{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,28]],"date-time":"2026-08-28T09:06:34Z","timestamp":1787907994075,"version":"build-2784847793"},"reference-count":32,"publisher":"MIT Press","issue":"2","license":[{"start":{"date-parts":[[2024,1,19]],"date-time":"2024-01-19T00:00:00Z","timestamp":1705622400000},"content-version":"vor","delay-in-days":18,"URL":"https:\/\/creativecommons.org\/licenses\/by-nc-nd\/4.0\/"}],"content-domain":{"domain":["direct.mit.edu"],"crossmark-restriction":true},"short-container-title":[],"published-print":{"date-parts":[[2024,6,1]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:p>Despite impressive advances in Natural Language Generation (NLG) and Large Language Models (LLMs), researchers are still unclear about important aspects of NLG evaluation. To substantiate this claim, I examine current classifications of hallucination and omission in data-text NLG, and I propose a logic-based synthesis of these classfications. I conclude by highlighting some remaining limitations of all current thinking about hallucination and by discussing implications for LLMs.<\/jats:p>","DOI":"10.1162\/coli_a_00509","type":"journal-article","created":{"date-parts":[[2024,1,19]],"date-time":"2024-01-19T19:51:06Z","timestamp":1705693866000},"page":"807-816","update-policy":"https:\/\/doi.org\/10.1162\/mitpressjournals.corrections.policy","source":"Crossref","is-referenced-by-count":13,"title":["The Pitfalls of Defining Hallucination"],"prefix":"10.1162","volume":"50","author":[{"given":"Kees van","family":"Deemter","sequence":"first","affiliation":[{"name":"Utrecht University, Department of Information and Computing Sciences. c.j.vandeemter@uu.nl"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"281","published-online":{"date-parts":[[2024,6,1]]},"reference":[{"key":"2024070814302799800_bib1","volume-title":"Intention, Plans, and Practical Reason","author":"Bratman","year":"1987"},{"key":"2024070814302799800_bib2","article-title":"Sparks of artificial general intelligence:\n                        Early experiments with GPT-4","author":"Bubeck","year":"2023","journal-title":"arXiv preprint\n                        arXiv:2303.12712"},{"key":"2024070814302799800_bib3","doi-asserted-by":"publisher","first-page":"70977","DOI":"10.1109\/ACCESS.2023.3294090","article-title":"Machine-generated text: A comprehensive\n                        survey of threat models and detection methods","volume":"11","author":"Crothers","year":"2023","journal-title":"IEEE\n                        Access"},{"key":"2024070814302799800_bib4","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2023.acl-long.3","article-title":"Detecting and mitigating hallucinations in\n                        machine translation: Model internal workings alone do well, sentence\n                        similarity even better","author":"Dale","year":"2022","journal-title":"arXiv preprint\n                        arXiv:2212.08597"},{"key":"2024070814302799800_bib5","doi-asserted-by":"publisher","first-page":"131","DOI":"10.18653\/v1\/2020.inlg-1.19","article-title":"Evaluating semantic accuracy of\n                        data-to-text generation with natural language inference","volume-title":"Proceedings of the 13th International Conference on Natural Language\n                        Generation (INLG)","author":"Du\u0161ek","year":"2020"},{"key":"2024070814302799800_bib6","doi-asserted-by":"publisher","first-page":"1530","DOI":"10.18653\/v1\/2021.findings-emnlp.132","article-title":"Entity-based semantic adequacy for\n                        data-to-text generation","volume-title":"Findings of the\n                        Association for Computational Linguistics: EMNLP 2021","author":"Faille","year":"2021"},{"key":"2024070814302799800_bib7","doi-asserted-by":"publisher","first-page":"44","DOI":"10.15294\/elt.v12i1.64069","article-title":"Artificial intelligence (AI) technology in OpenAI ChatGPT\n                        application: A review of ChatGPT in writing English essay","volume-title":"ELT Forum: Journal of English Language Teaching","author":"Fitria","year":"2023"},{"key":"2024070814302799800_bib8","doi-asserted-by":"publisher","first-page":"65","DOI":"10.1613\/jair.5477","article-title":"Survey of the state of the art in natural\n                        language generation: Core tasks, applications and\n                    evaluation","volume":"61","author":"Gatt","year":"2018","journal-title":"Journal of Artificial Intelligence\n                        Research"},{"key":"2024070814302799800_bib9","doi-asserted-by":"publisher","first-page":"41","DOI":"10.1163\/9789004368811_003","article-title":"Logic and conversation","volume-title":"Speech\n                        Acts","author":"Grice","year":"1975"},{"key":"2024070814302799800_bib10","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2020.coling-main.218","article-title":"Have your text and use it too! End-to-end\n                        neural data-to-text generation with semantic fidelity","author":"Harkous","year":"2020","journal-title":"arXiv preprint arXiv:2004.06577"},{"key":"2024070814302799800_bib11","article-title":"A survey on hallucination in large\n                        language models: Principles, taxonomy, challenges, and open\n                        questions","author":"Huang","year":"2023","journal-title":"arXiv preprint\n                    arXiv:2311.05232"},{"key":"2024070814302799800_bib12","doi-asserted-by":"publisher","first-page":"383","DOI":"10.18653\/v1\/2022.gem-1.33","article-title":"A survey of recent error annotation schemes for automatically\n                        generated text","volume-title":"Proceedings of the 2nd Workshop\n                        on Natural Language Generation, Evaluation, and Metrics (GEM)","author":"Huidrom","year":"2022"},{"issue":"12","key":"2024070814302799800_bib13","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1145\/3571730","article-title":"Survey of hallucination in natural language\n                        generation","volume":"55","author":"Ji","year":"2023","journal-title":"ACM Computing Surveys"},{"key":"2024070814302799800_bib14","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/W17-3204","article-title":"Six challenges for neural machine\n                        translation","author":"Koehn","year":"2017","journal-title":"arXiv preprint\n                        arXiv:1706.03872"},{"key":"2024070814302799800_bib15","doi-asserted-by":"publisher","DOI":"10.1017\/CBO9780511813313","volume-title":"Pragmatics","author":"Levinson","year":"1983"},{"key":"2024070814302799800_bib16","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2023.emnlp-main.51","article-title":"We\u2019re afraid language models aren\u2019t modeling\n                        ambiguity","author":"Liu","year":"2023","journal-title":"arXiv preprint arXiv:2304.14399"},{"key":"2024070814302799800_bib17","article-title":"Data-to-text generation for severely under-resourced\n                        languages with GPT-3.5: A bit of help needed from Google\n                        Translate","author":"Lorandi","year":"2023","journal-title":"arXiv preprint\n                    arXiv:2308.09957"},{"key":"2024070814302799800_bib18","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2020.acl-main.173","article-title":"On faithfulness and factuality in\n                        abstractive summarization","author":"Maynez","year":"2020","journal-title":"arXiv preprint\n                        arXiv:2005.00661"},{"key":"2024070814302799800_bib19","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2022.acl-long.394","article-title":"Human evaluation and correlation with\n                        automatic metrics in consultation note generation","author":"Moramarco","year":"2022","journal-title":"arXiv preprint arXiv:2204.00447"},{"key":"2024070814302799800_bib20","article-title":"Paraconsistent logic","volume-title":"The Stanford Encyclopedia of Philosophy","author":"Priest","year":"2022"},{"key":"2024070814302799800_bib21","article-title":"Summarization is (almost) dead","author":"Pu","year":"2023","journal-title":"arXiv\n                        preprint arXiv:2309.09558"},{"key":"2024070814302799800_bib22","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2021.naacl-main.92","article-title":"The curious case of hallucinations in\n                        neural machine translation","author":"Raunak","year":"2021","journal-title":"arXiv preprint\n                        arXiv:2104.06683"},{"key":"2024070814302799800_bib23","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/D18-1437","article-title":"Object hallucination in image\n                        captioning","author":"Rohrbach","year":"2018","journal-title":"arXiv preprint\n                    arXiv:1809.02156"},{"key":"2024070814302799800_bib24","first-page":"5689","article-title":"FuseCap: Leveraging large language models\n                        for enriched fused image captions","volume-title":"Proceedings of\n                        the IEEE\/CVF Winter Conference on Applications of Computer Vision","author":"Rotstein","year":"2024"},{"key":"2024070814302799800_bib25","doi-asserted-by":"publisher","first-page":"286","DOI":"10.1007\/978-3-642-15675-5_25","article-title":"A logical account of\n                    lying","volume-title":"Logics in Artificial Intelligence: 12th\n                        European Conference, JELIA 2010, Helsinki, Finland, September 13\u201315,\n                        2010. Proceedings 12","author":"Sakama","year":"2010"},{"key":"2024070814302799800_bib26","doi-asserted-by":"publisher","first-page":"158","DOI":"10.18653\/v1\/2020.inlg-1.22","article-title":"A gold standard methodology for evaluating\n                        accuracy in data-to-text systems","volume-title":"Proceedings of\n                        the 13th International Conference on Natural Language Generation","author":"Thomson","year":"2020"},{"key":"2024070814302799800_bib27","volume-title":"Not Exactly: In Praise of Vagueness","author":"Van Deemter","year":"2010"},{"issue":"2","key":"2024070814302799800_bib28","doi-asserted-by":"publisher","first-page":"466","DOI":"10.1111\/tops.12492","article-title":"Editors\u2019 review and introduction:\n                        Lying in logic, language, and cognition","volume":"12","author":"van Ditmarsch","year":"2020","journal-title":"Topics in\n                        Cognitive Science"},{"key":"2024070814302799800_bib29","doi-asserted-by":"publisher","DOI":"10.3384\/nejlt.2000-1533.2023.4529","article-title":"Barriers and enabling factors for error analysis in NLG\n                        research","volume":"11","author":"Van Miltenburg","year":"2023","journal-title":"Northern European Journal of Language\n                        Technology"},{"key":"2024070814302799800_bib30","article-title":"A neural conversational model","author":"Vinyals","year":"2015","journal-title":"CML\n                        Deep Learning Workshop"},{"key":"2024070814302799800_bib31","doi-asserted-by":"publisher","DOI":"10.26615\/978-954-452-092-2_133","article-title":"Evaluating generative models for\n                        graph-to-text generation","author":"Yuan","year":"2023","journal-title":"arXiv preprint\n                        arXiv:2307.14712"},{"key":"2024070814302799800_bib32","article-title":"Siren\u2019s song in the AI ocean: A\n                        survey on hallucination in large language models","author":"Zhang","year":"2023","journal-title":"arXiv preprint arXiv:2309.01219"}],"container-title":["Computational Linguistics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/direct.mit.edu\/coli\/article-pdf\/50\/2\/807\/2457437\/coli_a_00509.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/direct.mit.edu\/coli\/article-pdf\/50\/2\/807\/2457437\/coli_a_00509.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,7,8]],"date-time":"2024-07-08T14:30:41Z","timestamp":1720449041000},"score":1,"resource":{"primary":{"URL":"https:\/\/direct.mit.edu\/coli\/article\/50\/2\/807\/119144\/The-Pitfalls-of-Defining-Hallucination"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024]]},"references-count":32,"journal-issue":{"issue":"2","published-online":{"date-parts":[[2024,6,1]]},"published-print":{"date-parts":[[2024,6,1]]}},"URL":"https:\/\/doi.org\/10.1162\/coli_a_00509","relation":{},"ISSN":["0891-2017","1530-9312"],"issn-type":[{"value":"0891-2017","type":"print"},{"value":"1530-9312","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2024]]},"published":{"date-parts":[[2024]]}}}