{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,18]],"date-time":"2026-08-18T01:55:38Z","timestamp":1787018138060,"version":"build-2736575974"},"reference-count":25,"publisher":"MIT Press","issue":"3","license":[{"start":{"date-parts":[[2024,6,10]],"date-time":"2024-06-10T00:00:00Z","timestamp":1717977600000},"content-version":"vor","delay-in-days":161,"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":[[2024,9,1]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>What do language models (LMs) do with language? They can produce sequences of (mostly) coherent strings closely resembling English. But do those sentences mean something, or are LMs simply babbling in a convincing simulacrum of language use? We address one aspect of this broad question: whether LMs\u2019 words can refer, that is, achieve \u201cword-to-world\u201d connections. There is prima facie reason to think they do not, since LMs do not interact with the world in the way that ordinary language users do. Drawing on the externalist tradition in philosophy of language, we argue that those appearances are misleading: Even if the inputs to LMs are simply strings of text, they are strings of text with natural histories, and that may suffice for LMs\u2019 words to refer.<\/jats:p>","DOI":"10.1162\/coli_a_00522","type":"journal-article","created":{"date-parts":[[2024,6,10]],"date-time":"2024-06-10T10:18:24Z","timestamp":1718014704000},"page":"1191-1200","update-policy":"https:\/\/doi.org\/10.1162\/mitpressjournals.corrections.policy","source":"Crossref","is-referenced-by-count":34,"title":["Do Language Models\u2019 Words Refer?"],"prefix":"10.1162","volume":"50","author":[{"given":"Matthew","family":"Mandelkern","sequence":"first","affiliation":[{"name":"Department of Philosophy, New York University. mandelkern@nyu.edu"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tal","family":"Linzen","sequence":"additional","affiliation":[{"name":"Department of Linguistics and, Center for Data Science, New York University. linzen@nyu.edu"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"281","published-online":{"date-parts":[[2024,9,1]]},"reference":[{"key":"2024092014244234300_bib1","doi-asserted-by":"publisher","first-page":"5769","DOI":"10.18653\/v1\/2022.findings-emnlp.423","article-title":"Language models as agent models","volume-title":"Findings of the Association for Computational Linguistics: EMNLP 2022","author":"Andreas","year":"2022"},{"key":"2024092014244234300_bib2","doi-asserted-by":"publisher","first-page":"610","DOI":"10.1145\/3442188.3445922","article-title":"On the dangers of stochastic parrots: Can language models be too big?","volume-title":"Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency","author":"Bender","year":"2021"},{"key":"2024092014244234300_bib3","doi-asserted-by":"publisher","first-page":"5185","DOI":"10.18653\/v1\/2020.acl-main.463","article-title":"Climbing towards NLU: On meaning, form, and understanding in the age of data","volume-title":"Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics","author":"Bender","year":"2020"},{"key":"2024092014244234300_bib4","doi-asserted-by":"publisher","first-page":"3079","DOI":"10.1007\/s10670-021-00491-w","article-title":"Sharing our concepts with machines","volume":"88","author":"Butlin","year":"2021","journal-title":"Erkenntnis"},{"key":"2024092014244234300_bib5","doi-asserted-by":"publisher","DOI":"10.1093\/oso\/9780192894724.001.0001","volume-title":"Making AI Intelligible","author":"Cappelen","year":"2021"},{"key":"2024092014244234300_bib6","article-title":"Visualizing and measuring the geometry of BERT","author":"Coenen","year":"2019","journal-title":"Arxiv preprint arxiv 1906.02715"},{"key":"2024092014244234300_bib7","first-page":"747","article-title":"Societal grounding is essential to meaningful language use","volume-title":"Proceedings of the 21st National Conference on Artificial Intelligence and the 18th Innovative Applications of Artificial Intelligence Conference, AAAI-06\/IAAI-06","author":"DeVault","year":"2006"},{"key":"2024092014244234300_bib8","volume-title":"The Varieties of Reference","author":"Evans","year":"1982"},{"key":"2024092014244234300_bib9","doi-asserted-by":"publisher","first-page":"825","DOI":"10.1162\/tacl_a_00400","article-title":"Let\u2019s play mono-poly: BERT can reveal words\u2019 polysemy level and partitionability into senses","volume":"9","author":"Gar\u00ed Soler","year":"2021","journal-title":"Transactions of the Association for Computational Linguistics"},{"issue":"9","key":"2024092014244234300_bib10","doi-asserted-by":"publisher","first-page":"486","DOI":"10.5840\/2011108925","article-title":"What are words? Comments on Kaplan (1990), on Hawthorne and Lepore, and on the issue","volume":"108","author":"Hawthorne","year":"2011","journal-title":"Journal of Philosophy"},{"issue":"1","key":"2024092014244234300_bib11","doi-asserted-by":"publisher","first-page":"93","DOI":"10.1093\/aristoteliansupp\/64.1.93","article-title":"Words","volume":"64","author":"Kaplan","year":"1990","journal-title":"Aristotelian Society Supplementary Volume"},{"key":"2024092014244234300_bib12","doi-asserted-by":"publisher","first-page":"255","DOI":"10.1111\/j.1475-4975.1977.tb00045.x","article-title":"Speaker\u2019s reference and semantic reference","volume":"29","author":"Kripke","year":"1977","journal-title":"Midwest Studies in Philosophy"},{"key":"2024092014244234300_bib13","volume-title":"Naming and Necessity","author":"Kripke","year":"1980"},{"key":"2024092014244234300_bib14","unstructured":"Lederman, Harvey and KyleMahowald. 2024. Are language models more like libraries or like librarians? Bibliotechnism, the novel reference problem, and the attitudes of LLMs. Manuscript, University of Texas at Austin."},{"key":"2024092014244234300_bib15","doi-asserted-by":"publisher","first-page":"176","DOI":"10.18653\/v1\/2022.conll-1.13","article-title":"Entailment semantics can be extracted from an ideal language model","volume-title":"Proceedings of the 26th Conference on Computational Natural Language Learning (CoNLL)","author":"Merrill","year":"2022"},{"key":"2024092014244234300_bib16","article-title":"The vector grounding problem","author":"Mollo","year":"2023","journal-title":"arXiv preprint arXiv:2304.01481"},{"key":"2024092014244234300_bib17","doi-asserted-by":"publisher","DOI":"10.1017\/CBO9781139151269.002","volume-title":"Speech Communities","author":"Morgan","year":"2014"},{"key":"2024092014244234300_bib18","unstructured":"Ostertag, Gary\n          . 2024. Large language models and externalism about reference: Some negative results. Manuscript, Icahn School of Medicine and Mount Sinai & The Graduate Center, CUNY."},{"key":"2024092014244234300_bib19","doi-asserted-by":"publisher","first-page":"573","DOI":"10.1002\/9780470756591.ch23","article-title":"The speech community","volume-title":"The Handbook of Language Variation and Change","author":"Patrick","year":"2004"},{"issue":"381","key":"2024092014244234300_bib20","doi-asserted-by":"publisher","first-page":"20220041","DOI":"10.1098\/rsta.2022.0041","article-title":"Symbols and grounding in large language models","volume":"28","author":"Pavlick","year":"2023","journal-title":"Philosophical Transactions of the Royal Society A"},{"key":"2024092014244234300_bib21","article-title":"Meaning without reference in large language models","author":"Piantadosi","year":"2022"},{"key":"2024092014244234300_bib22","article-title":"Is it possible for language models to achieve language understanding?","author":"Potts","year":"2020"},{"key":"2024092014244234300_bib23","doi-asserted-by":"publisher","first-page":"131","DOI":"10.1017\/CBO9780511625251.014","article-title":"The meaning of \u2018meaning\u2019","volume-title":"Language, Mind, and Knowledge","author":"Putnam","year":"1975"},{"key":"2024092014244234300_bib24","doi-asserted-by":"publisher","DOI":"10.1017\/CBO9780511625398","volume-title":"Reason, Truth, and History","author":"Putnam","year":"1981"},{"issue":"101174","key":"2024092014244234300_bib25","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1016\/j.cogsys.2023.101174","article-title":"Does ChatGPT have semantic understanding?","volume":"83","author":"Titus","year":"2024","journal-title":"Cognitive Systems Research"}],"container-title":["Computational Linguistics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/direct.mit.edu\/coli\/article-pdf\/50\/3\/1191\/2470888\/coli_a_00522.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/direct.mit.edu\/coli\/article-pdf\/50\/3\/1191\/2470888\/coli_a_00522.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,9,20]],"date-time":"2024-09-20T10:24:51Z","timestamp":1726827891000},"score":1,"resource":{"primary":{"URL":"https:\/\/direct.mit.edu\/coli\/article\/50\/3\/1191\/121670\/Do-Language-Models-Words-Refer"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024]]},"references-count":25,"journal-issue":{"issue":"3","published-online":{"date-parts":[[2024,9,1]]},"published-print":{"date-parts":[[2024,9,1]]}},"URL":"https:\/\/doi.org\/10.1162\/coli_a_00522","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]]}}}