Elamur Research⌕
☰
home / papers / 2016 / ZeuScansion: A tool for scansion of English poetry

ZeuScansion: A tool for scansion of English poetry

Paper · 2016 · Manex Agirrezabal, Aitzol Astigarraga, Bertol Arrieta et al.

Summary

ZeuScansion automates scansion of English verse. Scansion means marking which syllables of a line carry stress, grouping them into metrical feet, and naming the overall meter. The authors build the system as a chain of finite-state transducers that tokenizes the line, tags parts of speech, looks up dictionary stress, and guesses stress for unknown words by comparing them with similarly spelled known words. Dictionary stress alone often contradicts the rhythm actually heard in a line. To handle this, the system applies a simplified form of Groves' rules, which remove stress from some lexically stressed syllables. A scoring scheme then splits each line into feet, weighing both how well candidate feet match the stress pattern and how long they are. Finally, the system reports the dominant meter of the whole poem. Evaluated against human-scanned poetry, where several readings can count as correct, ZeuScansion reaches 86.78% per-syllable accuracy. The tool is open-source under GNU GPL, so it gives a reproducible, rule-based baseline for computational metrics.

Key points

  • The pipeline has four finite-state stages (tokenization, POS tagging, stress placement, and stress prediction for out-of-vocabulary words) and marks three stress levels: primary, secondary and unstressed.
  • The system recognizes twelve kinds of feet. Four have two syllables (pyrrhus, iamb, trochee, spondee), and eight have three, namely the molossus, cretic, anapest, dactyl, tribrach, antibacchius, amphibrach and bacchius.
  • On a human-scanned gold standard that allows several valid scansions per line, per-syllable accuracy is 86.78%.
  • At poem level, the tool labels Shakespeare's Sonnets as iambic pentameter and Longfellow's The Song of Hiawatha as trochaic tetrameter.
  • The code is released on GitHub in two repositories, zeuscansion and a separate stress guesser called athenarhythm. The authors contrast their tool with Scandroid, which forces every line into either iambic or anapestic meter.

Interesting ideas

  • I1. The system first assigns each word its fixed dictionary stress, using POS to tell apart pairs like the noun and verb "record", and only then lets explicit context rules demote stress, so each change has a named cause.
  • I2. Unknown words get their stress from known words with similar spelling, on the assumption that similar spelling predicts a similar stress pattern, and this guesser is a separate, swappable module.
  • I3. The gold standard accepts several valid scansions for one line, because experts disagree on ambiguous lines.
  • I4. Meter is decided over the whole poem by looking for the dominant foot pattern, so individual lines can deviate without changing the poem's overall label.

For Mova

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

  • I1 (tokens): Lexical features in crate en could follow the same two-layer design, with a lexicon default plus logged context overrides, so every final tag or feature can be traced to the rule that changed it.
  • I2 (tokens): Unknown-word morphology and lemma guessing could borrow the analogy approach and show the nearest known neighbour as the explanation, while keeping the guesser pluggable so a stronger one can replace it.
  • I3 (learn): Evaluation could score against a set of acceptable outputs, for example equally valid attachments or role labels, so the system is not penalised for defensible alternatives and real errors stand out.
  • I4 (back): When generating verse or rhythmic text, a generated passage could be checked for a global target meter while tolerating some line-level variation, and the deviations could be reported as explicit, explainable differences.

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

🎓University of the Basque CountryLeioa · Spain40 works · 16,234 citations🎓University of Colorado BoulderBoulder · United States52 works · 49,647 citations🎓University of HelsinkiHelsinki · Finland33 works · 4,705 citations

📚 Shelves 5 Computer Science · 9,086 works Artificial Intelligence · 8,889 works Arts and Humanities · 9,296 works Language and Linguistics · 9,206 works Literature and Literary Theory · 340 works

🧭 Topics 1 English · 1,373 works

Cite

Cite as Manex Agirrezabal, Aitzol Astigarraga, Bertol Arrieta and Mans Hulden. 2016. ZeuScansion: A tool for scansion of English poetry. Journal of Language Modelling.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.15398/jlm.v4i1.102

@article{agirrezabal2016zeuscansion,
  title = {{ZeuScansion: A tool for scansion of English poetry}},
  author = {Manex Agirrezabal, Aitzol Astigarraga, Bertol Arrieta and Mans Hulden},
  journal = {Journal of Language Modelling},
  year = {2016},
  doi = {10.15398/jlm.v4i1.102},
  url = {https://doi.org/10.15398/jlm.v4i1.102}
}

References

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

  1. Manex Agirrezabal, Bertol Arrieta, Aitzol Astigarraga, and Mans Hulden (2013), POS-tag based poetry generation with WordNet, Proceedings of the 2013 European Workshop on Natural Language Generation, pp. 162–166. find
  2. Manex Agirrezabal, Jeffrey Heinz, Mans Hulden, and Bertol Arrieta (2014), Assigning stress to out-of-vocabulary words: three approaches, International Conference on Artificial Intelligence, 27:105–110. find
  3. Wystan H. Auden (1960), The more loving one, https://www.poets.org/poetsorg/poem/more-loving-one . ↗
  4. Kenneth R. Beesley and Lauri Karttunen (2003), Finite-state morphology: Xerox tools and techniques, CSLI. find
  5. Terry V. F. Brogan (1981), English versification, 1570-1980: a reference guide with a global appendix, Johns Hopkins University Press. find
  6. Lewis Carroll (1869), Phantasmagoria and Other Poems, London : Macmillan, https://archive.org/details/phantasmagoriaot00carrrich . ↗
  7. Lewis Carroll (1916), Alice’s Adventures in Wonderland and Through the Looking Glass, George W. Jacobs & Company, Philadelphia. find
  8. Chih-Chung Chang and Chih-Jen Lin (2011), LIBSVM: A library for support vector machines, ACM Transactions on Intelligent Systems and Technology, 2:27:1–27:27, software available at http://www.csie.ntu.edu.tw/~cjlin/libsvm . ↗
  9. Noam Chomsky and Morris Halle (1968), The sound pattern of English, Harper & Row. find
  10. Alfred Corn (1997), The poem’s heartbeat: a manual of prosody, Copper Canyon Press. find
  11. Rong-En Fan, Kai-Wei Chang, Cho-Jui Hsieh, Xiang-Rui Wang, and Chih-Jen Lin (2008), LIBLINEAR: A library for large linear classification, The Journal of Machine Learning Research, 9:1871–1874, software available at http://www.csie.ntu.edu.tw/~cjlin/liblinear . ↗
  12. Robert Frost (1979), The poetry of Robert Frost: the collected poems, complete and unabridged, Henry Holt and Company. find
  13. Paul Fussell (1965), Poetic meter and poetic form, McGraw Hill. find
  14. Natalie Gerber (2013), Stress-based metrics revisited: a comparative exercise in scansion systems and their implications for iambic pentameter, Thinking Verse, III:131–168. find
  15. Erica Greene, Tugba Bodrumlu, and Kevin Knight (2010), Automatic analysis of rhythmic poetry with applications to generation and translation, in Proceedings of the 2010 Conference on Empirical Methods in Natural Language Processing, pp. 524–533. find
  16. Peter L. Groves (1998), Strange music: the metre of the English heroic line, volume 74 of ELS Monograph Series, English Literary Studies, University of Victoria. find
  17. Péter Halácsy, András Kornai, and Csaba Oravecz (2007), HunPos: an open source trigram tagger, in Proceedings of the 45th Annual Meeting of the Association for Computational Linguistics (Interactive Poster and Demonstration Sessions), pp. 209–212. find
  18. Morris Halle and Jean-Roger Vergnaud (1987), An essay on stress, MIT Press. find
  19. Michael Hart (1971), Project Gutenberg, https://www.gutenberg.org/ . ↗
  20. Charles O. Hartman (1996), Virtual muse: experiments in computer poetry, Wesleyan University Press. find
  21. Charles O. Hartman (2005), The Scandroid 1.1, http://oak.conncoll.edu/cohar/Programs.htm . ↗
  22. Mans Hulden (2006), Finite-state syllabification, in Anssi Yli-Jyrä, Lauri Karttunen, and Juhani Karhumäki, editors, Finite-State Methods and Natural Language Processing, volume 4002 of Lecture Notes in Computer Science, pp. 86–96, Springer. find
  23. Mans Hulden (2009), Foma: a finite-state compiler and library, in Proceedings of the 12th Conference of the European Chapter of the Association for Computational Linguistics (Demonstrations Session), pp. 29–32. find
  24. Harry M. Logan (1988), Computer analysis of sound and meter in poetry, College Literature, 15(1):19–24. find
  25. Henry W. Longfellow (1855), The Song of Hiawatha, David Bogue, https://archive.org/details/songhiawathathe00longrich . ↗
  26. Mitchell P. Marcus, Mary Ann Marcinkiewicz, and Beatrice Santorini (1993), Building a large annotated corpus of English: the Penn Treebank, Computational Linguistics, 19(2):313–330. ↗ our card
  27. William G. M. McAleese (2007), Improving scansion with syntax: an investigation into the effectiveness of a syntactic analysis of poetry by computer using phonological scansion theory, Technical report, Department of Computing Faculty of Mathematics, Computing and Technology The Open University. ↗
  28. Harriet Monroe (1917), The new poetry: an anthology, The Macmillan Company, https://archive.org/details/newpoetryananth01hendgoog . ↗
  29. Augustus T. Murray (1925), Homer: The Iliad, Heinemann, https://archive.org/details/iliadmurray01homeuoft . ↗
  30. Josef Novak, Nobuaki Minematsu, and Keikichi Hirose (2012), WFST-based grapheme-to-phoneme conversion: open source tools for alignment, model-building and decoding, in Proceedings of the 10th Internaltional Workshop on Finite-State Methods and Natural Language Processing, pp. 45–49. find
  31. Marc R. Plamondon (2006), Virtual verse analysis: analysing patterns in poetry, Literary and Linguistic Computing, 21(1):127–141. find
  32. Margaret Robertson, editor (2007), Poems published in 1820 by John Keats, Project Gutenberg, http://www.gutenberg.org/ebooks/23684 . ↗
  33. Emmanuel Roche and Yves Schabes (1995), Deterministic part-of-speech tagging with finite-state transducers, Computational Linguistics, 21(2):227–253. find
  34. David E. Rumelhart and James L. McClelland (1985), On learning the past tenses of English verbs, Technical report, Institute for Cognitive Science, University of California, San Diego. find
  35. Terrence J. Sejnowski and Charles R. Rosenberg (1987), Parallel networks that learn to pronounce English text, Complex Systems, 1(1):145–168. find
  36. William Shakespeare (1806), Romeo and Juliet, John Cawthorn. find
  37. William Shakespeare (1904), The tragedy of Hamlet, Cambridge University Press. find
  38. William Shakespeare (2011), Shakespeare’s sonnets, Project Gutenberg, https://archive.org/details/shakespearessonn01041gut . ↗
  39. Timothy Steele (1999), All the fun’s in how you say a thing: an explanation of meter and versification, Ohio University Press. find
  40. Wallace Stevens (1923), Harmonium, Academy of American Poets. find
  41. Herbert F. Tucker (2011), Poetic data and the news from poems: a *For Better for Verse* memoir, Victorian Poetry, 49(2):267–281. ↗
  42. Robert L. Weide (1998), The CMU pronunciation dictionary, release 0.6, Carnegie Mellon University. This work is licensed under the Creative Commons Attribution 3.0 Unported License. http://creativecommons.org/licenses/by/3.0/ ↗

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/agirrezabal-2016-zeuscansion-tool-scansion-english-poetry/