Research Outputs & Projects

This study exploits a special historical case-openings of treaty ports in 19th-century China to examine how upper-tail human capital, quantified via book creation, impacted modernization when facing external pressures. Employing a prefecture-level panel dataset from 1840 to 1904, the study establishes book density, indicative of knowledge endowment, as a significant and positive predictor of modern firm entry following the opening of treaty ports. To understand the mechanism, a critical aspect lies in understanding the Civil Service Examination (keju), an indigenous institution that historically dominated talent accumulation and allocation in China. By integrating data with keju, we find that exposure to Western influence mobilized the segment of upper-tail human capital at the bottom or outside of the keju system into entrepreneurship. This paper illustrates the dynamics between indigenous institutions and external pressures.

Christian missionaries are shown to have contributed to the historical rise of international trade by bridging the information gap between Europe and unfamiliar overseas lands from the 16th century onward. To prove that they unintentionally mitigated information barriers for early traders, we focus on the experience of historical China where European missionaries arrived from 1580 onward but maritime foreign trade was largely banned between 1371 and 1842. Our analysis demonstrates that following China’s forced opening for international trade in 1842, regions with longer past missionary presence typically bought foreign goods earlier, imported more in terms of both value and goods diversity, and exported more local products, as these places had more foreign interactions and appeared more frequently in the missionaries’ letters and publications back in Europe. Our findings substantiate the importance of the information channel through which the missionaries accelerated trade globalization.

Extracting geographic information from historical texts presents unique challenges. To address these challenges, this study leverages generative large language models (LLMs) to extract historical toponyms and their corresponding location references from texts. The coordinates of the extracted toponyms are then identified by a historical geocoder, which also calculates their maximum error distances based on the location references, indicating the degree of uncertainty. Both the extraction and geocoding processes are integrated into a novel tool named ‘His-Geo’ (https://github.com/yukiyuqichen/His-Geo). To evaluate the results, this study also curates a manually annotated dataset, the Early China Historical Geographic Corpus (CHGC-Early), filling the gap in the absence of geographic data for early China in existing gazetteers and providing a benchmark dataset for training and evaluating approaches for tasks related to geographic information extraction from premodern Chinese texts. The evaluation results show a satisfactory 0.831 F1 score for the GPT-4o model, demonstrating the remarkable capability of generative large language models in extracting geographic information from lengthy, unstructured texts that encompass diverse and sometimes conflicting views.

量化歷史和數字人文是近年來快速發展的交叉學科。 在大數據、人工智能時代，兩者都促進了人文學科的發展。 同時，量化歷史研究跟數字人文兩者之間也是相互促進、互為補充的。 與數位人文相比，量化歷史的重點集中在資料庫的利用，特別是結合量化分析方法、社會科學理論來研究重要的歷史現象，發現其中的規律、驗證因果關係並提供系統解釋。 也就是說，數字人文一般是基於數據對現象之間的相關性做出估算、判斷，而量化歷史研究則更進一步探究歷史現象之間的因果關係並從統計學上做系統驗證。 隨著更多歷史資料庫的建立並為學界開放，量化分析方法的應用也隨之增加，變得更為廣泛。

China’s center of socioeconomic activities was in the North prior to the Tang dynasty but is in the South today. We demonstrate that Arab and Persian Muslim traders triggered that transition when they came to China in the late seventh century, by lifting maritime trade along the South Coast and re-creating the South. Between 742 and 1393 CE, prefectures with better access to maritime trade, or higher porcelain trade participation, experienced significantly higher population growth, but the predictive coefficient weakened substantially after the maritime trade ban of 1371 CE. These findings are robust after controlling for many confounding factors.

當前，對人工智能造成的衝擊，歷史學者的態度並不一致，或熱烈擁抱，或批判吸收，或警惕反思，總體而言，是在被動地回應。 但是，僅僅被動回應是不夠的。 歷史學及其理論不僅應當重視人工智能，更應該介入和影響其發展。 早在1976年，約瑟夫·魏岑鮑姆（Joseph Weizenbaum）就提出，歷史學有必要介入人工智能，特別是保存那些無法成為人工智能輸入數據的史料。 2019年，在斯坦福大學成立了「以人為本的人工智能」研究所，該研究所認識到，人工智能在改善人類狀況方面有著非凡前景，但前提是能夠成功引導其朝著負責任的方向發展。 為達此目的，研究所吸納了不少歷史學者和其他人文學者。 這一機構之所以能贏得人工智能領域的世界級聲譽，正是其理念和舉措的結果。 近年來，也有中國學者注意到這個問題，主張將知識、技能、感性經驗、價值觀念融入數位人文研究方法，其中包括人工智能的方法，惜未充分展開。
歷史學及其理論之所以能夠介入和影響人工智能，特別是當前代表人工智能前沿方向的大語言模型（以下簡稱「大模型」）的發展，很重要的一個因素在於兩者之間工作原理的相似和相通性。 在一定意義上說，歷史學就是用語言將「數據、記憶、關於過去的證據性遺迹、文獻和遺物」變成歷史的自覺省思。 大模型實際上是對人類社會積累的語言材料的吸收和綜合，或者說，是以兼收並蓄、有時可能是雜亂無章的方式吸收綜合其可以獲得的所有文本。 由此可見，大模型的工作方式與歷史學者的研究存在很大的相似性，即基於過去的各種史料，進行加工和提煉，從而獲得比較系統化的知識。 正如約書華·斯特恩菲爾德（Joshua Sternfeld）所說，收集模型的訓練數據就像是收集史料，模型根據輸入產生輸出就像是根據證據產生歷史，總而言之，「我們越是深入人工智能（大模型）學習的過程，事實上，這個過程就開始和做歷史研究的工作越發相似」。
因此，歷史學及其理論與人工智能（大模型）之間存在深度交融的可能和必要，但具體如何實現，既需要認識前者，也需要深入瞭解後者的運作機理。

Survival cannibalism persisted across human societies until recently. What drove the decline in cannibalism and other forms of violence? Using data from the 1470–1910 period, this paper documents that in historical China, the Confucian clan—an institutionalized kinship network—acted as an informal internal market to facilitate intra-clan resource pooling and risk-sharing, thus reducing the need for cannibalism during times of drought-related famine. The risk mitigation role of the clan remains robust after controlling for economic development and other factors and ruling out alternative channels. Thus, kinship networks and their associated culture contributed to human civilizational development before the advent of formal markets.

已有對近代鄉村不平等的研究主要關注地權分配，好少關注收入、消費分配，特別是同時討論三者之間嘅關係。 利用滿鐵調查32個村莊嘅數據，發現從戶均角度睇，土地不平等的變動對收入不平等和消費不平等的變動有一定的解釋力，收入不平等的變動對消費不平等的變動則有較大的解釋力。 但從人均角度睇，土地不平等對收入不平等和消費不平等嘅解釋力以及收入不平等對消費不平等嘅解釋力都好有限。 另外，也發現土地租佃市場和借貸市場的發展，有助於減少收入和消費嘅不平等。

社會網路分析是目前被廣泛應用的研究方法，特別是在一些交叉學科之中。 本文以量化歷史研究中的社會網路分析為例，基於政治、軍事、思想文化、經濟等領域的案例，介紹其原理與應用，說明其在説明理解社會結構中的人，建立微觀個體與宏觀社會現象之間的聯繫，揭示社會複雜特徵等方面的價值。