Surveying the Dead Minds: Historical-Psychological Text Analysis with Contextualized Construct Representation (CCR) for Classical Chinese

TL;DR

The paper tackles the challenge of extracting psychological constructs from historical/classical Chinese texts by introducing Contextualized Construct Representation (CCR). It combines cross-lingual questionnaire conversion with indirect supervised contrastive learning to fine-tune contextual transformers on a newly built Chinese Historical Psychology Corpus (C-HI-PSY). Across semantic similarity, questionnaire item classification, and psychological measurement tasks, CCR outperforms traditional DDR and prompting GPT-4, with benchmarking against externally annotated data supporting validity. The work enables theory-driven, interpretable psychological analysis in historical texts and offers open datasets and methods to advance historical psychology and NLP research.

Abstract

In this work, we develop a pipeline for historical-psychological text analysis in classical Chinese. Humans have produced texts in various languages for thousands of years; however, most of the computational literature is focused on contemporary languages and corpora. The emerging field of historical psychology relies on computational techniques to extract aspects of psychology from historical corpora using new methods developed in natural language processing (NLP). The present pipeline, called Contextualized Construct Representations (CCR), combines expert knowledge in psychometrics (i.e., psychological surveys) with text representations generated via transformer-based language models to measure psychological constructs such as traditionalism, norm strength, and collectivism in classical Chinese corpora. Considering the scarcity of available data, we propose an indirect supervised contrastive learning approach and build the first Chinese historical psychology corpus (C-HI-PSY) to fine-tune pre-trained models. We evaluate the pipeline to demonstrate its superior performance compared with other approaches. The CCR method outperforms word-embedding-based approaches across all of our tasks and exceeds prompting with GPT-4 in most tasks. Finally, we benchmark the pipeline against objective, external data to further verify its validity.

Figures (11)

• Figure 1: Comparison of the best performance among the DDR, CCR, and prompting methods on three tasks in the C-HI-PSY test set. (STS: Semantic Textual Similarity, PM: Psychological Measure, QIC: Questionnaire Item Classification)
• Figure 2: Pipeline of cross-lingual questionnaire conversion and contextualized construct representation for classical Chinese.
• Figure 3: Pipeline of triplet sampling and contrastive learning. CLM stands for contextual language model.
• Figure 4: Performance variation with sampling methods and thresholds.
• Figure 5: Comparison of model performance using the CCR method on the three tasks in the C-HI-PSY test set before and after fine-tuning. (Model A: bert-ancient-chinese, B: guwenbert-base, C: guwenbert-large, D: paraphrase-multilingual-MiniLM-L12-v2, E: text2vec-base-chinese, F: text2vec-base-chinese-paraphrase, G: text2vec-large-chinese)