mlsanityData to 14 Oct 2026

Can Language Models Be Tricked by Language Illusions? Easier with Syntax, Harder with Semantics

arXiv 2311.01386 ↗First posted 2 Nov 2023cs.CLcited by 11

Not on the map. Why?

Abstract (arXiv)

Language models (LMs) have been argued to overlap substantially with human beings in grammaticality judgment tasks. But when humans systematically make errors in language processing, should we expect LMs to behave like cognitive models of language and mimic human behavior? We answer this question by investigating LMs' more subtle judgments associated with "language illusions" -- sentences that are vague in meaning, implausible, or ungrammatical but receive unexpectedly high acceptability judgments by humans. We looked at three illusions: the comparative illusion (e.g. "More people have been to Russia than I have"), the depth-charge illusion (e.g. "No head injury is too trivial to be ignored"), and the negative polarity item (NPI) illusion (e.g. "The hunter who no villager believed to be trustworthy will ever shoot a bear"). We found that probabilities represented by LMs were more likely to align with human judgments of being "tricked" by the NPI illusion which examines a structural dependency, compared to the comparative and the depth-charge illusions which require sophisticated semantic understanding. No single LM or metric yielded results that are entirely consistent with human behavior. Ultimately, we show that LMs are limited both in their construal as cognitive models of human language processing and in their capacity to recognize nuanced but critical information in complicated language materials.

Key facts. Cited by 11 papers in the map (data through 14 Oct 2026). Builds on 27 of its 62 references that are in the map. First posted on arXiv 2 Nov 2023.

Builds on

27 of 62 references.

References · 27

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  2. 2020Language Models are Few-Shot Learnerscited by 32,419
  3. 2019RoBERTa: A Robustly Optimized BERT Pretraining Approachcited by 13,372
  4. 2019BLiMP: The Benchmark of Linguistic Minimal Pairs for Englishcited by 564
  5. 2016Assessing the Ability of LSTMs to Learn Syntax-Sensitive Dependenciescited by 354
  6. 2019What BERT is not: Lessons from a new suite of psycholinguistic diagnostics for language modelscited by 320
  7. 2020COGS: A Compositional Generalization Challenge Based on Semantic Interpretationcited by 263
  8. 2018Targeted Syntactic Evaluation of Language Modelscited by 242
  9. 2020A Systematic Assessment of Syntactic Generalization in Neural Language Modelscited by 165
  10. 2019Neural Language Models as Psycholinguistic Subjects: Representations of Syntactic Statecited by 138
  11. 2018What do RNN Language Models Learn about Filler-Gap Dependencies?cited by 86
  12. 2023Language models are not naysayers: An analysis of language models on negation benchmarkscited by 74
15 more
  1. 2021Frequency Effects on Syntactic Rule Learning in Transformerscited by 59
  2. 2019Investigating BERT's Knowledge of Language: Five Analysis Methods with NPIscited by 58
  3. 2022Event knowledge in large language models: the gap between the impossible and the unlikelycited by 45
  4. 2018Do Language Models Understand Anything? On the Ability of LSTMs to Understand Negative Polarity Itemscited by 36
  5. 2021Language Models Use Monotonicity to Assess NPI Licensingcited by 31
  6. 2022Syntactic Surprisal From Neural Models Predicts, But Underestimates, Human Processing Difficulty From Syntactic Ambiguitiescited by 30
  7. 2023A Better Way to Do Masked Language Model Scoringcited by 27
  8. 2022RuCoLA: Russian Corpus of Linguistic Acceptabilitycited by 26
  9. 2020Contextualized Word Embeddings Encode Aspects of Human-Like Word Sense Knowledgecited by 18
  10. 2023ScoNe: Benchmarking Negation Reasoning in Language Models With Fine-Tuning and In-Context Learningcited by 18
  11. 2020Discourse structure interacts with reference but not syntax in neural language modelscited by 17
  12. 2022SLING: Sino Linguistic Evaluation of Large Language Modelscited by 15
  13. 2023Construction Grammar Provides Unique Insight into Neural Language Modelscited by 13
  14. 2020Recurrent babbling: evaluating the acquisition of grammar from limited input datacited by 9
  15. 2022Representing Affect Information in Word Embeddingscited by 3

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 1; 2025: 5; 2026: 5.

11 papers.

  1. 2025How Linguistics Learned to Stop Worrying and Love the Language Modelscited by 51
  2. 2024Revenge of the Fallen? Recurrent Models Match Transformers at Predicting Human Language Comprehension Metricscited by 6
  3. 2025Language models as tools for investigating the distinction between possible and impossible natural languagescited by 0
  4. 2025Large language models have learned to use languagecited by 0
  5. 2025Large language models are not about natural languagecited by 0
  6. 2025Linguists should learn to love speech-based deep learning modelscited by 0
  7. 2026Are Language Models Models?cited by 0
  8. 2026What can LLMs tell us about the mechanisms behind polarity illusions in humans? Experiments across model scales and training stepscited by 0
  9. 2026Across the Levels of Analysis: Explaining Predictive Processing in Humans Requires More Than Machine-Estimated Probabilitiescited by 0
  10. 2026You Can't Fight in Here! This is BBS!cited by 0
  11. 2026Not-So-Strange Love: Language Models and Generative Linguistic Theories are More Compatible than They Appearcited by 0