Why Large Language Models Are Poor Theories of Human Linguistic Cognition: A Reply to Piantadosi
In a recent manuscript entitled “Modern language models refute Chomsky’s approach to language”, Steven Piantadosi proposes that large language models such as GPT-3 can serve as serious theories of human linguistic cognition. In fact, he maintains that these models are significantly better linguistic...
Main Author: | |
---|---|
Format: | Article |
Language: | English |
Published: |
PsychOpen GOLD/ Leibniz Institute for Psychology
2023-12-01
|
Series: | Biolinguistics |
Subjects: | |
Online Access: | https://doi.org/10.5964/bioling.13153 |
_version_ | 1797338144899596288 |
---|---|
author | Roni Katzir |
author_facet | Roni Katzir |
author_sort | Roni Katzir |
collection | DOAJ |
description | In a recent manuscript entitled “Modern language models refute Chomsky’s approach to language”, Steven Piantadosi proposes that large language models such as GPT-3 can serve as serious theories of human linguistic cognition. In fact, he maintains that these models are significantly better linguistic theories than proposals emerging from within generative linguistics. The present note explains why this claim is wrong. |
first_indexed | 2024-03-08T09:26:46Z |
format | Article |
id | doaj.art-3f6c37e2ef7240a29b4cf44c38f85518 |
institution | Directory Open Access Journal |
issn | 1450-3417 |
language | English |
last_indexed | 2024-03-08T09:26:46Z |
publishDate | 2023-12-01 |
publisher | PsychOpen GOLD/ Leibniz Institute for Psychology |
record_format | Article |
series | Biolinguistics |
spelling | doaj.art-3f6c37e2ef7240a29b4cf44c38f855182024-01-31T08:19:46ZengPsychOpen GOLD/ Leibniz Institute for PsychologyBiolinguistics1450-34172023-12-011710.5964/bioling.13153bioling.13153Why Large Language Models Are Poor Theories of Human Linguistic Cognition: A Reply to PiantadosiRoni Katzir0https://orcid.org/0000-0002-0241-1896Department of Linguistics, Tel Aviv University, Tel Aviv, IsraelIn a recent manuscript entitled “Modern language models refute Chomsky’s approach to language”, Steven Piantadosi proposes that large language models such as GPT-3 can serve as serious theories of human linguistic cognition. In fact, he maintains that these models are significantly better linguistic theories than proposals emerging from within generative linguistics. The present note explains why this claim is wrong.https://doi.org/10.5964/bioling.13153large language modelsgenerative linguisticslearninggeneralizationtypologycompetenceperformance |
spellingShingle | Roni Katzir Why Large Language Models Are Poor Theories of Human Linguistic Cognition: A Reply to Piantadosi Biolinguistics large language models generative linguistics learning generalization typology competence performance |
title | Why Large Language Models Are Poor Theories of Human Linguistic Cognition: A Reply to Piantadosi |
title_full | Why Large Language Models Are Poor Theories of Human Linguistic Cognition: A Reply to Piantadosi |
title_fullStr | Why Large Language Models Are Poor Theories of Human Linguistic Cognition: A Reply to Piantadosi |
title_full_unstemmed | Why Large Language Models Are Poor Theories of Human Linguistic Cognition: A Reply to Piantadosi |
title_short | Why Large Language Models Are Poor Theories of Human Linguistic Cognition: A Reply to Piantadosi |
title_sort | why large language models are poor theories of human linguistic cognition a reply to piantadosi |
topic | large language models generative linguistics learning generalization typology competence performance |
url | https://doi.org/10.5964/bioling.13153 |
work_keys_str_mv | AT ronikatzir whylargelanguagemodelsarepoortheoriesofhumanlinguisticcognitionareplytopiantadosi |