--- title: "Using Lexical Chains for Text Summarization — Barzilay 1997" description: "The paper looks for a middle path in automatic summarization." authors: ["Regina Barzilay", "Michael Elhadad"] year: 1997 venue: "Intelligent Scalable Text Summarization" source: "https://aclanthology.org/W97-0703/" 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/barzilay-1997-lexical-chains-text-summarization/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/barzilay-1997-lexical-chains-text-summarization/) [home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [1997](https://md.research.elamur.ai/free/en/papers/1997/) / Using Lexical Chains for Text Summarization # Using Lexical Chains for Text Summarization **Authors:** [Regina Barzilay](https://md.research.elamur.ai/free/en/authors/regina-barzilay/), [Michael Elhadad](https://md.research.elamur.ai/free/en/authors/michael-elhadad/) · **Year:** 1997 · **Venue:** Intelligent Scalable Text Summarization · **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/W97-0703/) · [PDF](https://aclanthology.org/W97-0703.pdf) · [OpenAlex](https://openalex.org/W1573514622) **Citations:** 855 citations (openalex-snapshot · 2026-10-14) · [134](#cited-by) in our library, of the 140884 papers whose bibliographies we read ## Summary The paper looks for a middle path in automatic summarization. Full semantic interpretation of a text is too costly for arbitrary input. Plain word-frequency tables cannot tell what a text is really about. The authors use lexical chains instead: sequences of semantically related words that trace topic progression. They build these chains from widely available resources: the WordNet thesaurus, a part-of-speech tagger, a shallow parser that identifies nominal groups, and a segmentation algorithm adapted from Hearst (1994). Summarization runs as a pipeline: 1. Segment the text. 2. Build the chains. 3. Select strong chains. 4. Extract the sentences that represent them. A key contribution is a new chain-construction algorithm. It avoids the greedy word-sense disambiguation used by Hirst and St-Onge, which commits to a sense too early and can be misled. The paper reports empirical results on identifying strong chains and significant sentences, producing indicative summaries. This matters because it shows that useful summaries can come from shallow, robust text structure without deep understanding. ## Key points - Lexical cohesion builds on Halliday and Hasan (1976) and Hoey (1991). It is treated as the most frequent and most easily detectable kind of cohesion, unlike coherence relations (as in RST), which are hard to compute. - Earlier work by Morris and Hirst (1991) used Roget's Thesaurus and covered over 90% of intuitive lexical relations, but was never implemented. - WordNet has more than 118,000 word forms, and about 17% of them are polysemous. Stairmand notes that roughly half of the nouns in news text are polysemous. - Hirst and St-Onge use three relation types: - extra-strong: repetition, with no distance limit; - strong: a direct WordNet relation, within a 7-sentence window; - medium-strong: a longer WordNet path, within 3 sentences. - In a worked example on a Bayesian-networks article, the strongest chains scored 14.0 (Microsoft/company), 9.0 (Bayesian system/network), 7.0, 6.0 and 3.0. Multiword nominal groups such as "bayesian-network" appear as chain members. ## Interesting ideas - **Delay disambiguation until context accumulates.** The "machine"/"person" example shows the failure of greedy sense choice: the wrong sense is locked in before later words like "device" and "pump" can correct it. Keeping several sense interpretations alive applies to any incremental NLP decision. - **Use cohesion as a cheap stand-in for discourse structure.** Related words tend to cluster within discourse units, so chains give a computable signal where full rhetorical analysis is out of reach. This is attractive for low-resource languages that have a WordNet-like resource but no discourse parser. - **Treat noun groups, not single words, as units.** The shallow parser lets compounds like "bayesian-network" join chains as single terms. This sharpens topic signals in technical and domain texts. - **Score topics, then extract sentences.** Ranking whole chains first and only then picking representative sentences separates topic detection from extraction. Each stage can be evaluated or replaced on its own. *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: [Natural language processing](https://md.research.elamur.ai/free/en/shelves/nlp/) Topics: [Word sense disambiguation](https://md.research.elamur.ai/free/en/topics/word-sense-disambiguation/) ## Cited by Cited by 134 papers of our library cite it (of the 140884 whose bibliographies we read); 11 of them have a page here, 123 not yet. - [Random Walks for Knowledge-Based Word Sense Disambiguation](https://md.research.elamur.ai/free/en/papers/agirre-2014-random-walks-knowledge-based-word/) - [Using Coreference Chains for Text Summarization](https://md.research.elamur.ai/free/en/papers/azzam-1999-coreference-chains-text-summarization/) - [Information Fusion in the Context of Multi-Document Summarization](https://md.research.elamur.ai/free/en/papers/barzilay-1999-information-fusion-context-multi-document/) - [Extracting Paraphrases from a Parallel Corpus](https://md.research.elamur.ai/free/en/papers/barzilay-2001-extracting-paraphrases-parallel-corpus/) - [Sentence Fusion for Multidocument News Summarization](https://md.research.elamur.ai/free/en/papers/barzilay-2005-sentence-fusion-multidocument-news-summarization/) - [CODRA: A Novel Discriminative Framework for Rhetorical Analysis](https://md.research.elamur.ai/free/en/papers/joty-2015-codra-novel-discriminative-framework-rhetorical/) - [Ranking Sentences for Extractive Summarization with Reinforcement Learning](https://md.research.elamur.ai/free/en/papers/narayan-2018-ranking-sentences-extractive-summarization-reinforcement/) - [Planning with Learned Entity Prompts for Abstractive Summarization](https://md.research.elamur.ai/free/en/papers/narayan-2021-planning-learned-entity-prompts-abstractive/) - [A Critique and Improvement of an Evaluation Metric for Text Segmentation](https://md.research.elamur.ai/free/en/papers/pevzner-2002-critique-improvement-evaluation-metric-text/) - [Lexical Chaining for Measuring Discourse Coherence Quality in Test-taker Essays](https://md.research.elamur.ai/free/en/papers/somasundaran-2014-lexical-chaining-measuring-discourse-coherence/) - [Unsupervised Recognition of Literal and Non-Literal Use of Idiomatic Expressions](https://md.research.elamur.ai/free/en/papers/sporleder-2009-unsupervised-recognition-literal-non-literal/) ## Cite The paper itself, not this page. BibTeX from the library's record of the work. ```bibtex @inproceedings{barzilay1997using, title = {{Using Lexical Chains for Text Summarization}}, author = {Barzilay, Regina and Elhadad, Michael}, booktitle = {Intelligent Scalable Text Summarization}, year = {1997}, doi = {10.7916/d85b09vz}, url = {https://doi.org/10.7916/d85b09vz} } ``` - [← previous Towards resolution of bridging descriptions](https://md.research.elamur.ai/free/en/papers/vieira-1997-resolution-bridging-descriptions/) - [next → What brings a higher-order entity into focus of attention? Sentential pronouns in English and Norwegian.](https://md.research.elamur.ai/free/en/papers/borthen-1997-what-brings-higher-order-entity/) - [↑ back to papers of 1997](https://md.research.elamur.ai/free/en/papers/1997/) ## References 6 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet. 1. Black, Wflham J 1994 Parsing, lmgmstlc resources and semantic analysis, for abstzactmg and categorization Halhday, lqhchael and Ruqatya Hasan 1976 Cohesion in Enghsh Longman, London Hearst, Marti A 1994 Multi-paragraph segmentation of exposltoZT text In Proceedmgs of the 3~nd Annual Meetmg of the As… [find](https://scholar.google.com/scholar?q=Black%2C+Wflham+J+1994+Parsing%2C+lmgmstlc+resources+and+semantic+analysis%2C+for+abstzactmg+and+categorization+Halhday%2C+lqhchael+and+Ruqatya+Hasan+1976+Cohesion+in+Enghsh+Longman%2C+London+Hearst%2C+Marti+A+19) 2. McKeown, Kathleen and Dragonur Radev 1995 Gen-eratmg summaries of multiple news articles In SIGIR 95 Proceedings Mdler, George A , P,~chard Beckwlth, Chnstiane Fell-bans, Derek Gross, and Kathenne J 1990 Introduction to WordNet An on-lme lexxcal database Internat,onal Journal of Lazwcographg (specia… [find](https://scholar.google.com/scholar?q=McKeown%2C+Kathleen+and+Dragonur+Radev+1995+Gen%2Deratmg+summaries+of+multiple+news+articles+In+SIGIR+95+Proceedings+Mdler%2C+George+A+%2C+P%2C%7Echard+Beckwlth%2C+Chnstiane+Fell%2Dbans%2C+Derek+Gross%2C+and+Kathenne+J+1) 3. Morns, J and G Hzrst 1991 Lexzcal coheszon computed by thesanra] relations as an re&cater of the structure of the text Computahonal Lmgutsttos, 17(1) pp 21.--45 [find](https://scholar.google.com/scholar?q=Morns%2C+J+and+G+Hzrst+1991+Lexzcal+coheszon+computed+by+thesanra%5D+relations+as+an+re%26cater+of+the+structure+of+the+text+Computahonal+Lmgutsttos%2C+17%281%29+pp+21%2E%2D%2D45) 4. Onq, Ken3h Sunuta Ks•no, and Mnke Seql 1994 Abstract genezation based on rhetorical structure extruc-tzon In Proceedings of the International Conference on Computahonal Lmgutshcs (Cohng 9~), pages 344-- 348, Japan Pa~ce, C D and G D Husk 1991 Towards the automatic recognntlon of anaphonc features m… [find](https://scholar.google.com/scholar?q=Onq%2C+Ken3h+Sunuta+Ks%E2%80%A2no%2C+and+Mnke+Seql+1994+Abstract+genezation+based+on+rhetorical+structure+extruc%2Dtzon+In+Proceedings+of+the+International+Conference+on+Computahonal+Lmgutshcs+%28Cohng+9%7E%29%2C+pages+344) 5. Patce, Chris D 1990 Constructing hterature abstracts by computer techmqu~ and prospects lnformahon Processmg and Management, 26(1) 171-186 [find](https://scholar.google.com/scholar?q=Patce%2C+Chris+D+1990+Constructing+hterature+abstracts+by+computer+techmqu%7E+and+prospects+lnformahon+Processmg+and+Management%2C+26%281%29+171%2D186) 6. Statrmand, Mark A 1996 A Computahonal Analysts of Lexscal Cohesion wtth Apphcattons :n Informatton Re-trievai Ph D thesis, Center for Computational Lm-gmstics, UMIST, Manchester [find](https://scholar.google.com/scholar?q=Statrmand%2C+Mark+A+1996+A+Computahonal+Analysts+of+Lexscal+Cohesion+wtth+Apphcattons+%3An+Informatton+Re%2Dtrievai+Ph+D+thesis%2C+Center+for+Computational+Lm%2Dgmstics%2C+UMIST%2C+Manchester)