The Illusion of Causality in LLMs: A Developmentally Grounded Analysis of Semantic Scaffolding and Benchmark–Capability Mismatches

Recent benchmarks increasingly report that large language models (LLMs) exhibit human-like causal reasoning abilities, including counterfactual inference and intervention planning. However, many such evaluations rely on domains that are heavily represented in training data and embed strong semantic cues, raising the possibility that apparent causal competence may reflect semantic pattern recombination rather than structure-sensitive causal reasoning. Drawing on human developmental theories of causal induction, this perspective argues that genuine causal understanding requires robustness to novelty and reliance on conditional structure rather than semantic familiarity. To illustrate the testability of this claim, the paper includes a pilot demonstration using synthetic causal micro-worlds. Identical numerical evidence was presented to a LLM under two conditions: semantically meaningful variable labels and non-semantic coded labels. Across paired cases, the model reliably selected the correct causal structure when labels were meaningful, but frequently misidentified or exhibited instability in causal model selection under coded labels, despite producing locally coherent explanations. These divergences emerged most clearly in diagnostically ambiguous settings requiring suppression of misleading marginal associations. The results align with the claim that semantic scaffolding can support and stabilize apparent causal competence in LLMs without implying structure-sensitive reasoning.

• Publication year

  2026

• Venue

  Machine Learning and Knowledge Extraction

• Publication date

  2026-03-02

• Fields of study

  Not labeled

• Identifiers
  DOI 10.3390/make8030057
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• The development of human causal learning and reasoning

  2024cited by this paper
• Dissociating language and thought in large language models: a cognitive perspective

  2023cited by this paper
• Using cognitive psychology to understand GPT-3

  2022cited by this paper
• Toward Causal Representation Learning

  2021cited by this paper
• Climbing towards NLU: On Meaning, Form, and Understanding in the Age of Data

  2020cited by this paper
• Building Machines that Learn and Think Like People

  2018cited by this paper
• Developmental psychology : Bandura , Ross and Ross ( 1961 ) Transmission of aggression through imitation of aggressive models

  2015cited by this paper
• Observing the unexpected enhances infants’ learning and exploration

  2015cited by this paper
• Reconstructing constructivism: causal models, Bayesian learning mechanisms, and the theory theory.

  2012cited by this paper
• Where science starts: spontaneous experiments in preschoolers' exploratory play.

  2011cited by this paper
• The Blicket Within: Preschoolers' Inferences About Insides and Causes

  2007cited by this paper
• Serious fun: preschoolers engage in more exploratory play when evidence is confounded.

  2007cited by this paper
• Children's causal inferences from indirect evidence: Backwards blocking and Bayesian reasoning in preschoolers

  2004cited by this paper
• Looking and search measures of object knowledge in preschool children.

  2003cited by this paper
• Detecting blickets: how young children use information about novel causal powers in categorization and induction.

  2000cited by this paper
• Dissociations in infant memory: rethinking the development of implicit and explicit memory.

  1997cited by this paper

• No citing papers are available for this paper.