The Icelandic Parsed Historical Corpus (IcePaHC)

What it is about

IcePaHC is a hand-corrected parsed corpus of about one million words of Icelandic. It samples 60 texts from the late 12th century to the 2000s, about 100,000 words per century, mostly narrative and religious prose. Icelandic inflection has hardly changed over that time, so one annotation scheme covers all ten centuries. All texts were first converted to modern spelling, which let a modern tagger and lemmatiser pre-annotate them. The scheme is the Penn historical-corpora bracketing used for Old, Middle and Early Modern English, with lemmas and case added. As a result, the same CorpusSearch queries run on Icelandic and on historical English, and the paper shows the opening of the Gospel of John parsed in Icelandic, Faroese and Tyndale's English side by side. Humans marked the clause boundaries before full annotation. The authors admit that much of the annotation was not double-checked, so the corpus is meant for quantitative work; one release had 51 clauses with two direct objects out of 4,727 double-object clauses, and the next release fixes them through structural well-formedness checks. A Faroese spin-off (FarPaHC, 250,000 words planned) starts with parallel Bible texts. The paper ends with ten "user freedoms" for language resources.

Ideas

• I1. Parallel Bible trees as a cross-period test: the Gospel of John and Acts exist in Penn-style parses for Early Modern English, Icelandic and Faroese. Run Mova's parser on the KJV and Tyndale verses of those books and check the UD trees against a Penn-to-UD conversion of the PPCEME verses (evaluation only, if the licence allows). The result would be a per-verse LAS on biblical English with a gold tree, which 12. How Mova reads this English today: pipeline, fixes, failures does not have.
• I2. Copy the "structural constraint" error check: write treeagree check gates that a well-formed UD tree must pass. Examples: no clause with two obj, no iobj without an obj, no verb with two nsubj. Count the violations on the school silver and on the Torah parses. The paper finds about 1% of double-object clauses mislabelled this way, and Mova's numbers can be compared with that.
• I3. Normalise spelling first, then annotate: IcePaHC's choice to modernise spelling so that a modern tagger works is the same bet as en/src/emode.rs. The paper does not measure what this costs (lost case endings and spelling evidence). Mova can measure it on a KJV sample: tagging accuracy with and without normalisation, run twice.
• I4. Keep the original form next to the normalised one, as IcePaHC's dash-separated lemma does (word-lemma), so that the roundtrip can restore the source spelling (tools/roundtrip --exact).

For Mova

1. The period, the register and the texts Mova reads (corpora), 02-spelling-and-tokens.md (normalisation) and 12-mova-pipeline.md (biblical test sets). Code: tools/treeagree (gate checks, I2), en/src/emode.rs (I3), tools/roundtrip (I4). Measures: LAS on parallel Gospel verses; the count of double-obj and orphan-iobj violations per 1,000 clauses.

Cite

The paper itself, not this card. BibTeX from the library's record of the work. DOI 10.63317/2icn3f8r7r2r

@inproceedings{rognvaldsson2012icelandic,
  title = {{The Icelandic Parsed Historical Corpus (IcePaHC)}},
  author = {Rögnvaldsson, Eiríkur and Ingason, Anton Karl and Sigurdhsson, Einar Freyr and Wallenberg, Joel},
  booktitle = {Proceedings of the Eighth International Conference on Language Resources and Evaluation (LREC'12)},
  year = {2012},
  doi = {10.63317/2icn3f8r7r2r},
  url = {https://doi.org/10.63317/2icn3f8r7r2r}
}

References

28 references, 2 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Beck, J.E. (2011). Penn Parsed Corpora of Historical Greek (PPCHiG). (http://www.ling.upenn.edu/~janabeck/greek-corpora.html). ↗
2. Beck, J.E.; Ecay, A. and Ingason, A.K. (2011). Annotald, version 11.11. [Software for treebank annotation.] (http://github.com/janabeck/Annotald). ↗
3. Bernharðsson, H. (1999). Málblöndun í sautjándu aldar uppskriftum íslenskra miðaldahandrita. Reykjavík: Institute of Linguistics, University of Iceland. find
4. Bick, E. (2003). Arboretum, a Hybrid Treebank for Danish. In J. Nivre and E. Hinrichs (Eds.), TLT 2003. Proceedings of the Second Workshop on Treebanks and Linguistic Theories, 14-15 November 2003, Växjö, Sweden. Växjö: Växjö University Press, pp. 9--20. find
5. Galves, C., Britto, H.( 2002). The Tycho Brahe Corpus of Historical Portuguese. Department of Linguistics, University of Campinas. Online publication, first edition. (http://www.tycho.iel.unicamp.br/~tycho/). ↗
6. Ingason, A.K.; Helgadóttir, S.; Loftsson, H. and Rögnvaldsson, E. (2008). A Mixed Method Lemmatization Algorithm Using a Hierarchy of Linguistic Identities (HOLI). In A. Raante and B. Nordström (Eds.), Advances in Natural Language Processing. (Lecture Notes in Computer Science, Vol. 5221.) Berlin: Springer, pp. 205--216. find
7. Ingason, A.K.; Sigurðsson, E.F. and Wallenberg, J. (2011). Distinguishing Change and Stability: a Quantitative Study of Icelandic Oblique Subjects. Paper presented at DiGS 13, University of Pennsylvania, Philadelphia, June 3rd, 2011. find
8. Kroch, A.; Santorini, B. and Delfs, L. (2004). The Penn-Helsinki Parsed Corpus of Early Modern English (PPCEME). Department of Linguistics, University of Pennsylvania. CD-ROM, first edition, (http://www.ling.upenn.edu/hist-corpora/). ↗
9. Kroch, A., Taylor, A. (2000). The Penn-Helsinki Parsed Corpus of Middle English (PPCME2). Department of Linguistics, University of Pennsylvania. CD-ROM, second edition, (http://www.ling.upenn.edu/hist-cor-pora/). ↗
10. Kromann, M.T. (2003). The Danish Dependency Treebank and the DTAG Treebank Tool. In J. Nivre and E. Hinrichs (Eds.), TLT 2003. Proceedings of the Second Workshop on Treebanks and Linguistic Theories, 14-15 November 2003, Växjö, Sweden. Växjö: Växjö University Press, pp. 217--220. find
11. Light, C. (2010). Parsed Corpus of Early New High German. (http://enhgcorpus.wikispaces.com/home). ↗
12. Light, C., Wallenberg, J. (2011). On the Use of Passives across Germanic. Paper presented at DiGS 13, University of Pennsylvania, Philadelphia, June 4th, 2011. find
13. Loftsson, H. (2008). Tagging Icelandic text: A linguistic rule-based approach. Nordic Journal of Linguistics 31(1), pp. 47--72. ↗
14. Loftsson, H., Rögnvaldsson, E. (2007). IceParser: An Incremental Finite-State Parser for Icelandic. In J. Nivre, H.-J. find
15. Kaalep, K. Muischnek and M. Koit (Eds.), NODALIDA 2007 Conference Proceedings. Tartu: University of Tartu, pp. 128--135. find
16. Martineau, F.; Hirschbühler, P.; Kroch, A. and Morin, Y.C. (2010). Corpus MCVF (parsed corpus), Modéliser le changement: les voies du français, Département de français, University of Ottawa. CD-ROM, first edition (http://www.arts.uottawa.ca/voies/voies_fr.html). ↗
17. Muhonen, K., Purtonen, T.K. (2011). Defining the Annotation Scheme of a Treebank: The End-Use Perspective. In Z. Vetulani (Ed.), Human language technologies as a challenge for computer science and linguistics. Proceedings of the 5th Language and Technology Conference (LTC 2011), November 25-27, 2011. Poznań, pp. 309--313. find
18. Nivre, J.; Nilsson, J. and Hall, J. (2006). Talbanken05: A Swedish Treebank with Phrase Structure and Dependency Annotation. In Proceedings of the Fifth International Conference on Language Resources and Evaluation (LREC). Genoa, pp. 1392--1395. find
19. Palmer, D.D., Hearst, M.A. (1994). Adaptive sentence boundary disambiguation. In Proceedings of the fourth conference on Applied natural language processing. Stroudsburg, PA: Association for Computational Linguistics, pp. 78--83. find our card
20. Randall, B. (2005). CorpusSearch 2 Users Guide. University of Pennsylvania, Philadelphia. (http:// corpussearch.sourceforge.net/CS-manual/ Contents.html). find
21. Rögnvaldsson, E., Helgadóttir, S. (2011). Morphosyntactic Tagging of Old Icelandic Texts and Its Use in Studying Syntactic Variation and Change. In C. Sporleder, A.P.J van den Bosch and K.A. Zervanou (Eds.), Language Technology for Cultural Heritage: Selected Papers from the LaTeCH Workshop series. Berlin: Springer, pp. 63--76. ↗
22. Rögnvaldsson, E.; Ingason, A.K. and Sigurðsson, E.F. (2011). Coping with Variation in the Icelandic Parsed Historical Corpus (IcePaHC). In J.B. Johannessen (Ed.), Language Variation Infrastructure. Papers on selected projects. Oslo Studies in Language 3.2. Oslo: University of Oslo, pp. 97--111. find our card
23. Rögnvaldsson, E.; Ingason, A.K.; Sigurðsson, E.F and Wallenberg, J. (2011). Creating a Dual-Purpose Treebank. In Proceedings of the ACRH Workshop, Heidelberg, 5 Jan. 2012. Journal for Language Technology and Computational Linguistics 26(2), pp. 141--152. ↗
24. Santorini, B. (1997/2008). The Penn Yiddish Corpus. University of Pennsylvania. For details, contact: beatrice@babel.ling.upenn.edu. find
25. Santorini, B. (2010). Annotation manual for the Penn historical corpora and the PCEEC. University of Pennsylvania, Philadelphia. (http://www.ling.upenn.edu/hist-corpora/annotation/index.html). ↗
26. Sapp, C. (2011). A Relative Pronoun in Old Norse? Paper presented at DiGS 13, University of Pennsylvania, Philadelphia, June 5th, 2011. find
27. Taylor, A.; Warner, A.; Pintzuk, S. and Beths, F. (2003). The York-Toronto-Helsinki Parsed Corpus of Old English Prose. University of York. (http://www.users.york.ac.uk/~lang22/YcoeHome1.htm). ↗
28. Wallenberg, J.; Ingason, A.K.; Sigurðsson, E.F. and Rögnvaldsson, E. (2011). Icelandic Parsed Historical Corpus (IcePaHC). Version 0.9. (http://www.linguist.is/icelandic_treebank). ↗

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/rognvaldsson-2012-icelandic-parsed-historical-corpus-icepahc/