Do Language Models Know What Not to Say? Causal Evidence for Statistical Preemption in LLMs

• URL: http://arxiv.org/abs/2605.23039v1
• Date: Thu, 21 May 2026 21:06:43 GMT
• Title: Do Language Models Know What Not to Say? Causal Evidence for Statistical Preemption in LLMs
• Abstract summary: Construction Grammar proposes statistical preemption: exposure to a conventional form preempts structurally possible but unattested alternatives.<n>We present a computational study that dissociates statistical preemption from the competing entrenchment hypothesis in large language models.
• Score: 13.891522069967507
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: How do learners acquire knowledge of what is unacceptable without negative evidence? Construction Grammar proposes statistical preemption: exposure to a conventional form (e.g., "donated the books to the library") preempts structurally possible but unattested alternatives ("*donated the library the books"). We present a computational study that, for the first time, directly dissociates statistical preemption from the competing entrenchment hypothesis in large language models within a single converging design. Across four experiments spanning 120 English verb-construction pairings (dative, causative, locative), we show that (1) LLM surprisal patterns correlate strongly with human acceptability judgments ($r = 0.79$), validated against three independent behavioral datasets; (2) these patterns are driven by competing-form frequency rather than overall verb frequency, confirmed by non-circular partial correlations; (3) preemption sensitivity scales as a power law with model size; and (4) a controlled fine-tuning intervention causally demonstrates that manipulating competing-form frequencies shifts preemption behavior in the predicted direction, with reverse-direction controls ruling out frequency-sensitivity confounds. These results provide converging evidence that neural language models acquire negative linguistic knowledge through distributional competition, the core mechanism posited by Construction Grammar.

Related papers

• DERELAB: Probing Defeasible Reasoning and Confirmation Bias in LLMs with a Generative Benchmark [6.835919520017402]
  Defeasible reasoning is a type of reasoning where inferences are drawn from plausible current evidence, but can be retracted upon the introduction of newer evidence.<n>We introduce DeReLab, a generative framework that produces multi-turn belief-updating conversations from parameterized graph structures.<n>We evaluate nine open and proprietary large language models and find that nearly all exhibit a systematic tendency to accept congruent evidence while resisting incongruent updates.
  arXiv  Detail & Related papers  (2026-08-31T08:10:29Z)
• How Unlikely Is "Unlikely"? Assessing Verbal Probability Perception Across Large Language Models [1.435579512127234]
  We present a systematic cross-model evaluation using a word-to-number mapping task grounded in established human benchmarks.<n>Eleven uncertainty expressions were presented to 19 models under two conditions, forced single-number response and explanation elicitation.<n>Models show a systematic upward bias for negative expressions such as unlikely'' and improbable''
  arXiv  Detail & Related papers  (2026-08-26T19:00:41Z)
• Causally Evaluating the Learnability of Formal Language Tasks [59.73001113068827]
  We rigorously investigate the relationship between data frequency and learnability.<n>We derive Kullback-Leibler divergence metrics to measure the learnability of specific sub-tasks.<n>Our experiments show that evaluating learnability without causal intervention leads to incorrect conclusions.
  arXiv  Detail & Related papers  (2026-06-08T17:58:36Z)
• Rethinking the Idiomaticity Decomposability Hypothesis: Evidence from Distributional Learning [15.730654942954267]
  We propose a model-internal measure of decomposability and relate it to human ratings, syntactic flexibility, and predictability while tracking learning during pretraining.<n>We show that stabilisation of representations in models is not explained by frequency alone. Instead surprisal decomposability, and frequency all contribute with decomposability showing the strongest training-dependent effect.
  arXiv  Detail & Related papers  (2026-06-02T15:59:22Z)
• Evaluating Evidence Grounding Under User Pressure in Instruction-Tuned Language Models [6.431677598656395]
  In contested domains, instruction-tuned language models must balance user-alignment pressures against faithfulness to the in-context evidence.<n>We conduct fine-grained ablations over evidence composition and uncertainty cues across 19 instruction-tuned models spanning 0.27B to 32B parameters.
  arXiv  Detail & Related papers  (2026-03-20T17:38:23Z)
• LLM Reasoning Predicts When Models Are Right: Evidence from Coding Classroom Discourse [0.18268488712787334]
  Large Language Models (LLMs) are increasingly deployed to automatically label and analyze educational dialogue at scale.<n>We investigate whether reasoning generated by LLMs can be used to predict the correctness of a model's own predictions.<n>We analyze 30,300 teacher utterances from classroom dialogue, each labeled by multiple state-of-the-art LLMs with an instructional move construct and an accompanying reasoning.
  arXiv  Detail & Related papers  (2026-02-10T14:38:13Z)
• Planted in Pretraining, Swayed by Finetuning: A Case Study on the Origins of Cognitive Biases in LLMs [51.00909549291524]
  Large language models (LLMs) exhibit cognitive biases.<n>These biases vary across models and can be amplified by instruction tuning.<n>It remains unclear if these differences in biases stem from pretraining, finetuning, or even random noise.
  arXiv  Detail & Related papers  (2025-07-09T18:01:14Z)
• A Closer Look at Bias and Chain-of-Thought Faithfulness of Large (Vision) Language Models [58.32070787537946]
  Chain-of-thought (CoT) reasoning enhances performance of large language models.<n>We present the first comprehensive study of CoT faithfulness in large vision-language models.
  arXiv  Detail & Related papers  (2025-05-29T18:55:05Z)
• ExpliCa: Evaluating Explicit Causal Reasoning in Large Language Models [75.05436691700572]
  We introduce ExpliCa, a new dataset for evaluating Large Language Models (LLMs) in explicit causal reasoning.<n>We tested seven commercial and open-source LLMs on ExpliCa through prompting and perplexity-based metrics.<n>Surprisingly, models tend to confound temporal relations with causal ones, and their performance is also strongly influenced by the linguistic order of the events.
  arXiv  Detail & Related papers  (2025-02-21T14:23:14Z)
• Do Large Language Models Show Biases in Causal Learning? [3.0264418764647605]
  Causal learning is the cognitive process of developing the capability of making causal inferences based on available information.<n>This research investigates whether large language models (LLMs) develop causal illusions.
  arXiv  Detail & Related papers  (2024-12-13T19:03:48Z)
• The Devil is in the Neurons: Interpreting and Mitigating Social Biases in Pre-trained Language Models [78.69526166193236]
  Pre-trained Language models (PLMs) have been acknowledged to contain harmful information, such as social biases. We propose sc Social Bias Neurons to accurately pinpoint units (i.e., neurons) in a language model that can be attributed to undesirable behavior, such as social bias. As measured by prior metrics from StereoSet, our model achieves a higher degree of fairness while maintaining language modeling ability with low cost.
  arXiv  Detail & Related papers  (2024-06-14T15:41:06Z)
• Avoiding Inference Heuristics in Few-shot Prompt-based Finetuning [57.4036085386653]
  We show that prompt-based models for sentence pair classification tasks still suffer from a common pitfall of adopting inferences based on lexical overlap. We then show that adding a regularization that preserves pretraining weights is effective in mitigating this destructive tendency of few-shot finetuning.
  arXiv  Detail & Related papers  (2021-09-09T10:10:29Z)
• Exploring Lexical Irregularities in Hypothesis-Only Models of Natural Language Inference [5.283529004179579]
  Natural Language Inference (NLI) or Recognizing Textual Entailment (RTE) is the task of predicting the entailment relation between a pair of sentences. Models that understand entailment should encode both, the premise and the hypothesis. Experiments by Poliak et al. revealed a strong preference of these models towards patterns observed only in the hypothesis.
  arXiv  Detail & Related papers  (2021-01-19T01:08:06Z)
• HypoNLI: Exploring the Artificial Patterns of Hypothesis-only Bias in Natural Language Inference [38.14399396661415]
  We derive adversarial examples in terms of the hypothesis-only bias. We investigate two debiasing approaches which exploit the artificial pattern modeling to mitigate such hypothesis-only bias.
  arXiv  Detail & Related papers  (2020-03-05T16:46:35Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.