This article is concerned with the investigation of variation, completed change, and change in progress, which are leading to a radical difference between everyday spoken European French and the standard written variety. We will focus on one small part of French grammar -- the subject pronouns -- and detail how grammatical and semantic change is contributing to an incipient diglossic situation in French, a situation which must be recognized by linguists and language teachers alike. Standard treatments (linguistic, reference, pedagogical grammars) of French agree that the subject personal pronouns/clitics (used with a verb) are: je ‘I’, tu ‘you’ (familiar/ singular), il, elle ‘he, she, it’, nous ‘we’, vous ‘you’ (plural, polite), ils, elles ‘they’, and, among others, the indefinite on ‘one’. This accounts quite well for written French, but it is quite inaccurate for everyday spoken French. Through the analysis of a corpus of everyday spoken European French, we have found that the subject clitics (especially the first and second person, and third person to a certain extent) have become grammatical prefixes. In addition, as we will show in detail, nous has all but disappeared as a subject clitic and has been replaced by on-. At the same time, the use of on- for ‘one’ is much less frequent than before: there has been a reversal of the basic/marginal relation in its meaning, such that the meaning ‘we’ occurs in by far the majority of its uses, and the meaning ‘one’ is now only a marginal meaning. There are, however, vague uses of on, which could be interpreted as either ‘we’ or ‘one’ -- thus showing the path of change from the one to the other both diachronically and synchronically. This indefinite meaning is now shifting over increasingly to tu (and only to a very small extent to vous), so much so that in our corpus, tu seems to have two basic meanings, split almost 50-50 between ‘you’ and ‘one’. This is inherently an unstable situation and probably presages more changes to come. It is clear, therefore, that more good corpus work is needed for a fuller understanding of spoken European French. Paradoxically, in addition, good corpus work is also needed on written French of many different varieties, since the reference and pedagogical grammars that focus on the written language tend to be based on the written French of only the ‘best’ authors and ‘good usage’. And finally, corpus-based reference works and textbooks are essential if we expect our students to develop any real, pragmatically appropriate, communicative proficiency in French.