texts: The Twofold Length Problem. Language Learning, 74: 638–671. https://doi.org/10.1111/lang.12630 ↗
  4. Blühdorn, H. (2019). Modalpartikeln und Akzent im Deutschen. Linguistische Berichte, 259. Hamburg: Buske. 275–317. https://nbn-resolving.org/urn:nbn:de:bsz:mh39-91746 ↗
  5. Broll, S. and Schneider, R. (2023). Empirische Verortung konzeptioneller Nähe/Mündlichkeit inner- und außerhalb schriftsprachlicher Korpora. Journal for Language Technology and Computational Linguistics, 36(1). 113–150. https://doi.org/10.21248/jlcl.36.2023.240 ↗
  6. Covington, M. and McFall, J. (2010). Cutting the Gordian knot: The moving-average type-token ratio (MATTR). Journal of Quantitative Linguistics, 17, 94–100. https://doi.org/10.1080/09296171003643098 ↗
  7. Dahary, S., Edana, A., Apartsin, A., and Aperstein, Y. (2025). From Joy to Fear: A Benchmark of Emotion Estimation in Pop Song Lyrics. ArXiv Preprint 2509.05617. https://arxiv.org/pdf/2509.05617 ↗
  8. Diewald, G. (2007): Abtönungspartikel. In Hoffmann, L. (Ed.), Handbuch der deutschen Wortarten. Berlin, New York: de Gruyter. 117– 142. https://doi.org/10.1515/9783110217087.117 Emerson, G. and Declerck. ↗
  9. T. (2014). SentiMerge: Combining Sentiment Lexicons in a Bayesian Framework. In Proceedings of Workshop on Lexical and Grammatical Resources for Language Processing, 30–38, Dublin, Ireland: Association for Computational Linguistics and Dublin City University. https://doi.org/10.3115/v1/W14-5805 ↗
  10. Fehle, J., Schmidt, T., and Wolff, C. (2021). Lexicon-based Sentiment Analysis in German: Systematic Evaluation of Resources and Preprocessing Techniques. In Proceedings of the 17th Conference on Natural Language Processing (KONVENS 2021), Düsseldorf. https://doi.org/10.5283/EPUB.50833 ↗
  11. Guhr, O., Schumann, A.-K., Bahrmann, F., and Böhme, H. J. (2020). Training a Broad-Coverage German Sentiment Classification Model for Dialog Systems. In N. Calzolari et al. (Eds.), Proceedings of the 12th Language Resources and Evaluation Conference (LREC), 1627–1632. Marseille: European Language Resources Association (ELRA). https://aclanthology.org/2020.lrec-1.202/ ↗
  12. Hentschel, E. (2010). Partikelprofile literarischer Texte. In T. Harden, E. Hentschel (Eds.), 40 Jahre Partikelforschung. Tübingen: Stauffenburg, 97–118. find
  13. Hunke, T., Huber, F., and Steffens, J. (2025). The Evolution of Song Lyrics: An NLP-Based Analysis of Popular Music in Germany from 1954 to 2022. Music & Science, 8. https://doi.org/10.1177/20592043251331155 ↗
  14. Koch, P. and Oesterreicher, W. (2012). Language of immediacy – Language of distance: Orality and literacy from the perspective of language theory and linguistic history. In C. Lange, B. Weber, and G. Wolf (Eds.), Communicative spaces: Variation, contact, and change, 441– 473. Frankfurt: Lang. https://doi.org/10.15496/publikation-20415 ↗
  15. Kraus, M., Feuerriegel, S. (2019). Sentiment analysis based on rhetorical structure theory: Learning deep neural networks from discourse trees. Expert Systems with Applications, 118, 65-79, https://doi.org/10.1016/j.eswa.2018.10.002 ↗
  16. Kreyer, R. and Mukherjee, J. (2007). The style of pop song lyrics: a corpus-linguistic pilot study. Anglia, 125(1). 31–58. https://doi.org/10.1515/ANGL.2007.31 ↗
  17. Landis, J. R. and Koch, G. G. (1977): The Measurement of Observer Agreement for Categorical Data. Biometrics, 33(1), 159–174. https://doi.org/10.2307/2529310 ↗
  18. Liu, B. (2015). Sentiment Analysis: Mining Opinions, Sentiments, and Emotions. Cambridge, University Press. https://doi.org/10.1017/CBO9781139084789 ↗
  19. Malvern, D., Richards, B., Chipere, N., and Durán, P. (2024). Lexical Diversity and Language Development. Quantification and Assessment. London : Palgrave Macmillan. https://doi.org/10.1057/9780230511804 ↗
  20. Mccarthy, P. and Jarvis, S. (2010). Mtld, vocd-d, and hd-d: A validation study of sophisticated approaches to lexical diversity assessment. Behavior research methods, 42, 381–92. https://doi.org/10.3758/BRM.42.2.381 ↗
  21. Öhman, E. (2021). The Validity of Lexicon-based Sentiment Analysis in Interdisciplinary Research. In Proceedings of the Workshop on Natural Language Processing for Digital Humanities, 7–12, NIT Silchar, India. NLP Association of India (NLPAI). https://aclanthology.org/2021.nlp4dh-1.2/ ↗
  22. Pang, B., Lee, L. (2008). Opinion Mining and Sentiment Analysis. Now Foundations and Trends. https://doi.org/10.1561/1500000011 ↗
  23. Parada-Cabaleiro, E., Mayerl, M., Brandl, S., Skowron, M., Schedl, M., Lex, E., and Zangerle, E. (2024). Song lyrics have become simpler and more repetitive over the last five decades. Scientifc Reports, 14 (5531). https://doi.org/10.1038/s41598-024-55742-x ↗
  24. Schneider, R. (in preparation). Beyond Standard: Intelligent Modeling of Creative Language Using the German Song Corpus. In L. Herzberg, C. Mair, and A. Witt (Eds.), Corpus linguistics 2040: Which data, which methods, which models? Digital Linguistics, 6, Berlin/Boston: De Gruyter. find
  25. Schneider, R. (2026). Linguistic resources for the study of pop culture. In Valentin Werner, Cecilia Cutler, and Andrew Moody (Eds.), Handbook of Language and Pop Culture. Berlin, Boston: De Gruyter Mouton. find
  26. Schneider, R. (2020). A Corpus Linguistic Perspective on Contemporary German Pop Lyrics with the Multi-Layer Annotated “Songkorpus”. In N. Calzolari et al. (Eds.), Proceedings of the 12th Language Resources and Evaluation Conference (LREC), 835–841. Marseille: European Language Resources Association (ELRA). https://aclanthology.org/2020.lrec-1.105/ ↗
  27. Schoonjans, S. (2018). Modalpartikeln als multimodale Konstruktionen. Berlin/Boston: deGruyter. https://doi.org/10.1515/9783110566260 ↗
  28. Scott, M. (2004). WordSmith Tools Version 4.0. Oxford: Oxford University Press. find
  29. Storø, S. R. (2022). Die Annotation der Modalpartikeln im GeWiss-Korpus. Eine syntaktische und semantisch-pragmatische Analyse der PTKMA-Annotation. Deutsche Sprache. Zeitschriften für Theorie, Praxis und Dokumentation, 22 (2), 124–149. https://doi.org/10.1515/ds-2022-0014 ↗
  30. Sluyter-Gäthje, H. and Trilcke, P. (2022). Poesie als Fehler. Ein 'Tool Misuse'-Experiment zur Prozessierung von Lyrik. In Proceedings 8. Tagung des Verbands Digital Humanities im deutschsprachigen Raum (DHd), Potsdam. https://doi.org/10.5281/zenodo.6328201 ↗
  31. Trippel, T. (2025). Metadata for research data. In P. Bański, U. Heid, and L. Herzberg (Eds.), Harmonizing language data: Standards for linguistic resources. Digital Linguistics, 4, Berlin/Boston: De Gruyter 251–279. https://doi.org/10.1515/9783112208212-011 ↗
  32. Werner, V., Hiramoto, M., and Flanagan, P. (2025). Language and pop culture. Setting the agenda, Journal of Language and Pop Culture, 1(1). 1–17. https://doi.org/10.1075/jlpop.24034.wer 6. ↗
  33. Language Resource References Kudo, T. (2005). CRF++: Yet Another CRF Toolkit. http://taku910.github.io/crfpp/ ↗
  34. Schneider, R. (2025). Songkorpus - Linguistic Corpus of German Song Lyrics. Release 7.0. https://songkorpus.de ↗

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/schneider-2026-pop-lyrics-through-time-challenges/