This study addresses the impact of a definition paradigm shift on the use of anglicisms in the French-speaking written press. Built around a diachronic analysis (2000-2015) of the numerous differences linked to a change of definition, it also ventures into the impact such a change may have on comparative analysis focusing on geographical origin (France vs Quebec) or the nature of the newspaper (“of record” vs “popular”). It aims at contributing to the literature though two different aspects:(1) This study is the only conceptual work to highlight what sort of impact a change in the definition of “English borrowing” can have on the frequency of anglicism usage. It shows through an opposition with/without French equivalents, that an analysis focused only on anglicisms for which French provides an equivalent – those Forest and Boudreau (1998) consider “unnecessary” or “erroneous” (Villers, 2009) – reveals an important drop in the usage frequency compared to an analysis based on all anglicisms. It also unveils that, out of three anglicisms extracted from my corpus, two have no equivalents in French and are de facto unavoidable.(2) This research is, to my knowledge, one of the first in comparative study to show, with empirical evidence, how definition affects not only usage frequency, but also dictates the way newspapers develop over time depending on different criterion. Based on two different angles of comparison (geographical origin and nature of the journal), I also show that a change in the definition leads to different evolution trajectories, different relationships between our newspapers and different diachronic variations. To answer my research questions, I built one of the largest corpora ever used for a study on anglicisms. With more than 330 million words, it gathers all the articles published by four different newspapers for the years 2000, 2005, 2010 and 2015: Le Monde and Le Parisien for France, and Le Devoir and La Presse for Quebec. I also used two lists of keywords composed of 5416 and 2934 anglicisms taken from two general dictionaries (Le Petit Robert 2016 and Le Multidictionnaire de la langue française–4th edition 2013) and two specialized dictionaries (Höfler’s Dictionnaire des anglicismes–1982 and Le Colpron, dictionnaire des anglicismes–4th edition). By going far beyond what previous studies have done, it offers a more advanced overview than usual.First, my results show that the usage frequency is generally low (0.72% for all anglicisms; 0.28% for anglicisms with equivalents only; 0.44% for anglicisms without equivalents only) both in France and Quebec, which is consistent with the findings of previous studies. However, I find clear differences in terms of results for each analysis, as the frequency rate for anglicisms with equivalents is 1.5 times lower than the frequency rate for anglicisms without equivalent. This proves that the definition of the very concept of “anglicism” directly influences the frequency of usage. Second, I observe that this change of definition also affects comparative analysis focused on specific criteria such as the geographical origin or the nature of the newspapers. My study thus reveals great nuances in terms of trend over 15 years according to which definition is used: for instance, results for Quebec newspapers are slowly decreasing over time when anglicisms with equivalents are concerned, – especially for Le Devoir, whereas they increase when anglicisms do not have an equivalent. When it comes to the nature of the newspaper, Le Parisien, which is the newspaper that uses the most anglicisms of the four, forms a much more heterogeneous duo with La Presse when only anglicisms with equivalents are taken into account. They also give prominence to Le Devoir and Le Monde whose evolution trends are close to identical when the broader definition of Anglicism is used but a bit more asymmetric when equivalents come into play.