Article: Counterfactual causal inference for attribution of L2 Chinese grammatical errors Journal: International Journal of Information and Communication Technology (IJICT) 2026 Vol.27 No.49 pp.79 - 97 Abstract: Accurately attributing Chinese second language grammatical errors is crucial for optimising teaching strategies. However, traditional methods are prone to being disturbed by confounding factors such as the learner's level, making it difficult to distinguish between superficial correlation and true causation. To address this, this paper introduces the framework of counterfactual causal inference for the first time. By simulating 'correction' interventions on specific grammatical points, it aims to identify the root causes of the errors. Experiments based on a large-scale public Chinese proficiency test dynamic composition corpus show that this method achieves an accuracy rate of 87.5% in error attribution, an improvement of 8.2% over the best baseline model; its causal effect ranking quality reaches 0.92, significantly outperforming traditional correlation analysis. This method provides interpretable and verifiable causal insights for Chinese second language teaching, and can directly serve the construction of personalised learning paths. Inderscience Publishers - linking academia, business and industry through research

Title: Counterfactual causal inference for attribution of L2 Chinese grammatical errors

Authors: Hongbo Ma

Addresses: School of Chinese Language and Literature, Anyang Normal University, Anyang, 455002, China; School of Chinese Language and Literature, Capital Normal University, Beijing, 100048, China

Abstract: Accurately attributing Chinese second language grammatical errors is crucial for optimising teaching strategies. However, traditional methods are prone to being disturbed by confounding factors such as the learner's level, making it difficult to distinguish between superficial correlation and true causation. To address this, this paper introduces the framework of counterfactual causal inference for the first time. By simulating 'correction' interventions on specific grammatical points, it aims to identify the root causes of the errors. Experiments based on a large-scale public Chinese proficiency test dynamic composition corpus show that this method achieves an accuracy rate of 87.5% in error attribution, an improvement of 8.2% over the best baseline model; its causal effect ranking quality reaches 0.92, significantly outperforming traditional correlation analysis. This method provides interpretable and verifiable causal insights for Chinese second language teaching, and can directly serve the construction of personalised learning paths.

Keywords: second language acquisition; attribution of grammatical errors; dual machine learning; DML.

DOI: 10.1504/IJICT.2026.153618

International Journal of Information and Communication Technology, 2026 Vol.27 No.49, pp.79 - 97

Received: 17 Jan 2026
Accepted: 18 Feb 2026
Published online: 18 May 2026 *

Title: Counterfactual causal inference for attribution of L2 Chinese grammatical errors

Keep up-to-date