Digital Humanities in the TIME-US Project: Richness and Contribution of Interdisciplinary Methods for Labour History

TL;DR

This paper discusses the integration of interdisciplinary methods and advanced digital tools, including AI, in labour history research within the TIME-US project, highlighting new approaches to studying women's work in history.

Contribution

It presents the innovative application of digital humanities and AI techniques to labour history, emphasizing interdisciplinary methods for analyzing historical gendered labor data.

Findings

Demonstrates the use of AI to analyze historical labor data.

Highlights the importance of interdisciplinary approaches in digital humanities.

Shows how digital methods can uncover previously invisible aspects of women's labor history.

Abstract

In 2015, the Annales journal, traditionally open to interdisciplinary approaches in history, referred to 'the current historiographical moment [as] call [ing] for an experimentation of approaches'. 1 Although this observation did not exclusively refer to the new possibilities offered by the technological advancements of the time -particularly in the field of artificial intelligence 2 -it was nonetheless motivated by these rapid and numerous changes, which also affect the historiographical landscape. A year earlier, St\'ephane Lamass\'e and Philippe Rygiel spoke of the 'new frontiers of the historian', frontiers opened a few years earlier by the realisation of the unprecedented impact of new technologies on historical practices, leading to a 'mutation des conditions de production et de diffusion des connaissances historiques, voire de la nature de celles-ci' ('transformation of the conditions of production and dissemination of historical knowledge, and even the nature of this knowledge'). 3 It was in this fertile ground, conducive to the cross-fertilisation of approaches, that the TIME-US project was born in 2016. TIME-US is directly the result of this awareness and reflects the transformations induced by major technological advancements, disrupting not only our daily practices but also our historical practices. 1 Annales 2015, 216. 2 For example, convolutional neural networks, which have revolutionised the field of artificial intelligence, began to gain popularity just before the 2010s. 3 Translated by the author. Lamass\'e and Rygiel 2014. To quantify women's work in the past, labour historians cannot rely on the classic sources of their discipline, which allow to produce large statistical data series, systematically treatable in the form of databases. What to do when such data are not available? Should the task simply be abandoned? As Maria {\AA}gren points out, the invisibility of women's participation in the labour market does not mean non-existence 8 ; there must therefore be traces of it. To quantify women's economic activity, Sara Horrell and Jane Humphries, for example, turned to household budgets from 59 different sources (from Parliamentary Papers to autobiographical texts), which had never before been systematically used to identify women's work patterns and their contribution to family income. 9 In her study A Bitter Living: Women, Markets, and Social Capital in Early Modern Germany published in 2003, Sheilagh Ogilvie used information contained in court records to identify activities carried out by women and the time spent on these activities. Court records were not intended to record such information; yet, in their testimonies, witnesses often described in detail the activities they were engaged in while a crime was unfolding before their eyes. Sheilagh Ogilvie thus identified nearly 3000 such observations. 10 These works have opened two main avenues for the TIME-US project. First, making already digitised sources accessible in homogeneous corpora. 11 Following the example of previous research, TIME-US mobilised varied sources containing traces of professional activities carried out by women in France during the period studied: these include both printed (posters and petitions, working-class newspapers, and contemporary surveys on workers) and handwritten sources (labour court decisions, police reports, company archives, personal archives, surveys, petitions). 12 One of the project's objectives was to gather and 8 {\AA}gren 2018a, 144. 9 Horrell and Humphries 1995.