The focus of this chapter is on how machine learning (ML) impacts the analysis of culture insociology. It shows how ML has greatly advanced the analysis of culture, with new tools enabling amassive and fine-grained extraction of information from textual and audiovisual troves as well as dataanalysis, operationalizing long-standing cultural sociology concepts. It also indicates that this renewedinterest is building on already fertile ground, as sociologists of culture have long used and reflected onformal models when analyzing culture. The chapter suggests that as the toolbox of ML approachesexpands, so will the need for methodological reflection on the datasets and algorithms used, analyzed,and interpreted. The chapter also suggests that ML techniques can serve as catalysts to generate newtheoretical insights. The chapter’s conclusion discusses the potential of ML research to generate newtheoretical insights abductively and advocates for methodological reflexivity.

 The last decade witnessed a spectacular rise in the volume of available textual data. With this new abundance came the question of how to analyze it. In the social sciences, scholars mostly resorted to two well-established approaches, human annotation on sampled data on the one hand (either performed by the researcher, or outsourced to microworkers), and quantitative methods on the other. Each approach has its own merits – a potentially very fine-grained analysis for the former, a very scalable one for the latter – but the combination of these two properties has not yielded highly accurate results so far. Leveraging recent advances in sequential transfer learning, we demonstrate via an experiment that an expert can train a precise, efficient automatic classifier in a very limited amount of time. We also show that, under certain conditions, expert-trained models produce better annotations than humans themselves. We demonstrate these points using a classic research question in the sociology of journalism, the rise of a “horse race” coverage of politics. We conclude that recent advances in transfer learning help us augment ourselves when analyzing unstructured data.

"To anyone familiar with the French political scene, the 2017 election season felt atypical. Normally, during these moments of intense attention on politics, the main issues often revolve around broader issues. Security, taxation, immigration, or more recently the environment, are more likely to capture the public's attention than the careers of political representatives. But not in 2017. That year the attention was focused on the biography of the political personnel, and the term "les professionnel de la politique" [professional politicians] became a ubiquitous insult in public debate. But neither in France nor abroad was this attention to the political elite fully unprecedented. Rather, it marked the latest revival of an insult that is as old as the compensation of politicians for their work. The criticism of professional politicians had been widely revived over the previous decade. An apparently trivial example provides a good illustration of this. In blog post from September 2014, Michèle Delaunay, a former Socialist MP, lamented the arrival of a new generation into politics. According to her, those moving into national politics in the 2010s shared one trait: the vast majority of them had had no professional experience outside this milieu. Although she did not use the term "professional" or "career politicians", as such, her target was clear. "They graduate from Sciences Po [the school of the political elite], take an administrative exam, or not, they look around ... Then they get a position as a staffer or a local government job. The luckiest, or cleverest of them end up as a top aide to a cabinet member. In this ever so slightly limited world, they catch the bug." Describing the stages in a well-oiled career, Delaunay, a former oncologist, observed at the end of her career that her youngest colleagues had less and less experience outside politics. The image she used is telling. She wrote that "they go into the tunnel early and never come out." According to her, the consequences are immense. Once they set out on this path, these ambitious young people "lose touch with reality and the sense of the common good." As they get older, they behave as if they are "beyond even the most basic rules""--Provided by publisher.

The release of ChatGPT at the end of 2022 thrust large language models (LLMs) into the limelight. By enabling its users to query the model directly in natural language, ChatGPT democratized access to these models — a welcome development. Since then, ChatGPT and similar tools such as Bard, Claude and Bing AI have shown their versatility and efficiency on a wide variety of tasks.

Combien perçoivent vraiment les députés français pour leur activité parlementaire ? À cette question en apparence anodine, il est pourtant difficile d’apporter une réponse claire pour la majeure partie du xxe siècle. À partir d’un travail inédit mené dans diverses archives de l’Assemblée nationale, cet article propose une première estimation du revenu que les députés ont pu retirer de leur mandat, du début du xxe siècle à nos jours. Il démontre ainsi la fécondité de ces données pour l’historien, en ce qu’il contribue à répondre à trois interrogations liées : pourquoi est-il si complexe d’accéder à une information normalement publique ? Comment le revenu des parlementaires situe-t-il les élus dans l’échelle des revenus de la population française ? Enfin, que nous disent ces évolutions de l’indemnité (en termes de niveau comme de nature) sur le type d’activité qu’est la députation ? Ce faisant, l’étude éclaire certaines transformations de fond du métier politique.

How much are members of the French parliament paid? No clear answer can be provided to this apparently trivial question for most of the twentieth century. Based on new research carried out in the archives of the lower chamber, the Assemblée nationale, this article offers an estimate of the revenues politicians have collected from their parliamentary activity since the early twentieth century. This rich seam of historical data is used to address three interrelated questions: Why is it so complex to access supposedly public information? Where do French parliamentarians’ earnings place them on the scale of the active population? Finally, what do the evolutions identified (in terms of both value and kind) reveal about the professional activity of these representatives? In probing these questions, the article sheds new light on political careers in France.