Elamur Research⌕
☰
home / papers / 2012 / The Rocky Road towards a Swedish FrameNet - Creating SweFN

The Rocky Road towards a Swedish FrameNet - Creating SweFN

Written for the Research library (claude-opus-5-5) from the paper text, in our own words: the license "CC BY-NC-SA 3.0 (ACL Anthology, before 2016)" does not allow redistributing the text itself. Original — aclanthology.org/L12-1135.pdf

Summary

The paper reports on building SweFN (Swedish FrameNet, Svenskt FrasNät), a free lexical resource based on frame semantics and modelled on the Berkeley FrameNet. SweFN is designed to be useful to people as well as to language technology applications, for example generating and understanding text and extracting information from it. Its long-term target is 50,000 entries with morphological, syntactic and semantic information. The authors reuse Berkeley frames, frame names and frame elements, but every lexical unit must be linked to SALDO, a Swedish association lexicon. They also add domain labels and SIMPLE semantic types to the frames.

The paper discusses three main difficulties:

  • Mismatched sense inventories. SALDO and the frames do not always divide word senses the same way, as the polysemous noun 'brott' shows.
  • Compounds. Swedish compounding is highly productive, so compounds are annotated with the head as the frame-evoking unit and the modifier as a frame element.
  • Language and culture gaps. Some English frames fit Swedish poorly, which required new, split or renamed frames.

Reusing an existing FrameNet speeds up construction and supports cross-lingual uses such as machine translation, but it requires care where the two languages or cultures diverge.

Key points

  • By March 2012, SweFN had 561 frames and more than 18,700 lexical units drawn from SALDO. Over 1,400 new units have been proposed for addition to SALDO.
  • SweFN is released under CC-BY-SA 3.0 and LGPL 3.0. It belongs to the larger SweFN++ lexicon complex and covers general language plus the medical and art domains.
  • The noun 'brott' has seven senses and should evoke at least seven frames, yet SALDO lists only three entries and only two frames currently include it.
  • Annotated compounds, such as those in the Assistance frame (e.g. fattig|hjälp, 'help for the poor'), are meant to provide training data for automatic compound disambiguation. Example sentences are chosen following the GDEX criteria of Kilgarriff et al.
  • 31 frames have no Berkeley counterpart, including 8 entirely new ones such as Animals, Plants, Languages and Furniture. Jury deliberation was renamed Deliberation because Sweden has no juries.

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: Linguistics & languages

Topics: Semantic role labeling

Cited by

12 papers of our library cite it (of the 140673 whose bibliographies we read); 2 of them have a page here, 10 not yet.

Cite

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

@inproceedings{heppin2012rocky,
  title = {{The Rocky Road towards a Swedish FrameNet - Creating SweFN}},
  author = {Heppin, Karin Friberg and Gronostaj, Maria Toporowska},
  booktitle = {Proceedings of the Eighth International Conference on Language Resources and Evaluation (LREC'12)},
  year = {2012},
  doi = {10.63317/4bqbn4aoezg9},
  url = {https://doi.org/10.63317/4bqbn4aoezg9}
}

References

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

  1. B.T. Sue Atkins and Michael Rundell. 2008. The Oxford Guide to Practical Lexicography. Oxford University Press. ↗
  2. Lars Borin, Dana Dannélls, Markus Forsberg, Maria Toporowska Gronostaj, and Dimitrios Kokkinakis. 2009. Thinking green: Toward Swedish FrameNet++. find
  3. Lars Borin, Dana Dannélls, Markus Forsberg, Maria Toporowska Gronostaj, and Dimitrios Kokkinakis. 2010. The past meets the present in Swedish FrameNet++. The examples are from the Berkley FrameNet. find
  4. Lars Borin. 2010. Med Zipf mot framtiden - en integrerad lexikonresurs för svensk språkteknologi. LexicoNordica, 17:35–54. ↗
  5. Charles J. Fillmore, Christopher R. Johnson, and Miriam R.L.Petruck. 2003. Background to Framenet. International Journal of Lexicography, 16(3). ↗
  6. Adam Kilgarriff, Miloš Husák, Katy McAdam, Michael Rundell, and Pavel Rychlý. 2008. GDEX: Automatically finding good dictionary examples in a corpus. In Proceedings of the XIII EURALEX International Congress. find our card
  7. A. Lenci, N. Bel, F. Busa, N. Calzolari, E. Gola, and M. Monachini. 2000. SIMPLE: A general framework for the development of multilingual lexicons. Lexicography, 13(3). ↗ our card
  8. J. Ruppenhofer, M. Ellsworth, M.R.L.Petruck, C.R. Johnson, and J. Scheffczyk. 2010. FrameNet II: Extended theory and practice. Morgan Kaufman Publishers, <https://framenet2.icsi.berkeley.edu/docs/r1.5/book.pdf>. Figure 1: The frame Assistance. The fields shown in this figure are: Name of the frame, Domain, Semantic type, Core elements, Periferal elements, and Example sentences. The frame is continued in figure 2. Figure 2: The continuation of frame Assistance (from figure 1). The fields shown in this figure are: Example sentences (continuation), Instantiated compound patterns, Compound examples, Lexical units with links to the SALDO lexicon, Suggestions for new entries in SALDO, Notes, Creator, Date of creation, and Date of modification. ↗