This thesis presents a comparative study of French language commentary from two time periods, the late nineteenth and the early twenty-first centuries, and from two perspectives, expert and lay. It analyses four sources of language commentary to undertake two main comparisons. The first is a diachronic comparison of two language advice services: Le Courrier de Vaugelas (1868-1881) and the Courrier des internautes (2011-present, part of the Dire, Ne pas dire section of the Académie française's website). Both sources publish readers' questions about language and a response from an expert, allowing for the analysis of commentary from an expert and lay perspective. Expert language commentary has been well studied, in the form of usage guides, remarques, chroniques de langage and dictionary prefaces, for instance, but these sources are primarily monologic. Analysis of the two dialogic Q+A sources provides insight into both lay and expert commentary and the interaction between them. The second comparison, a synchronic comparison, analyses the blog posts and user comments from two websites on the topic of language and correctness: Langue sauce piquante (2004-present) and Bescherelle ta mère (2014-present). Langue sauce piquante contains both expert and lay commentary. Bescherelle ta mère, on the other hand, is an exclusively lay space and its audience comprises not language enthusiasts (as is the case with the other three sources) but 'ordinary' people, due to both the type of content featured and its accessibility via Facebook. Language commentary from the nineteenth and the twenty-first century has received less scholarly attention than, for instance, the seventeenth century which marked the beginning of the remarqueur tradition, and the twentieth century, the period in which language columns were at their most popular. However, both the time periods analysed here are times of significant change for the language. In the late nineteenth century, the introduction of free compulsory education in the French language began to increase the number of people interested in questions of language. Turning to the twenty-first century, the lay-lay language commentary which we might assume was occurring most frequently in spoken language and was therefore inaccessible to researchers, has become accessible online. This thesis exploits the opportunities presented by the internet to examine lesser-studied lay-lay language commentary. This thesis combines quantitative and qualitative analysis to examine the areas of the language which, to judge by the four sources studied here, cause French speakers difficulty or simply interest, how linguistic authority is created and negotiated, the recurring tropes in discussions of correct language, the use of purist and prescriptivist imagery, and, finally, the implicit and explicit language ideologies expressed in lay and expert language commentary. It shows that standard language ideology and prescriptivism run through the nineteenth- and twenty-first-century sources, and argues that they have become a part of popular culture in lay online spaces. Whilst there are some striking similarities across the forms of language commentary from two time periods and two mediums (print to online), analysis also suggests that, in some cases, traditional commentary has taken a more extreme form online.