---
title: "Comparing Corpus-based to Web-based Lookup Techniques for"
description: "The paper asks whether a Web-based lookup is really needed to spot English words embedded in… 19 references, cited by 5 papers here."
authors: ["Beatrice Alex"]
year: 2008
venue: "Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC'08)"
source: "https://aclanthology.org/L08-1388/"
licence: "CC BY-NC 4.0 (ELRA, LREC proceedings)"
licence_url: "https://creativecommons.org/licenses/by-nc/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/alex-2008-comparing-corpus-based-web-based/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/alex-2008-comparing-corpus-based-web-based/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2008](https://md.research.elamur.ai/free/en/papers/2008/page/2/) / Comparing Corpus-based to Web-based Lookup Techniques for Automatic English Inclusion Detection

# Comparing Corpus-based to Web-based Lookup Techniques for Automatic English Inclusion Detection

**Authors:** [Beatrice Alex](https://md.research.elamur.ai/free/en/authors/beatrice-alex/) · **Year:** 2008 · **Venue:** [Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC'08)](https://md.research.elamur.ai/free/en/venues/lrec/) ([LREC 2008](https://md.research.elamur.ai/free/en/venues/lrec/2008/)) · **Publisher:** [European Language Resources Association (ELRA)](https://md.research.elamur.ai/free/en/publishers/european-language-resources-association-elra/) · **Licence:** [CC BY-NC 4.0 (ELRA, LREC proceedings)](https://creativecommons.org/licenses/by-nc/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/L08-1388/) · [PDF](https://aclanthology.org/L08-1388.pdf) · [OpenAlex](https://openalex.org/W1522653827)

**Citations:** 10 citations (OpenAlex, 2026-10-14) · [5](https://md.research.elamur.ai/free/en/papers/alex-2008-comparing-corpus-based-web-based/cited-by/) in our library, of the 149,439 papers whose bibliographies we read

## Summary
The paper asks whether a Web-based lookup is really needed to spot English words embedded in German text, or whether static corpora could do the same job. The existing English inclusion classifier combines three parts. A CELEX lexicon lookup handles known words. A search engine module labels unknown tokens by comparing hit counts for English and German, normalised by estimated Web corpus sizes. Rule-based post-processing resolves the remaining cases. In the experiments the search engine module is swapped for a corpus search module. It labels a token English when its relative frequency in an English corpus exceeds its relative frequency in a German one. Evaluation uses a German internet & telecoms development set. The first setup pairs the WSJ part of the Penn Treebank with NEGRA+TIGER. A second setup uses growing subsets of Gigaword. Performance rises with more data, and the publication date of the data also matters. Even about 40m tokens fall short of the Web module. The Web wins because it is both huge and constantly refreshed, although it is slower and limited by query quotas.

## Key points
- The classifier has three stages. CELEX German/English lexicon lookup gives high precision. The search engine module gives high recall on unknown words. Rules handle homographs, function words, currencies, units, abbreviations and person names.
- On internet & telecoms data, about 6% of tokens are English inclusions. The full classifier reaches F1 82.17 on the dev set and 83.18 on the unseen test set, with precision above 92% and recall around 73–76%.
- The first corpus experiment uses roughly size-matched 1990s corpora: Penn Treebank WSJ (about 1.2m tokens) and NEGRA+TIGER (about 1.1m tokens), both turned into frequency lists.
- The second experiment uses increasing Gigaword subsets, with and without post-processing, and plots them against the search engine module baseline (Figure 1).
- Matching the Web module would require a corpus well beyond 40m tokens in size. Fixed corpora age quickly and miss new borrowings.

## Interesting ideas
- Language is decided per token by comparing frequencies normalised by corpus size across two corpora, rather than by raw counts, so corpora of different sizes stay comparable.
- The pipeline is cascaded: a high-precision lexicon handles known words, and only unknown words go to a high-recall but costly lookup, which saves expensive queries.
- The Web module is replaced by a controlled alternative with the rest of the architecture held fixed, which isolates exactly what the swapped part contributes.
- Learning curves over growing corpus subsets, plotted against a fixed reference line, show how much data a static resource would need to catch up.
- The age of the resource is treated as its own variable, because frozen corpora miss neologisms and new borrowings.

## For Mova

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

- **I1** (tokens): The tagger could flag foreign or code-switched tokens, or rare words, by comparing normalised frequencies in reference corpora before assigning a POS tag or lemma.
- **I2** (tokens): The lemmatizer and morphology could consult the cheap dictionary first and send only out-of-vocabulary words to slower resources or an LLM.
- **I3** (learn): Evaluation could swap single components, such as the knowledge source behind a frame or the parser, while freezing everything else, to measure each part's real value.
- **I4** (learn): Silver-data experiments could plot parser or repair-rule quality against data size, with the best witness parser as a horizontal baseline.
- **I5** (world): Compiled knowledge from WordNet or Wikidata could store the date of each entry, and the world builder could treat unknown recent terms as possibly new rather than wrong.

*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 Edinburgh](https://md.research.elamur.ai/free/en/universities/university-of-edinburgh/) — University, Edinburgh, United Kingdom · 359 works · 77,352 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 (1): [English](https://md.research.elamur.ai/free/en/topics/english/) · 2,263 works

## Cite

Cite as: Beatrice Alex. 2008. Comparing Corpus-based to Web-based Lookup Techniques for Automatic English Inclusion Detection. Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC'08).

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

```bibtex
@inproceedings{alex2008comparing,
  title = {{Comparing Corpus-based to Web-based Lookup Techniques for Automatic English Inclusion Detection}},
  author = {Beatrice Alex},
  booktitle = {Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC'08)},
  year = {2008},
  doi = {10.63317/4k7dzf67acvm},
  url = {https://doi.org/10.63317/4k7dzf67acvm}
}
```

- [← previous Syntactic Models for Structural Word Insertion and Deletion during Translation](https://md.research.elamur.ai/free/en/papers/menezes-2008-syntactic-models-structural-word-insertion/)
- [next → ProPOSEL: A Prosody and POS English Lexicon for Language Engineering](https://md.research.elamur.ai/free/en/papers/brierley-2008-proposel-prosody-pos-english-lexicon/)
- [↑ back to papers of 2008](https://md.research.elamur.ai/free/en/papers/2008/page/2/)

## References

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

1. Beatrice Alex, Amit Dubey, and Frank Keller. 2007. Using foreign inclusion detection to improve parsing performance. In Proceedings of EMNLP-CoNLL 2007, pages 151–160, Prague, Czech Republic. F1-score Corpus Size in Millions of Tokens Figure 1: Performance gains using the corpus search module (CSM) with/without post-processing (PP) with increasing sub-sets of the Gigaword corpus, compared to the performance of the search engine module (SEM) represented by the horizontal lines. [find](https://scholar.google.com/scholar?q=Beatrice+Alex%2C+Amit+Dubey%2C+and+Frank+Keller%2E+2007%2E+Using+foreign+inclusion+detection+to+improve+parsing+performance%2E+In+Proceedings+of+EMNLP%2DCoNLL+2007%2C+pages+151%E2%80%93160%2C+Prague%2C+Czech+Republic%2E+F1%2Dscore)
2. Beatrice Alex. 2005. An unsupervised system for identifying English inclusions in German text. In Proceedings of the 43rd Annual Meeting of the Association for Computational Linguistics (ACL 2005), Student Research Workshop, pages 133–138, Ann Arbor, Michigan, USA. [find](https://scholar.google.com/scholar?q=Beatrice+Alex%2E+2005%2E+An+unsupervised+system+for+identifying+English+inclusions+in+German+text%2E+In+Proceedings+of+the+43rd+Annual+Meeting+of+the+Association+for+Computational+Linguistics+%28ACL+2005%29%2C+St)
3. Beatrice Alex. 2006. Integrating language knowledge resources to extend the English inclusion classifier to a new language. In Proceedings of the 5th International Conference on Language Resources and Evaluation (LREC 2006), pages 595– 600, Genoa, Italy. [find](https://scholar.google.com/scholar?q=Beatrice+Alex%2E+2006%2E+Integrating+language+knowledge+resources+to+extend+the+English+inclusion+classifier+to+a+new+language%2E+In+Proceedings+of+the+5th+International+Conference+on+Language+Resources+and)
4. Gisle Andersen. 2005. Assessing algorithms for automatic extraction of anglicisms in Norwegian texts. In Proceedings of the International Conference on Corpus Linguistics (CL2005), Birmingham, UK. Available online at: [http://www.corpus.bham.ac.uk/PCLC/](http://www.corpus.bham.ac.uk/PCLC/). [↗](http://www.corpus.bham.ac.uk/PCLC/)
5. Werner Betz. 1936. Der Einfluß des Lateinischen auf den althochdeutschen Sprachschatz. Winter, Heidelberg. [find](https://scholar.google.com/scholar?q=Werner+Betz%2E+1936%2E+Der+Einflu%C3%9F+des+Lateinischen+auf+den+althochdeutschen+Sprachschatz%2E+Winter%2C+Heidelberg%2E)
6. Sabine Brants, Stefanie Dipper, Silvia Hansen, Wolfgang Lezius, and George Smith. 2002. The TIGER Treebank. In Proceedings of the Workshop on Treebanks and Linguistic Theories, September 20-21 (TLT02), pages 24–41, Sozopol, Bulgaria. [find](https://scholar.google.com/scholar?q=Sabine+Brants%2C+Stefanie+Dipper%2C+Silvia+Hansen%2C+Wolfgang+Lezius%2C+and+George+Smith%2E+2002%2E+The+TIGER+Treebank%2E+In+Proceedings+of+the+Workshop+on+Treebanks+and+Linguistic+Theories%2C+September+20%2D21+%28TLT02%29)
7. Broder Carstensen. 1979. Evidente und latente Einflüsse des Englischen auf das Deutsche. In Peter Braun, editor, Fremdwort-Diskussion, pages 90–94. Fink, Munich. [find](https://scholar.google.com/scholar?q=Broder+Carstensen%2E+1979%2E+Evidente+und+latente+Einflu%CC%88sse+des+Englischen+auf+das+Deutsche%2E+In+Peter+Braun%2C+editor%2C+Fremdwort%2DDiskussion%2C+pages+90%E2%80%9394%2E+Fink%2C+Munich%2E)
8. David Duckworth. 1977. Zur terminologischen und systematischen Grundlage der Forschung auf dem Gebiet der englisch-deutschen Interferenz. In Herbert Kolb and Hartmut Lauffer, editors, Sprachliche Interferenz, pages 36–56. Niemeyer, Tübingen. [find](https://scholar.google.com/scholar?q=David+Duckworth%2E+1977%2E+Zur+terminologischen+und+systematischen+Grundlage+der+Forschung+auf+dem+Gebiet+der+englisch%2Ddeutschen+Interferenz%2E+In+Herbert+Kolb+and+Hartmut+Lauffer%2C+editors%2C+Sprachliche+Inte)
9. Paulseph-John Farrugia. 2005. Text to Speech Technologies for Mobile Telephony Services. Department of Computer Science and AI, University of Malta, Msida, Malta. MSc Thesis. [find](https://scholar.google.com/scholar?q=Paulseph%2DJohn+Farrugia%2E+2005%2E+Text+to+Speech+Technologies+for+Mobile+Telephony+Services%2E+Department+of+Computer+Science+and+AI%2C+University+of+Malta%2C+Msida%2C+Malta%2E+MSc+Thesis%2E)
10. Gregory Grefenstette and Julien Nioche. 2000. [Estimation of English and non-English language use on the WWW](https://md.research.elamur.ai/free/en/papers/grefenstette-2000-estimation-english-non-english-language/). In Proceedings of RIAO (Recherche d’Informations Assistée par Ordina-CSM without PP CSM with PP SEM without PP SEM with PP teur) 2000, pages 237–246, Paris, France. [↗](https://doi.org/10.48550/arxiv.cs/0006032)
11. Einar Haugen. 1950. The analysis of linguistic borrowing. Language, 26:210–231. [↗](https://doi.org/10.2307/410058)
12. Alan Kirkness. 1984. Aliens, denziens, hybrids and natives: foreign influence on the etymological structure of German vocabulary. In Charles V. J. Russ, editor, Foreign Influences on German. Lochee, Dundee. [find](https://scholar.google.com/scholar?q=Alan+Kirkness%2E+1984%2E+Aliens%2C+denziens%2C+hybrids+and+natives%3A+foreign+influence+on+the+etymological+structure+of+German+vocabulary%2E+In+Charles+V%2E+J%2E+Russ%2C+editor%2C+Foreign+Influences+on+German%2E+Lochee%2C+D)
13. J. C. Marcadet, V. Fischer, and C. Waast-Richard. 2005. A transformation-based learning approach to language identification for mixed-lingual text-to-speech synthesis. In Proceedings of the 9th European Conference on Speech Communication and Technology (Interspeech 2005 - Eurospeech), pages 2249–2252, Lisbon, Portugal. [find](https://scholar.google.com/scholar?q=J%2E+C%2E+Marcadet%2C+V%2E+Fischer%2C+and+C%2E+Waast%2DRichard%2E+2005%2E+A+transformation%2Dbased+learning+approach+to+language+identification+for+mixed%2Dlingual+text%2Dto%2Dspeech+synthesis%2E+In+Proceedings+of+the+9th+Europe)
14. Mitchell P. Marcus, Beatrice Santorini, and Mary Ann 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(2):313–330. [↗](https://doi.org/10.21236/ada273556)
15. Alexander Onysko. 2007. Anglicisms in German: Borrowing, Lexical Productivity and Written Codeswitching. De Gruyter, Berlin/New York. [find](https://scholar.google.com/scholar?q=Alexander+Onysko%2E+2007%2E+Anglicisms+in+German%3A+Borrowing%2C+Lexical+Productivity+and+Written+Codeswitching%2E+De+Gruyter%2C+Berlin%2FNew+York%2E)
16. Beat Pfister and Harald Romsdorfer. 2003. Mixed-lingual analysis for polyglot TTS synthesis. In Proceedings of the 8th European Conference on Speech Communication and Technology (Eurospeech 2003), pages 2037–2040, Geneva, Switzerland. [find](https://scholar.google.com/scholar?q=Beat+Pfister+and+Harald+Romsdorfer%2E+2003%2E+Mixed%2Dlingual+analysis+for+polyglot+TTS+synthesis%2E+In+Proceedings+of+the+8th+European+Conference+on+Speech+Communication+and+Technology+%28Eurospeech+2003%29%2C+pag)
17. Ariel Schwartz and Marti Hearst. 2003. A simple algorithm for identifying abbreviation definitions in biomedical text. In Proceedings of the Pacific Symposium on Biocomputing (PSB 2003), pages 451–462, Kauai, Hawaii. [find](https://scholar.google.com/scholar?q=Ariel+Schwartz+and+Marti+Hearst%2E+2003%2E+A+simple+algorithm+for+identifying+abbreviation+definitions+in+biomedical+text%2E+In+Proceedings+of+the+Pacific+Symposium+on+Biocomputing+%28PSB+2003%29%2C+pages+451%E2%80%93462)
18. Wojciech Skut, Thorsten Brants, Brigitte Krenn, and Hans Uszkoreit. 1998. A linguistically interpreted corpus of German newspaper text. In Proceedings of the Conference on Language Resources and Evaluation (LREC 1998), pages 705– 712, Granada, Spain. [↗](https://doi.org/10.48550/arxiv.cmp-lg/9807008)
19. Wenliang Yang. 1990. Anglizismen im Deutschen: am Beispiel des Nachrichtenmagazins Der Spiegel. Niemeyer, Tübingen. [find](https://scholar.google.com/scholar?q=Wenliang+Yang%2E+1990%2E+Anglizismen+im+Deutschen%3A+am+Beispiel+des+Nachrichtenmagazins+Der+Spiegel%2E+Niemeyer%2C+Tu%CC%88bingen%2E)

*Updated 1 November 2026*