---
title: "PaddyWaC — paper by Murphy and Stemle (2011)"
description: "Hand-built corpora of minor language varieties, such as ICE-Ireland at roughly a million words, are high quality but small and… 16 references."
authors: ["Brian Murphy", "Egon W. Stemle"]
year: 2011
venue: "Proceedings of the First Workshop on Algorithms and Resources for Modelling of Dialects and Language Varieties"
source: "https://aclanthology.org/W11-2603/"
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/murphy-2011-paddywac-minimally-supervised-web-corpus/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/murphy-2011-paddywac-minimally-supervised-web-corpus/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2011](https://md.research.elamur.ai/free/en/papers/2011/page/2/) / PaddyWaC: A Minimally-Supervised Web-Corpus of Hiberno-English

# PaddyWaC: A Minimally-Supervised Web-Corpus of Hiberno-English

**Authors:** Brian Murphy and [Egon W. Stemle](https://md.research.elamur.ai/free/en/authors/egon-w-stemle/) · **Year:** 2011 · **Venue:** [Proceedings of the First Workshop on Algorithms and Resources for Modelling of Dialects and Language Varieties](https://md.research.elamur.ai/free/en/venues/workshop-on-algorithms-and-resources-for-modelling-of-dialects-and-language/) ([Workshop on Algorithms and Resources for Modelling of Dialects and Language Varieties 2011](https://md.research.elamur.ai/free/en/venues/workshop-on-algorithms-and-resources-for-modelling-of-dialects-and-language/2011/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **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/W11-2603/) · [PDF](https://aclanthology.org/W11-2603.pdf) · [OpenAlex](https://openalex.org/W848342684)

**Citations:** 3 citations (OpenAlex, 2026-10-14)

## Summary
Hand-built corpora of minor language varieties, such as ICE-Ireland at roughly a million words, are high quality but small and costly to produce. Web harvesting can scale up, but a variety like Hiberno-English is hard to separate from the British and US English that also circulates on Irish sites. Murphy and Stemle used the BootCaT pipeline from the WaCky initiative to build seven prototype Hiberno-English corpora. They varied three settings: the seed set (Irish-specific terms or generic ukWaC seeds), the top-level-domain restriction, and an optional filter that drops pages with mostly American spelling. ukWaC and a newly crawled US-domain corpus served as comparison points. "Irishness" was measured by how often Ireland-related names appear and by four grammatical markers of Hiberno-English. Restricting to .ie mainly yields Irish topics written in largely standard grammar. Irish seed terms bring in more informal, dialect-rich text, but also some fake or quoted Irishisms. This matters for dialects that have no national domain of their own.

## Key points
- Seven Hiberno-English prototype corpora were built with BootCaT and compared against ukWaC, a US-domain web corpus, Scannell's .ie-filtered Crúbadán corpus and the hand-made ICE-Ireland. All were processed for the IMS Open Corpus Workbench (CWB) using the ukWaC processing chain.
- The experiments varied three dimensions:
  - the seeds: Irish seeds (IEs), used in pairs or triples, versus the ukWaC mid-frequency BNC seeds (UKs);
  - the TLD restriction: .ie or any domain;
  - a British-vs-American spelling filter (BrEn).
- Four structural markers were counted:
  - the contraction *amn't*;
  - the "after" perfective (*I'm after losing…*);
  - embedded questions with inversion and no *if/whether*;
  - unbound subject reflexives (*himself is…*).
- ICE-Ireland contains only 7 instances of the "after" construction. A restrictive query on the Irish-seed corpus found 26.
- Many hits from the Irish seeds were false positives. They came from pages that discuss Hiberno-English, and from scripts and fiction. Marker rates per million words stayed below ICE-Ireland, which is about 60% spoken.

## Interesting ideas
- Checking a corpus with diagnostic constructions: judge whether harvested text belongs to the target variety by counting a few syntactic markers findable with shallow POS patterns, rather than trusting its source label.
- Separating topical from structural signals: domain filtering reliably captures what text is about but not how it is written, so the two need separate metrics.
- Watching for metalinguistic contamination: pages that describe a dialect are packed with invented example sentences, which inflate counts and create false positives that need manual review.
- Choosing the source by register: national-domain pages lean toward edited institutional prose, while seed-driven search reaches forums and blogs where vernacular grammar appears.
- Iterative seed discovery: compare the first harvest against reference corpora of other varieties to find comparatively specific terms, then use them as new seeds.

## For Mova

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

- **I1** (learn): A small set of variety-specific construction probes could test whether silver data or new raw text truly matches the intended register before it enters training.
- **I2** (learn): Evaluation sets could report topic match and grammatical-register match as separate scores, so a corpus that is on-topic but formal is not mistaken for informal data.
- **I3** (tree): When treebank material is mined for rare constructions, sentences cited inside linguistic discussion should be flagged, since they are mentions of a construction rather than genuine usage.
- **I4** (tokens): Test text for the tokenizer and tagger could be gathered with informal seed terms to deliberately expose contractions like *amn't* and other non-standard forms.
- **I5** (world): Contrasting against a reference corpus could help build region-specific lexicon and proper-name lists, such as organisations and places, for the knowledge layer.

*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 Trento](https://md.research.elamur.ai/free/en/universities/university-of-trento/) — University, Trento, Italy · 87 works · 15,761 citations

📚 Shelves (5): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 17,965 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 17,450 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 17,296 works · [🗄️ Information Systems](https://md.research.elamur.ai/free/en/shelves/computer-science/information-systems/) · 1,730 works

🧭 Topics (1): [English](https://md.research.elamur.ai/free/en/topics/english/) · 2,379 works

## Cite

Cite as: Brian Murphy and Egon W. Stemle. 2011. PaddyWaC: A Minimally-Supervised Web-Corpus of Hiberno-English. Proceedings of the First Workshop on Algorithms and Resources for Modelling of Dialects and Language Varieties.

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

```bibtex
@inproceedings{murphy2011paddywac,
  title = {{PaddyWaC: A Minimally-Supervised Web-Corpus of Hiberno-English}},
  author = {Brian Murphy and Egon W. Stemle},
  booktitle = {Proceedings of the First Workshop on Algorithms and Resources for Modelling of Dialects and Language Varieties},
  year = {2011},
  url = {https://aclanthology.org/W11-2603/}
}
```

- [← previous Infinity in computable probability](https://md.research.elamur.ai/free/en/papers/mckubre-jordens-2011-infinity-computable-probability/)
- [next → Semantic Information and Derivation Rules for Robust Dialogue Act Detection in a Spoken Dialogue System](https://md.research.elamur.ai/free/en/papers/liang-2011-semantic-information-derivation-rules-robust/)
- [↑ back to papers of 2011](https://md.research.elamur.ai/free/en/papers/2011/page/2/)

## References

16 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Baroni, M. and Bernardini, S. (2004). [BootCaT: Bootstrapping corpora and terms from the web](https://md.research.elamur.ai/free/en/papers/baroni-2004-bootcat-bootstrapping-corpora-terms-web/). In (ELRA), E. L. R. A., editor, Proceedings of LREC 2004, Lisbon: ELDA., pages 1313–1316.
2. Baroni, M. and Bernardini, S., editors (2006). Wacky! Working papers on the Web as Corpus. [find](https://scholar.google.com/scholar?q=Baroni%2C+M%2E+and+Bernardini%2C+S%2E%2C+editors+%282006%29%2E+Wacky%21+Working+papers+on+the+Web+as+Corpus%2E)
3. 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(3):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+%282009%29%2E+The+WaCky+wide+web%3A+a+collection+of+very+large+linguistically+processed+web%2Dcrawled+corpora%2E+Language+Resources+and+Evaluation%2C+43%28)
4. Burnard, L. (1995). Users Reference Guide, British National Corpus, Version 1.0. Oxford University Computing Services/British National Corpus Consortium, Oxford. [find](https://scholar.google.com/scholar?q=Burnard%2C+L%2E+%281995%29%2E+Users+Reference+Guide%2C+British+National+Corpus%2C+Version+1%2E0%2E+Oxford+University+Computing+Services%2FBritish+National+Corpus+Consortium%2C+Oxford%2E)
5. Christ, O. (1994). A Modular and Flexible Architecture for an Integrated Corpus Query System. In Papers in Computational Lexicography (COMPLEX ’94), pages 22–32. [find](https://scholar.google.com/scholar?q=Christ%2C+O%2E+%281994%29%2E+A+Modular+and+Flexible+Architecture+for+an+Integrated+Corpus+Query+System%2E+In+Papers+in+Computational+Lexicography+%28COMPLEX+%E2%80%9994%29%2C+pages+22%E2%80%9332%2E)
6. Erjavec, I. S., Erjavec, T., and Kilgarriff, A. (2008). A web corpus and word sketches for Japanese. Information and Media Technologies, 3:529–551. [find](https://scholar.google.com/scholar?q=Erjavec%2C+I%2E+S%2E%2C+Erjavec%2C+T%2E%2C+and+Kilgarriff%2C+A%2E+%282008%29%2E+A+web+corpus+and+word+sketches+for+Japanese%2E+Information+and+Media+Technologies%2C+3%3A529%E2%80%93551%2E)
7. Finn, A., Kushmerick, N., and Smyth, B. (2001). Fact or fiction: Content classification for digital libraries. [find](https://scholar.google.com/scholar?q=Finn%2C+A%2E%2C+Kushmerick%2C+N%2E%2C+and+Smyth%2C+B%2E+%282001%29%2E+Fact+or+fiction%3A+Content+classification+for+digital+libraries%2E)
8. Greenbaum, S. (1996). Comparing English Worldwide. Clarendon Press. [find](https://scholar.google.com/scholar?q=Greenbaum%2C+S%2E+%281996%29%2E+Comparing+English+Worldwide%2E+Clarendon+Press%2E)
9. Guevara, E. (2010). NoWaC: a large web-based corpus for Norwegian. In Proceedings of the Sixth Web as Corpus Workshop (WAC6), pages 1–7. The Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Guevara%2C+E%2E+%282010%29%2E+NoWaC%3A+a+large+web%2Dbased+corpus+for+Norwegian%2E+In+Proceedings+of+the+Sixth+Web+as+Corpus+Workshop+%28WAC6%29%2C+pages+1%E2%80%937%2E+The+Association+for+Computational+Linguistics%2E)
10. Kallen, J. and Kirk, J. (2007). ICE-Ireland: Local variations on global standards. In Beal, J. C., Corrigan, K. P., and Moisl, H. L., editors, Creating and Digitizing Language Corpora: Synchronic Databases, volume 1, pages 121–162. Palgrave Macmillan, London. [find](https://scholar.google.com/scholar?q=Kallen%2C+J%2E+and+Kirk%2C+J%2E+%282007%29%2E+ICE%2DIreland%3A+Local+variations+on+global+standards%2E+In+Beal%2C+J%2E+C%2E%2C+Corrigan%2C+K%2E+P%2E%2C+and+Moisl%2C+H%2E+L%2E%2C+editors%2C+Creating+and+Digitizing+Language+Corpora%3A+Synchronic+Data)
11. Kirk, J. and Kallen, J. (2007). Assessing Celticity in a Corpus of Irish Standard English. In The Celtic languages in contact: papers from the workshop within the framework of the XIII International Congress of Celtic Studies, Bonn, 26-27 July 2007, page 270. [find](https://scholar.google.com/scholar?q=Kirk%2C+J%2E+and+Kallen%2C+J%2E+%282007%29%2E+Assessing+Celticity+in+a+Corpus+of+Irish+Standard+English%2E+In+The+Celtic+languages+in+contact%3A+papers+from+the+workshop+within+the+framework+of+the+XIII+International+C)
12. Levin, B. (1993). English Verb Classes and Alternations. University of Chicago Press, Chicago. [find](https://scholar.google.com/scholar?q=Levin%2C+B%2E+%281993%29%2E+English+Verb+Classes+and+Alternations%2E+University+of+Chicago+Press%2C+Chicago%2E)
13. Nelson, G., Wallis, S., and Aarts, B. (2002). Exploring natural language: working with the British component of the International Corpus of English. John Benjamins. [find](https://scholar.google.com/scholar?q=Nelson%2C+G%2E%2C+Wallis%2C+S%2E%2C+and+Aarts%2C+B%2E+%282002%29%2E+Exploring+natural+language%3A+working+with+the+British+component+of+the+International+Corpus+of+English%2E+John+Benjamins%2E)
14. Scannell, K. (2007). The Crúbadán project: Corpus building for under-resourced languages. In Fairon, C., Naets, H., Kilgarriff, A., and de Schryver, G.-M., editors, Building and Exploring Web Corpora: Proceedings of the 3rd Web as Corpus Workshop, volume 4, pages 5–15. [find](https://scholar.google.com/scholar?q=Scannell%2C+K%2E+%282007%29%2E+The+Cr%C3%BAbad%C3%A1n+project%3A+Corpus+building+for+under%2Dresourced+languages%2E+In+Fairon%2C+C%2E%2C+Naets%2C+H%2E%2C+Kilgarriff%2C+A%2E%2C+and+de+Schryver%2C+G%2E%2DM%2E%2C+editors%2C+Building+and+Exploring+Web+Corpora%3A)
15. Web (2008). The IMS Open Corpus Workbench (CWB). [find](https://scholar.google.com/scholar?q=Web+%282008%29%2E+The+IMS+Open+Corpus+Workbench+%28CWB%29%2E)
16. Yahoo! Inc. (1995). The Yahoo! Internet search engine. [find](https://scholar.google.com/scholar?q=Yahoo%21+Inc%2E+%281995%29%2E+The+Yahoo%21+Internet+search+engine%2E)

*Updated 2 November 2026*