---
title: "Lexicography — definition and key papers"
description: "Lexicography is the discipline of compiling, describing and designing dictionaries and related lexical resources."
canonical: "https://research.elamur.ai/free/en/concepts/lexicography/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/concepts/lexicography/)

[home](https://md.research.elamur.ai/free/en/) / [concepts](https://md.research.elamur.ai/free/en/concepts/) / Lexicography

# 📒 Lexicography

Concept · 30 key papers · [all concepts](https://md.research.elamur.ai/free/en/concepts/)

Lexicography is the discipline of compiling, describing and designing dictionaries and related lexical resources. It covers both the practical craft of writing entries and the theoretical study of how word knowledge should be selected, structured and presented to human users or to machines.

Also: dictionary making, dictionary compilation, practical lexicography, theoretical lexicography (metalexicography), electronic lexicography (e-lexicography), computational lexicography

## What it is

Lexicography has two sides. Practical lexicography is the work of producing a concrete dictionary: choosing a word list, dividing words into senses, writing definitions, picking examples and adding labels. Theoretical lexicography, often called metalexicography, studies dictionaries as objects. It asks about their typology, their users, the structure of entries and the criteria by which dictionaries are judged.

The basic unit is the entry. A typical entry contains several parts:

- a headword, or lemma (the citation form under which a word is listed), sometimes with variant or inflected forms;
- pronunciation and grammatical information;
- one or more senses, each with a definition or a translation equivalent;
- examples of use;
- labels that mark register, region or period.

Atkins gives a fuller inventory of entry components for bilingual dictionaries. It includes the following ([Atkins 1996](https://md.research.elamur.ai/free/en/papers/atkins-1996-bilingual-dictionaries/)):

- transcription in IPA (the International Phonetic Alphabet);
- many kinds of usage labels;
- sense indicators, which are short cues telling the user which meaning a translation belongs to;
- collocates in both languages;
- three kinds of examples: typical, problematic and idiomatic.

Today the field also covers resources that no human reads directly. These include lexical databases, machine-readable lexicons for parsers and taggers, and bilingual word lists produced by algorithms.

## The key idea

The central methodological idea of modern lexicography is to ground the description of a word in evidence of its actual use, and increasingly to let statistics organise that evidence. Church and Hanks proposed the association ratio, an adaptation of pointwise mutual information that measures how strongly two words attract each other in a corpus ([Church & Hanks 1990](https://md.research.elamur.ai/free/en/papers/church-1990-word-association-norms-mutual-information/)). Rychlý later proposed a score tuned to lexicographers' needs ([Rychlý 2008](https://md.research.elamur.ai/free/en/papers/rychly-2008-logdice/)):

```
I(x,y)       = log2 [ P(x,y) / ( P(x) · P(y) ) ]
logDice(x,y) = 14 + log2 [ 2 · f(x,y) / ( f(x) + f(y) ) ]
```

In the first line, P(x) and P(y) are the probabilities of encountering words x and y in the corpus, and log2 is the base-2 logarithm. P(x,y) is the probability that x is followed by y within a small window. Because it counts order, the ratio is asymmetric, unlike true mutual information. A value well above zero means the pair occurs together more often than chance would predict.

In the second line, f(x) and f(y) are the corpus frequencies of each word and f(x,y) is the frequency of the pair. The constant 14 sets the ceiling of the score: a pair whose words only ever occur together scores 14. Scores drop below zero only for very weak pairs ([Rychlý 2008](https://md.research.elamur.ai/free/en/papers/rychly-2008-logdice/)).

Church and Hanks used a window of 5 words and discarded pairs whose joint count f(x,y) was 5 or lower. They left the ratio asymmetric on purpose, so that word order is preserved. In 1988 Associated Press text of 44 million words, *save…from* appeared 176 times against 18 in the reverse order ([Church & Hanks 1990](https://md.research.elamur.ai/free/en/papers/church-1990-word-association-norms-mutual-information/)). The resulting ranked lists serve the lexicographer as an index. Concordance lines can be sorted by strong collocates and then grouped into senses based on usage ([Church & Hanks 1990](https://md.research.elamur.ai/free/en/papers/church-1990-word-association-norms-mutual-information/)).

The same principle extends to selecting examples. GDEX ranks corpus sentences automatically to propose good dictionary examples ([Kilgarriff et al. 2008](https://md.research.elamur.ai/free/en/papers/kilgarriff-2008-gdex-examples/)).

## History and founding papers

Dictionary making is old. The European tradition was shaped by landmarks such as Samuel Johnson's single-author dictionary (1755) and the Oxford English Dictionary. The OED was compiled on historical principles by many editors and volunteer readers. Ladislav Zgusta's Manual of Lexicography (1971) is often cited as a founding text of the theoretical side. The corpus turn is usually traced to John Sinclair's COBUILD project and the corpus-based dictionary it published in 1987.

Computational work on dictionaries began with machine-readable versions of printed ones. Alshawi parsed the sense definitions of the machine-readable LDOCE (Longman Dictionary of Contemporary English). These definitions are written in a restricted vocabulary of about 2000 words. His aim was to extract the semantic head and other classification features ([Alshawi 1987](https://md.research.elamur.ai/free/en/papers/alshawi-1987-processing-dictionary-definitions-phrasal-pattern/)). On 500 random definitions, the head was correct in 387 cases, or 77% ([Alshawi 1987](https://md.research.elamur.ai/free/en/papers/alshawi-1987-processing-dictionary-definitions-phrasal-pattern/)).

Jensen and Binot mined Webster's Seventh for cues such as "used for" to decide prepositional-phrase attachment ([Jensen & Binot 1987](https://md.research.elamur.ai/free/en/papers/jensen-1987-disambiguating-prepositional-phrase-attachments-line/)). Russell and colleagues built a dictionary and morphological analyser for English parsing ([Russell et al. 1986](https://md.research.elamur.ai/free/en/papers/russell-1986-dictionary-morphological-analyser-english/)). Church and Hanks's association ratio, described under the key idea above, then tied corpus statistics directly to lexicographic practice ([Church & Hanks 1990](https://md.research.elamur.ai/free/en/papers/church-1990-word-association-norms-mutual-information/)).

Lexical databases form a parallel line of work. WordNet, developed at Princeton by George Miller's group, organised English vocabulary as a network of word senses linked by semantic relations ([Miller 1994](https://md.research.elamur.ai/free/en/papers/miller-1994-wordnet/)). It became a widely used resource for disambiguation. Stetina and Nagao, for example, used a WordNet-based semantic distance to resolve attachment. They reached 88.1% accuracy on about 3,100 test quadruples, that is, verb–noun–preposition–noun tuples ([Stetina & Nagao 1997](https://md.research.elamur.ai/free/en/papers/stetina-1997-corpus-based-pp-attachment-ambiguity/)).

FrameNet, based on Fillmore's frame semantics, added syntactic realisations of meaning backed by annotated sentences. Boas designed a German counterpart linked to English through shared frames ([Boas 2002](https://md.research.elamur.ai/free/en/papers/boas-2002-bilingual-framenet-dictionaries-machine-translation/)).

Atkins, whose inventory of entry components appears above, argued that electronic bilingual dictionaries were still printed books in disguise. She proposed describing entries along six dimensions ([Atkins 1996](https://md.research.elamur.ai/free/en/papers/atkins-1996-bilingual-dictionaries/)):

- data type;
- information type;
- what the user can do with the information;
- mode of expression;
- the user's language;
- whether the user is encoding (producing text in the foreign language) or decoding (understanding it).

The automatic construction of bilingual dictionaries from existing dictionaries began with pivot methods. Tanaka and Umemura built a bilingual dictionary through a third language ([Tanaka & Umemura 1994](https://md.research.elamur.ai/free/en/papers/tanaka-1994-bilingual-dictionary-third-language/)). The approach was later applied to Japanese–Chinese via English ([Kaji et al. 2008](https://md.research.elamur.ai/free/en/papers/kaji-2008-automatic-construction-japanese-chinese-dictionary/)).

## How it is used today

Corpus-based dictionaries now operate at scale. The Leipzig Corpora Collection offers frequency lists, example sentences and significant co-occurrences. It had reached 136 online languages at the time of its report ([Goldhahn et al. 2012](https://md.research.elamur.ai/free/en/papers/goldhahn-2012-building-large-monolingual-dictionaries-leipzig/)).

Collaborative lexicography has produced Wiktionary, which has been examined as a possible rival to expert-built lexicons ([Meyer & Gurevych 2012](https://md.research.elamur.ai/free/en/papers/meyer-2012-wiktionary-new-rival/)). Its translation tables also feed automatic methods. Ács triangulated through pivot languages across 43 Wiktionary editions. A classifier, trained on automatically collected examples, then judged which triangulated candidate pairs were correct and reached an F1 of about 0.93. Coverage of a basic-vocabulary lexicon rose from 67.4% to 80.2% ([Acs 2014](https://md.research.elamur.ai/free/en/papers/acs-2014-wikt2dict-pivot/)). Probabilistic inference over many dictionaries has been used to compile a massive multilingual one ([Mausam et al. 2009](https://md.research.elamur.ai/free/en/papers/mausam-2009-panlingual-dictionary/)).

Electronic lexicography depends on shared encoding standards. TEI-Lex0 reduces the TEI dictionary guidelines to a baseline for exchanging retrodigitised dictionaries, meaning dictionaries converted from printed editions into digital form. One of its rules is that every form element must carry a type, such as lemma, inflected, paradigm or variant ([Bański et al. 2017](https://md.research.elamur.ai/free/en/papers/banski-2017-tei-lex0-forms/)).

Revisions of Moroccan, Syrian and Iraqi Arabic dictionaries are kept in a relational database. They are exported as XML that conforms to the Lexical Markup Framework (ISO 24,613) ([Graff & Maamouri 2012](https://md.research.elamur.ai/free/en/papers/graff-2012-developing-lmf-xml-bilingual-dictionaries/)). Bilingual dictionaries have also been used to transfer FrameNet to French ([Mouton et al. 2010](https://md.research.elamur.ai/free/en/papers/mouton-2010-framenet-translation-bilingual-dictionaries-evaluation/)).

Dictionaries are also valuable to NLP as supervision and as knowledge sources:

- **Reverse dictionaries.** Hill and colleagues trained encoders to map definitions onto headword embeddings and used them as a reverse dictionary ([Hill et al. 2016](https://md.research.elamur.ai/free/en/papers/hill-2016-learning-understand-phrases-embedding-dictionary/)). A SemEval task later compared definitions with embeddings directly ([Mickus et al. 2022](https://md.research.elamur.ai/free/en/papers/mickus-2022-codwoe-dictionaries-embeddings/)).
- **Bilingual dictionary induction.** This task aligns embedding spaces, often without parallel data and sometimes from a small seed lexicon. Even identically spelled words can serve as seeds ([Søgaard et al. 2018](https://md.research.elamur.ai/free/en/papers/sogaard-2018-limitations-unsupervised-bilingual-dictionary-induction/)). Its weak points are discussed under the open questions below.
- **Prompting for translation.** Large language models translate rare words better when dictionary hints are placed in the prompt. Compared with a prompt without hints, gains reach 13 BLEU points on German–English text from medical, legal, IT and Koran domains. BLEU is an automatic score of overlap with reference translations ([Ghazvininejad et al. 2023](https://md.research.elamur.ai/free/en/papers/ghazvininejad-2023-dictionary-based-phrase-level-prompting/)). Chained multilingual hints improve low-resource directions ([Lu et al. 2024](https://md.research.elamur.ai/free/en/papers/lu-2024-chain-dictionary-prompting-elicits-translation/)).
- **Named-entity tagging.** Domain dictionaries alone can train named-entity taggers through distant supervision, where labels come from automatically matching dictionary entries against raw text ([Shang et al. 2018](https://md.research.elamur.ai/free/en/papers/shang-2018-learning-named-entity-tagger-domain/)).
- **Specialised dictionaries.** Dictionaries of multiword expressions ([Shigeto et al. 2013](https://md.research.elamur.ai/free/en/papers/shigeto-2013-construction-english-mwe-dictionary-its/)) and of preposition patterns ([Litkowski 2014](https://md.research.elamur.ai/free/en/papers/litkowski-2014-pattern-dictionary-english-prepositions/)) support tagging and semantic analysis. Hajič compared hand-built dictionaries with training data for morphological tagging ([Hajic 2000](https://md.research.elamur.ai/free/en/papers/hajic-2000-morphological-tagging-data-vs-dictionaries/)).

"Dictionary learning" in sparse autoencoders, which decomposes neural activations into feature directions, is a separate and unrelated use of the word ([Rajamanoharan et al. 2024](https://md.research.elamur.ai/free/en/papers/rajamanoharan-2024-improving-dictionary-learning-gated-sparse/)).

## Open questions

Sense division is the first. Purely distributional scores cannot tell compositional pairs from idioms ([Church & Hanks 1990](https://md.research.elamur.ai/free/en/papers/church-1990-word-association-norms-mutual-information/)). Boas argues that splitting senses by frame gives finer distinctions than EuroWordNet-style senses ([Boas 2002](https://md.research.elamur.ai/free/en/papers/boas-2002-bilingual-framenet-dictionaries-machine-translation/)). Still, there is no agreement on how fine senses should be or how they should map across languages.

The second issue is how far automatic construction can be trusted. Unsupervised bilingual induction assumes that embedding spaces in different languages are nearly isomorphic, that is, close enough in shape for one to be mapped onto the other. That assumption often fails. With a fully unsupervised method, English–Finnish reached a P@1 of 0 in every frequency band and word group analysed. P@1 is the share of test words whose top-ranked translation is correct ([Søgaard et al. 2018](https://md.research.elamur.ai/free/en/papers/sogaard-2018-limitations-unsupervised-bilingual-dictionary-induction/)). Common benchmarks for this task may also give misleading results ([Kementchedjhieva et al. 2019](https://md.research.elamur.ai/free/en/papers/kementchedjhieva-2019-lost-in-evaluation/)). Pivot methods are vulnerable to polysemous pivot words ([Acs 2014](https://md.research.elamur.ai/free/en/papers/acs-2014-wikt2dict-pivot/)).

The third issue is quality control at scale. When there are hundreds of corpora, manual checking is impossible, so monitoring has to rely on statistical signals ([Goldhahn et al. 2012](https://md.research.elamur.ai/free/en/papers/goldhahn-2012-building-large-monolingual-dictionaries-leipzig/)).

Finally, the role of the dictionary is changing. As language models absorb lexical knowledge, it is unclear whether dictionaries will remain mainly reference works for people or become structured inputs for machines. Atkins's call to design entries around user needs ([Atkins 1996](https://md.research.elamur.ai/free/en/papers/atkins-1996-bilingual-dictionaries/)) may need to be rethought for readers that are not human.

## Key papers

The works this page cites, the most cited first; citations as counted by OpenAlex plus the papers of this library that cite the work, read 2026-10-23.

- [Word Association Norms, Mutual Information, and Lexicography](https://md.research.elamur.ai/free/en/papers/church-1990-word-association-norms-mutual-information/) — Kenneth Ward Church, Patrick Hanks · 1990 · 4,363 citations
- [WordNet: A Lexical Database for English](https://md.research.elamur.ai/free/en/papers/miller-1994-wordnet/) — George A. Miller · 1994 · 2,179 citations
- [Building Large Monolingual Dictionaries at the Leipzig Corpora Collection: From 100 to 200 Languages](https://md.research.elamur.ai/free/en/papers/goldhahn-2012-building-large-monolingual-dictionaries-leipzig/) — Dirk Goldhahn, Thomas Eckart, Uwe Quasthoff · 2012 · 533 citations
- [On the Limitations of Unsupervised Bilingual Dictionary Induction](https://md.research.elamur.ai/free/en/papers/sogaard-2018-limitations-unsupervised-bilingual-dictionary-induction/) — Anders Søgaard, Sebastian Ruder, Ivan Vulić · 2018 · 396 citations
- [A Lexicographer-Friendly Association Score](https://md.research.elamur.ai/free/en/papers/rychly-2008-logdice/) — Pavel Rychlý · 2008 · 312 citations
- [Learning Named Entity Tagger using Domain-Specific Dictionary](https://md.research.elamur.ai/free/en/papers/shang-2018-learning-named-entity-tagger-domain/) — Jingbo Shang, Liyuan Liu, Xiaotao Gu et al. · 2018 · 301 citations
- [GDEX: Automatically Finding Good Dictionary Examples in a Corpus](https://md.research.elamur.ai/free/en/papers/kilgarriff-2008-gdex-examples/) — Adam Kilgarriff, Miloš Husák, Katy McAdam et al. · 2008 · 238 citations
- [Learning to Understand Phrases by Embedding the Dictionary](https://md.research.elamur.ai/free/en/papers/hill-2016-learning-understand-phrases-embedding-dictionary/) — Felix Hill, Kyunghyun Cho, Anna Korhonen et al. · 2016 · 234 citations
- [Disambiguating Prepositional Phrase Attachments by Using On-Line Dictionary Definitions](https://md.research.elamur.ai/free/en/papers/jensen-1987-disambiguating-prepositional-phrase-attachments-line/) — Karen Jensen, Jean-Louis Binot · 1987 · 116 citations
- [Corpus Based PP Attachment Ambiguity Resolution with a Semantic Dictionary](https://md.research.elamur.ai/free/en/papers/stetina-1997-corpus-based-pp-attachment-ambiguity/) — Jiri Stetina, Makoto Nagao · 1997 · 115 citations
- [Bilingual Dictionaries: Past, Present and Future](https://md.research.elamur.ai/free/en/papers/atkins-1996-bilingual-dictionaries/) — B. T. S. Atkins · 1996 · 107 citations
- [Bilingual FrameNet Dictionaries for Machine Translation](https://md.research.elamur.ai/free/en/papers/boas-2002-bilingual-framenet-dictionaries-machine-translation/) — Hans C. Boas · 2002 · 106 citations
- [Dictionary-based Phrase-level Prompting of Large Language Models for Machine Translation](https://md.research.elamur.ai/free/en/papers/ghazvininejad-2023-dictionary-based-phrase-level-prompting/) — Marjan Ghazvininejad, Hila Gonen, Luke Zettlemoyer · 2023 · 95 citations
- [Processing Dictionary Definitions with Phrasal Pattern Hierarchies](https://md.research.elamur.ai/free/en/papers/alshawi-1987-processing-dictionary-definitions-phrasal-pattern/) — Hiyan Alshawi · 1987 · 84 citations
- [Chain-of-Dictionary Prompting Elicits Translation in Large Language Models](https://md.research.elamur.ai/free/en/papers/lu-2024-chain-dictionary-prompting-elicits-translation/) — Hongyuan Lu, Haoran Yang, Haoyang Huang et al. · 2024 · 81 citations
- [Morphological Tagging: Data vs. Dictionaries](https://md.research.elamur.ai/free/en/papers/hajic-2000-morphological-tagging-data-vs-dictionaries/) — Jan Hajic · 2000 · 55 citations
- [Construction of a Bilingual Dictionary Intermediated by a Third Language](https://md.research.elamur.ai/free/en/papers/tanaka-1994-bilingual-dictionary-third-language/) — Kumiko Tanaka, Kyoji Umemura · 1994 · 41 citations
- [Improving Dictionary Learning with Gated Sparse Autoencoders](https://md.research.elamur.ai/free/en/papers/rajamanoharan-2024-improving-dictionary-learning-gated-sparse/) — Senthooran Rajamanoharan, Arthur Conmy, Lewis Smith et al. · 2024 · 39 citations
- [Semeval-2022 Task 1: CODWOE – Comparing Dictionaries and Word Embeddings](https://md.research.elamur.ai/free/en/papers/mickus-2022-codwoe-dictionaries-embeddings/) — Timothee Mickus, Kees Van Deemter, Mathieu Constant et al. · 2022 · 38 citations
- [Pattern Dictionary of English Prepositions](https://md.research.elamur.ai/free/en/papers/litkowski-2014-pattern-dictionary-english-prepositions/) — Ken Litkowski · 2014 · 34 citations
- [Construction of English MWE Dictionary and its Application to POS Tagging](https://md.research.elamur.ai/free/en/papers/shigeto-2013-construction-english-mwe-dictionary-its/) — Yutaro Shigeto, Ai Azuma, Sorami Hisamoto et al. · 2013 · 34 citations
- [Compiling a Massive, Multilingual Dictionary via Probabilistic Inference](https://md.research.elamur.ai/free/en/papers/mausam-2009-panlingual-dictionary/) — Mausam, Stephen Soderland, Oren Etzioni · 2009 · 30 citations
- [Pivot-based multilingual dictionary building using Wiktionary](https://md.research.elamur.ai/free/en/papers/acs-2014-wikt2dict-pivot/) — Judit Acs · 2014 · 27 citations
- [A Dictionary and Morphological Analyser for English](https://md.research.elamur.ai/free/en/papers/russell-1986-dictionary-morphological-analyser-english/) — G.J. Russell, S.G. Pulman, G.D. Ritchie et al. · 1986 · 27 citations
- [Wiktionary: A new rival for expert-built lexicons? Exploring the possibilities of collaborative lexicography](https://md.research.elamur.ai/free/en/papers/meyer-2012-wiktionary-new-rival/) — Christian M. Meyer, Iryna Gurevych · 2012 · 24 citations
- [Lost in Evaluation: Misleading Benchmarks for Bilingual Dictionary Induction](https://md.research.elamur.ai/free/en/papers/kementchedjhieva-2019-lost-in-evaluation/) — Yova Kementchedjhieva, Mareike Hartmann, Anders Sgaard · 2019 · 23 citations
- [TEI-Lex0 Guidelines for the Encoding of Dictionary Information on Written and Spoken Forms](https://md.research.elamur.ai/free/en/papers/banski-2017-tei-lex0-forms/) — Piotr Bański, Jack Bowers, Tomaž Erjavec · 2017 · 16 citations
- [Developing LMF-XML Bilingual Dictionaries for Colloquial Arabic Dialects](https://md.research.elamur.ai/free/en/papers/graff-2012-developing-lmf-xml-bilingual-dictionaries/) — David Graff, Mohamed Maamouri · 2012 · 13 citations
- [FrameNet Translation Using Bilingual Dictionaries with Evaluation on the English-French Pair](https://md.research.elamur.ai/free/en/papers/mouton-2010-framenet-translation-bilingual-dictionaries-evaluation/) — Claire Mouton, Gaël de Chalendar, Benoit Richert · 2010 · 12 citations
- [Automatic Construction of a Japanese-Chinese Dictionary via English](https://md.research.elamur.ai/free/en/papers/kaji-2008-automatic-construction-japanese-chinese-dictionary/) — Hiroyuki Kaji, Shin'ichi Tamamura, Dashtseren Erdenebat · 2008 · 11 citations

## Related concepts

- [Lexical semantics](https://md.research.elamur.ai/free/en/concepts/lexical-semantics/) — Lexical semantics is the study of word meaning. It asks what words and multiword expressions denote, how a word's senses are organized, how…
- [Corpus linguistics](https://md.research.elamur.ai/free/en/concepts/corpus-linguistics/) — Corpus linguistics is the empirical study of language through large, principled collections of naturally occurring text or transcribed…
- [Word sense disambiguation](https://md.research.elamur.ai/free/en/concepts/word-sense-disambiguation/) — Word sense disambiguation (WSD) identifies which meaning of an ambiguous word is intended in a given context. It is usually framed as…
- [Ontology engineering](https://md.research.elamur.ai/free/en/concepts/ontology-engineering/) — Ontology engineering is the discipline that studies how to design, build, evaluate, maintain and reuse ontologies. An ontology here is an…

## More papers

Topics: [📖 Lexical semantics & word embeddings](https://md.research.elamur.ai/free/en/topics/lexical-semantics/) · [🧷 Idioms & multiword expressions](https://md.research.elamur.ai/free/en/topics/multiword-expressions/)

Shelves: [🗣️ Linguistics & languages](https://md.research.elamur.ai/free/en/shelves/linguistics/)

- [← previous concept Lexical semantics](https://md.research.elamur.ai/free/en/concepts/lexical-semantics/)
- [next concept → Minimum description length](https://md.research.elamur.ai/free/en/concepts/minimum-description-length/)
- [↑ back to concepts](https://md.research.elamur.ai/free/en/concepts/)