---
title: "Modeling the Complexity and Descriptive Adequacy of Construction Grammars — Dunn 2018"
description: "The paper proposes a way to compare construction grammars (CxG) objectively without gold annotation, relying on the… 26 references, cited by 7 papers here."
authors: ["Jonathan Dunn"]
year: 2018
venue: "Proceedings of the Society for Computation in Linguistics (SCiL) 2018"
source: "https://aclanthology.org/W18-0309/"
doi: "10.7275/r59p2ztb"
arxiv: "1904.05588"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/dunn-2018-modeling-complexity-descriptive-adequacy-construction/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/dunn-2018-modeling-complexity-descriptive-adequacy-construction/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2018](https://md.research.elamur.ai/free/en/papers/2018/page/4/) / Modeling the Complexity and Descriptive Adequacy of Construction Grammars

# Modeling the Complexity and Descriptive Adequacy of Construction Grammars

**Authors:** [Jonathan Dunn](https://md.research.elamur.ai/free/en/authors/jonathan-dunn/) · **Year:** 2018 · **Venue:** Proceedings of the Society for Computation in Linguistics (SCiL) 2018 · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/W18-0309/) · [PDF](https://aclanthology.org/W18-0309.pdf) · [DOI 10.7275/r59p2ztb](https://doi.org/10.7275/r59p2ztb) · [arXiv 1904.05588](https://arxiv.org/abs/1904.05588)

**Citations:** [7](https://md.research.elamur.ai/free/en/papers/dunn-2018-modeling-complexity-descriptive-adequacy-construction/cited-by/) in our library, of the 140,884 papers whose bibliographies we read

## Summary

The paper proposes a way to compare construction grammars (CxG) objectively without gold annotation, relying on the Minimum Description Length (MDL) principle. The author measures the complexity of a grammar as the size of its own encoding. Descriptive adequacy is measured as the size of the encoding of a corpus using that grammar. The sum of these two quantities becomes the objective function for finding the best grammar. Each slot of a construction is constrained at one of three levels: lexical (word forms), syntactic (Universal POS tags from RDRPosTagger) or semantic (clusters of word2vec vectors). The grammars are learned automatically from the large web corpora WaC and Aranea for English, Spanish, French, German and Italian. Candidate constructions are selected by tabu search together with association measures. Learning proceeds in three passes: first lexis only, then syntax only, and finally all levels together. Grammar quality is assessed by how well the grammar compresses test data compared with unencoded data. The learned grammars achieve substantial compression, and the multi-level CxG compresses best, although its set of constructions is less stable across subsamples. The value of the work is that construction grammars become reproducible and testable on data rather than on a linguist's intuition.

## Key points

- The objective function is the minimum of the sum L1(G) + L2(D|G). Quality is given as 1 − MDL(G)/MDL(U), where U is the unencoded test set: a value closer to 1 means stronger compression. A measure of grammar stability across restarts is computed separately.
- The semantic domains were obtained as follows: a word2vec skip-gram model with 500 dimensions was trained on approximately 1 billion words of web text for each language and then partitioned by k-means into 100 clusters.
- Three types of grammars are compared: CxG_LEX, CxG_SYN (whose constructions subsequently function as context-free rules for constituents) and CxG_FULL. There are on average only 22 lexical constructions per language. They compress little but are almost completely stable.
- Experimental design: two-fold cross-validation with a random split and 2 restarts each on different portions of the training data. Segments of 20 thousand to 1 million sentences were used for the different stages.
- The code and models are open: github.com/jonathandunn/c2xg. The appendix contains examples of learned English constructions, in particular verb phrases with an infinitive and noun phrases with a prepositional modifier.

*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](https://md.research.elamur.ai/free/en/shelves/linguistics/)

## Cite

Cite as: Jonathan Dunn. 2018. Modeling the Complexity and Descriptive Adequacy of Construction Grammars. Proceedings of the Society for Computation in Linguistics (SCiL) 2018.

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

```bibtex
@inproceedings{dunn2018modeling,
  title = {{Modeling the Complexity and Descriptive Adequacy of Construction Grammars}},
  author = {Jonathan Dunn},
  booktitle = {Proceedings of the Society for Computation in Linguistics (SCiL) 2018},
  year = {2018},
  doi = {10.7275/r59p2ztb},
  eprint = {1904.05588},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.7275/r59p2ztb}
}
```

- [← previous Exponence, allomorphy and haplology in the number and State morphology of Modern Hebrew](https://md.research.elamur.ai/free/en/papers/faust-2018-exponence-allomorphy-haplology-number-state/)
- [next → A Characterwise Windowed Approach to Hebrew Morphological Segmentation](https://md.research.elamur.ai/free/en/papers/zeldes-2018-characterwise-windowed-approach-hebrew-morphological/)
- [↑ back to papers of 2018](https://md.research.elamur.ai/free/en/papers/2018/page/4/)

## References

26 references, 3 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Nguyen, Dat Quoc; Nguyen, Dai Quoc; Pham, Dang Duc; and Pham, Son Bao. 2016. A Robust Transformation-Based Learning Approach Using Ripple Down Rules for Part-Of-Speech Tagging. AI Communications, 29(3): 409-422. [find](https://scholar.google.com/scholar?q=Nguyen%2C+Dat+Quoc%3B+Nguyen%2C+Dai+Quoc%3B+Pham%2C+Dang+Duc%3B+and+Pham%2C+Son+Bao%2E+2016%2E+A+Robust+Transformation%2DBased+Learning+Approach+Using+Ripple+Down+Rules+for+Part%2DOf%2DSpeech+Tagging%2E+AI+Communications%2C+29%283)
2. Baroni, M., Bernardini, S., Ferraresi, A., and Zanchetta, E. 2009. The WaCky Wide Web: A Collection of Very Large Linguistically Processed Web-crawled Corpora. Language Resources and Evaluation, 43: 209-226. [find](https://scholar.google.com/scholar?q=Baroni%2C+M%2E%2C+Bernardini%2C+S%2E%2C+Ferraresi%2C+A%2E%2C+and+Zanchetta%2C+E%2E+2009%2E+The+WaCky+Wide+Web%3A+A+Collection+of+Very+Large+Linguistically+Processed+Web%2Dcrawled+Corpora%2E+Language+Resources+and+Evaluation%2C+43%3A+2)
3. Benko, V. 2014. Aranea: Yet Another Family of (Comparable) Web Corpora. In Proceedings of Text, Speech and Dialogue. 17th International Conference. 257-264. [find](https://scholar.google.com/scholar?q=Benko%2C+V%2E+2014%2E+Aranea%3A+Yet+Another+Family+of+%28Comparable%29+Web+Corpora%2E+In+Proceedings+of+Text%2C+Speech+and+Dialogue%2E+17th+International+Conference%2E+257%2D264%2E)
4. Petrov, S.; Das, D.; and McDonald, R. 2012. A Universal Part-of-Speech Tagset. In Proceedings of the Eight International Conference on Language Resources and Evaluation. [our card](https://md.research.elamur.ai/free/en/papers/petrov-2012-universal-part-speech-tagset/)
5. Bryant, J. 2004. Scalable Construction-based Parsing and Semantic Analysis. In Proceedings of the Workshop on Scalable Natural Language Understanding (HLT-NAACL): 33-40. [find](https://scholar.google.com/scholar?q=Bryant%2C+J%2E+2004%2E+Scalable+Construction%2Dbased+Parsing+and+Semantic+Analysis%2E+In+Proceedings+of+the+Workshop+on+Scalable+Natural+Language+Understanding+%28HLT%2DNAACL%29%3A+33%2D40%2E)
6. Rehurek, R. and Sojka, P. 2010. Software Framework for Topic Modelling with Large Corpora. In Proceedings of the LREC 2010 Workshop on New Challenges for NLP Frameworks. [find](https://scholar.google.com/scholar?q=Rehurek%2C+R%2E+and+Sojka%2C+P%2E+2010%2E+Software+Framework+for+Topic+Modelling+with+Large+Corpora%2E+In+Proceedings+of+the+LREC+2010+Workshop+on+New+Challenges+for+NLP+Frameworks%2E)
7. Bybee, J. 2006. From Usage to Grammar: The Mind’s Response to Repetition. Language, 82(4): 711-733. [find](https://scholar.google.com/scholar?q=Bybee%2C+J%2E+2006%2E+From+Usage+to+Grammar%3A+The+Mind%E2%80%99s+Response+to+Repetition%2E+Language%2C+82%284%29%3A+711%2D733%2E)
8. Rissanen, J. 1978. Modeling by the Shortest Data Description. Automatica, 14: 465-471. [find](https://scholar.google.com/scholar?q=Rissanen%2C+J%2E+1978%2E+Modeling+by+the+Shortest+Data+Description%2E+Automatica%2C+14%3A+465%2D471%2E)
9. Chang, N.; De Beule, J.; and Micelli, V. 2012. Computational construction grammar: Comparing ECG and FCG. In Steels, L. (ed.), Computational Issues in Fluid Construction Grammar. Berlin: Springer. 259-288. [find](https://scholar.google.com/scholar?q=Chang%2C+N%2E%3B+De+Beule%2C+J%2E%3B+and+Micelli%2C+V%2E+2012%2E+Computational+construction+grammar%3A+Comparing+ECG+and+FCG%2E+In+Steels%2C+L%2E+%28ed%2E%29%2C+Computational+Issues+in+Fluid+Construction+Grammar%2E+Berlin%3A+Springer%2E+259)
10. Rissanen, J. 1986. Stochastic Complexity and Modeling. Annals of Statistics, 14: 1,080-1,100. [find](https://scholar.google.com/scholar?q=Rissanen%2C+J%2E+1986%2E+Stochastic+Complexity+and+Modeling%2E+Annals+of+Statistics%2C+14%3A+1%2C080%2D1%2C100%2E)
11. Dunn, J. 2017. Computational Learning of Construction Grammars. Language & Cognition, 9(2): 254-292. [find](https://scholar.google.com/scholar?q=Dunn%2C+J%2E+2017%2E+Computational+Learning+of+Construction+Grammars%2E+Language+%26+Cognition%2C+9%282%29%3A+254%2D292%2E)
12. Steels, L. 2004. Constructivist development of grounded construction grammar. In Proceedings of the 42nd Meeting of the Association for Computational Linguistics: 9-16. [our card](https://md.research.elamur.ai/free/en/papers/steels-2004-constructivist-development-grounded-construction-grammar/)
13. Dunn, J. Forthcoming. Finding Variants for Construction-Based Dialectometry: A Corpus-Based Approach to Regional CxGs. Cognitive Linguistics. [our card](https://md.research.elamur.ai/free/en/papers/dunn-2021-finding-variants-construction-based-dialectometry/)
14. Steels, L. 2012. Design methods for fluid construction grammar. In Steels, L. (ed), Computational Issues in Fluid Construction Grammar. Berlin: Springer. 3-36. [find](https://scholar.google.com/scholar?q=Steels%2C+L%2E+2012%2E+Design+methods+for+fluid+construction+grammar%2E+In+Steels%2C+L%2E+%28ed%29%2C+Computational+Issues+in+Fluid+Construction+Grammar%2E+Berlin%3A+Springer%2E+3%2D36%2E)
15. Forsberg, M.; Johansson, R.; Bckstrm, L.; Borin, L.; Lyngfelt, B.; Olofsson, J.; and Prentice, J. 2014. From Construction Candidates to Constructicon Entries: An experiment using semi-automatic methods for identifying constructions in corpora. Constructions and Frames, 6(1): 114-135. [find](https://scholar.google.com/scholar?q=Forsberg%2C+M%2E%3B+Johansson%2C+R%2E%3B+Bckstrm%2C+L%2E%3B+Borin%2C+L%2E%3B+Lyngfelt%2C+B%2E%3B+Olofsson%2C+J%2E%3B+and+Prentice%2C+J%2E+2014%2E+From+Construction+Candidates+to+Constructicon+Entries%3A+An+experiment+using+semi%2Dautomatic+method)
16. Wible, D. and Tsao, N. 2010. StringNet as a Computational Resource for Discovering and Investigating Linguistic Constructions. In Proceedings of the Workshop on Extracting and Using Constructions in Computational Linguistics (NAACL-HTL): 25-31. [find](https://scholar.google.com/scholar?q=Wible%2C+D%2E+and+Tsao%2C+N%2E+2010%2E+StringNet+as+a+Computational+Resource+for+Discovering+and+Investigating+Linguistic+Constructions%2E+In+Proceedings+of+the+Workshop+on+Extracting+and+Using+Constructions+in+C)
17. Glover, F. 1989. Tabu Search, Part 1. ORSA Journal on Computing, 1(3): 190-206. [find](https://scholar.google.com/scholar?q=Glover%2C+F%2E+1989%2E+Tabu+Search%2C+Part+1%2E+ORSA+Journal+on+Computing%2C+1%283%29%3A+190%2D206%2E)
18. Glover. F. 1990a. Tabu Search, Part 2. ORSA Journal on Computing, 2(1): 4-32. [find](https://scholar.google.com/scholar?q=Glover%2E+F%2E+1990a%2E+Tabu+Search%2C+Part+2%2E+ORSA+Journal+on+Computing%2C+2%281%29%3A+4%2D32%2E)
19. Glover. F. 1990b. Tabu Search: A Tutorial. Interfaces, 20(4): 74-94. [find](https://scholar.google.com/scholar?q=Glover%2E+F%2E+1990b%2E+Tabu+Search%3A+A+Tutorial%2E+Interfaces%2C+20%284%29%3A+74%2D94%2E)
20. Goldberg, A. 2006. Constructions at Work: The Nature of Generalization in Language. Oxford: Oxford University Press. [find](https://scholar.google.com/scholar?q=Goldberg%2C+A%2E+2006%2E+Constructions+at+Work%3A+The+Nature+of+Generalization+in+Language%2E+Oxford%3A+Oxford+University+Press%2E)
21. Goldsmith, J. 2001. Unsupervised Learning of the Morphology of a Natural Language. Computational Linguistics, 27(2): 153-198. [find](https://scholar.google.com/scholar?q=Goldsmith%2C+J%2E+2001%2E+Unsupervised+Learning+of+the+Morphology+of+a+Natural+Language%2E+Computational+Linguistics%2C+27%282%29%3A+153%2D198%2E)
22. Goldsmith, J. 2006. An Algorithm for the Unsupervised Learning of Morphology. Natural Language Engineering, 12(4): 353-371. [find](https://scholar.google.com/scholar?q=Goldsmith%2C+J%2E+2006%2E+An+Algorithm+for+the+Unsupervised+Learning+of+Morphology%2E+Natural+Language+Engineering%2C+12%284%29%3A+353%2D371%2E)
23. Grünwald, P. and Rissanen, J. 2007. The Minimum Description Length Principle. Cambridge, MA: The MIT Press. [find](https://scholar.google.com/scholar?q=Gru%CC%88nwald%2C+P%2E+and+Rissanen%2C+J%2E+2007%2E+The+Minimum+Description+Length+Principle%2E+Cambridge%2C+MA%3A+The+MIT+Press%2E)
24. Hopper, P. 1987. Emergent Grammar. In Proceedings of the 13th Annual Meeting of the Berkeley Linguistics Society, 139-157. [find](https://scholar.google.com/scholar?q=Hopper%2C+P%2E+1987%2E+Emergent+Grammar%2E+In+Proceedings+of+the+13th+Annual+Meeting+of+the+Berkeley+Linguistics+Society%2C+139%2D157%2E)
25. Kay, P. and Fillmore, C. 1999. Grammatical Constructions and Linguistic Generalizations: The Whats X Doing Y? Construction. Language, 75(1): 1-33. [find](https://scholar.google.com/scholar?q=Kay%2C+P%2E+and+Fillmore%2C+C%2E+1999%2E+Grammatical+Constructions+and+Linguistic+Generalizations%3A+The+Whats+X+Doing+Y%3F+Construction%2E+Language%2C+75%281%29%3A+1%2D33%2E)
26. Langacker, R. 2008. Cognitive Grammar: A Basic Introduction. Oxford: Oxford University Press. [find](https://scholar.google.com/scholar?q=Langacker%2C+R%2E+2008%2E+Cognitive+Grammar%3A+A+Basic+Introduction%2E+Oxford%3A+Oxford+University+Press%2E)