---
title: "A low-complexity, broad-coverage probabilistic Dependency"
description: "Schneider presents a fast, robust parser for English that outputs functional dependency structures… 16 references, cited by 2 papers here."
authors: ["Gerold Schneider"]
year: 2003
venue: "Proceedings of the HLT-NAACL 2003 Student Research Workshop"
source: "https://aclanthology.org/N03-3006/"
licence: "CC BY-NC-SA 3.0 (ACL Anthology, before 2016)"
licence_url: "https://creativecommons.org/licenses/by-nc-sa/3.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/schneider-2003-low-complexity-broad-coverage-probabilistic/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/schneider-2003-low-complexity-broad-coverage-probabilistic/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2003](https://md.research.elamur.ai/free/en/papers/2003/) / A low-complexity, broad-coverage probabilistic Dependency Parser for English

# A low-complexity, broad-coverage probabilistic Dependency Parser for English

**Authors:** [Gerold Schneider](https://md.research.elamur.ai/free/en/authors/gerold-schneider/) · **Year:** 2003 · **Venue:** Proceedings of the HLT-NAACL 2003 Student Research Workshop · **Licence:** [CC BY-NC-SA 3.0 (ACL Anthology, before 2016)](https://creativecommons.org/licenses/by-nc-sa/3.0/)

**Links:** [Publisher's page](https://aclanthology.org/N03-3006/) · [PDF](https://aclanthology.org/N03-3006.pdf) · [OpenAlex](https://openalex.org/W2053354966)

**Citations:** 6 citations (OpenAlex, 2026-10-14) · 2 in our library, of the 149,439 papers whose bibliographies we read

## Summary
Full syntactic parsing of large text collections is often too slow for practical knowledge-management uses such as information retrieval or data mining. Schneider presents a fast, robust parser for English that outputs functional dependency structures, meaning labelled relations between content-word heads that sit close to argument structure. The design splits the work by difficulty. A finite-state tagger-chunker (LTPos) handles part-of-speech tags and base chunks, and a restrictive hand-written Dependency Grammar then links only the chunk heads using CYK. Two supervised MLE models rank the candidate analyses:
- lexical head-to-head relation probabilities, backed off to Levin verb classes and top WordNet noun classes;
- a VP PCFG that captures subcategorization.

Training data come from the Penn Treebank, converted to dependencies with many tgrep queries. Accuracy on section 0 is a little below the state of the art, but the whole section is parsed in about 30 minutes. The work shows that linguistically grounded parsing can be cheap enough for unrestricted text.

## Key points
- The models are trained on Penn Treebank sections 2–24 and tested on section 0, which has 46,527 words. Parsing it takes about 30 minutes on an 800 MHz Pentium 3, and roughly 3 of those minutes go to tagging and chunking.
- Provisional precision/recall figures (the table is garbled in the extracted text, so this reading is approximate):
  - subject: about 77/70;
  - object: about 72/75;
  - PP-attachment: about 67/49;
  - preposition-vs-complement disambiguation: 80/78.
- CYK runs over chunk heads instead of words. The chunk-to-word ratio is 1.52 on section 0, and in a toy test unchunked input was about 4× slower. Using only the tagger's best tag avoids the Treebank average of 2.11 possible tags per token.
- The relations modelled include `subj`, `obj`, `obj2`, `adj`, `sentobj`, `pobj` and `modpp`, among others. Passive subjects get their own probability model, separate from active subjects. The PP-attachment model follows Collins and Brooks (1995).
- The parser is written in Prolog (SWI-Prolog and Sicstus) and has an XPCE graphical interface. When no full parse exists, it returns an optimal sequence of partial structures.

## Interesting ideas
- Handle passives as a surface pattern with their own subject statistics, with no trace or movement, because the verb form already marks the passive reliably.
- When the grammar cannot tell two categories apart reliably (verbal particles vs prepositions), merge them on purpose so the analysis is consistently wrong rather than randomly wrong.
- Back off sparse lexical counts to semantic classes: Levin classes for verbs and the top WordNet class of the most frequent sense for nouns. Linear interpolation was also tried and gave almost the same results.
- Recover from failed parses greedily: take the best-ranked longest partial analysis, then repeat the search on the uncovered spans to its left and right.
- The author notes that a Treebank-to-dependency conversion can be good enough for training statistics but too noisy for evaluation, since chunker and Treebank conventions differ.

## For Mova

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

- **I1** (tree): keep separate attachment statistics for passive `nsubj:pass` versus active `nsubj` in the perceptron parser instead of sharing one model.
- **I2** (frames): a consistent error can be repaired later by a single explainable phrasal-verb rule that maps the merged `pobj` back to a particle reading before semantic roles are assigned.
- **I3** (world): the compiled VerbNet/WordNet class inventory could serve directly as a back-off layer for attachment and role statistics on rare lemmas.
- **I4** (tree): a deterministic, explainable fallback for fragments and headlines, so downstream frames still receive maximal well-formed subtrees.
- **I5** (learn): score converted or silver treebanks separately from gold, and flag convention mismatches before reading LAS differences as real parser errors.

*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 Zurich](https://md.research.elamur.ai/free/en/universities/university-of-zurich/) — University, Zurich, Switzerland · 78 works · 12,260 citations

📚 Shelves (4): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 16,667 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 16,194 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 16,039 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 15,891 works

🧭 Topics (2): [English](https://md.research.elamur.ai/free/en/topics/english/) · 2,263 works · [🌳 Dependency parsing](https://md.research.elamur.ai/free/en/topics/dependency-parsing/) · 131 works

## Cite

Cite as: Gerold Schneider. 2003. A low-complexity, broad-coverage probabilistic Dependency Parser for English. Proceedings of the HLT-NAACL 2003 Student Research Workshop.

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

```bibtex
@inproceedings{schneider2003low,
  title = {{A low-complexity, broad-coverage probabilistic Dependency Parser for English}},
  author = {Gerold Schneider},
  booktitle = {Proceedings of the HLT-NAACL 2003 Student Research Workshop},
  year = {2003},
  doi = {10.3115/1073416.1073422},
  url = {https://doi.org/10.3115/1073416.1073422}
}
```

- [← previous Syntactic disambiguation using presupposition resolution](https://md.research.elamur.ai/free/en/papers/knott-2003-syntactic-disambiguation-presupposition-resolution/)
- [next → Application of corpus-based techniques to Amharic texts](https://md.research.elamur.ai/free/en/papers/fissaha-2003-application-corpus-based-techniques-amharic/)
- [↑ back to papers of 2003](https://md.research.elamur.ai/free/en/papers/2003/)

## References

16 references, 4 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. John Carroll, Guido Minnen, and Ted Briscoe. 1999. [Corpus annotation for parser evaluation](https://md.research.elamur.ai/free/en/papers/carroll-1999-corpus-annotation-parser-evaluation/). In Proceedings of the EACL-99 Post-Conference Workshop on Linguistically Interpreted Corpora, Bergen, Norway. [↗](https://doi.org/10.48550/arxiv.cs/9907013)
2. Michael Collins and James Brooks. 1995. Prepositional attachment through a backed-off model. In Proceedings of the Third Workshop on Very Large Corpora, Cambridge, MA. [find](https://scholar.google.com/scholar?q=Michael+Collins+and+James+Brooks%2E+1995%2E+Prepositional+attachment+through+a+backed%2Doff+model%2E+In+Proceedings+of+the+Third+Workshop+on+Very+Large+Corpora%2C+Cambridge%2C+MA%2E)
3. Michael Collins. 1999. Head-Driven Statistical Models for Natural Language Parsing. Ph.d. dissertation, University of Pennsylvania, Philadelphia, PA. [find](https://scholar.google.com/scholar?q=Michael+Collins%2E+1999%2E+Head%2DDriven+Statistical+Models+for+Natural+Language+Parsing%2E+Ph%2Ed%2E+dissertation%2C+University+of+Pennsylvania%2C+Philadelphia%2C+PA%2E)
4. Michael A. Covington. 1994. An empirically motivated reinterpretation of Dependency Grammar. Technical Report AI1994-01, University of Georgia, Athens, Georgia. [↗](https://doi.org/10.48550/arxiv.cmp-lg/9404004)
5. Christiane Fellbaum, editor. 1998. WordNet: An Electronic Lexical Database. MIT Press, Cambridge, MA. [find](https://scholar.google.com/scholar?q=Christiane+Fellbaum%2C+editor%2E+1998%2E+WordNet%3A+An+Electronic+Lexical+Database%2E+MIT+Press%2C+Cambridge%2C+MA%2E)
6. Beth C. Levin. 1993. English Verb Classes and Alternations: a Preliminary Investigation. University of Chicago Press, Chicago, IL. [find](https://scholar.google.com/scholar?q=Beth+C%2E+Levin%2E+1993%2E+English+Verb+Classes+and+Alternations%3A+a+Preliminary+Investigation%2E+University+of+Chicago+Press%2C+Chicago%2C+IL%2E)
7. Dekang Lin. 1995. A dependency-based method for evaluating broad-coverage parsers. In Proceedings of IJCAI-95, Montreal. [find](https://scholar.google.com/scholar?q=Dekang+Lin%2E+1995%2E+A+dependency%2Dbased+method+for+evaluating+broad%2Dcoverage+parsers%2E+In+Proceedings+of+IJCAI%2D95%2C+Montreal%2E)
8. Dekang Lin. 1998. Dependency-based evaluation of MINIPAR. In Workshop on the Evaluation of Parsing Systems, Granada, Spain. [find](https://scholar.google.com/scholar?q=Dekang+Lin%2E+1998%2E+Dependency%2Dbased+evaluation+of+MINIPAR%2E+In+Workshop+on+the+Evaluation+of+Parsing+Systems%2C+Granada%2C+Spain%2E)
9. Mitch Marcus, Beatrice Santorini, and M.A. Marcinkiewicz. 1993. [Building a large annotated corpus of English: the Penn Treebank](https://md.research.elamur.ai/free/en/papers/marcus-1993-building-large-annotated-corpus-english/). Computational Linguistics, 19:313–330. [↗](https://doi.org/10.21236/ada273556)
10. Andrei Mikheev. 1997. [Automatic rule induction for unknown word guessing](https://md.research.elamur.ai/free/en/papers/mikheev-1997-automatic-rule-induction-unknown-word/). Computational Linguistics, 23(3):405–423.
11. Guido Minnen, John Carroll, and Darren Pearce. 2000. Applied morphological generation. In Proceedings of the 1st International Natural Language Generation Conference (INLG), Mitzpe Ramon, Israel. [find](https://scholar.google.com/scholar?q=Guido+Minnen%2C+John+Carroll%2C+and+Darren+Pearce%2E+2000%2E+Applied+morphological+generation%2E+In+Proceedings+of+the+1st+International+Natural+Language+Generation+Conference+%28INLG%29%2C+Mitzpe+Ramon%2C+Israel%2E)
12. Gabriele Musillo and Khalil Sima’an. 2002. Towards comparing parsers from different linguistic frameworks. In Proceedings of LREC 2002 Beyond PAR-SEVAL Workshop, Las Palmas, Spain. [find](https://scholar.google.com/scholar?q=Gabriele+Musillo+and+Khalil+Sima%E2%80%99an%2E+2002%2E+Towards+comparing+parsers+from+different+linguistic+frameworks%2E+In+Proceedings+of+LREC+2002+Beyond+PAR%2DSEVAL+Workshop%2C+Las+Palmas%2C+Spain%2E)
13. Judita Preiss. 2003. Using grammatical relations to compare parsers. In Proceedings of EACL 03, Budapest, Hungary. [find](https://scholar.google.com/scholar?q=Judita+Preiss%2E+2003%2E+Using+grammatical+relations+to+compare+parsers%2E+In+Proceedings+of+EACL+03%2C+Budapest%2C+Hungary%2E)
14. Christer Samuelsson and Atro Voutilainen. 1997. Comparing a linguistic and a stochastic tagger. In Proceedings of of ACL/EACL Joint Conference, Madrid. [↗](https://doi.org/10.3115/979617.979649)
15. Pasi Tapanainen and Timo Järvinen. 1997. [A non-projective dependency parser](https://md.research.elamur.ai/free/en/papers/tapanainen-1997-non-projective-dependency-parser/). In Proceedings of the 5th Conference on Applied Natural Language Processing, pages 64–71. Association for Computational Linguistics.
16. Lucien Tesnière. 1959. Eléments de Syntaxe Structurale. Librairie Klincksieck, Paris. [↗](https://doi.org/10.5281/zenodo.19430142)

*Updated 1 November 2026*