This paper deals with the apparent optionality of French wh-in-situ. I present the theoretical proposal of Cheng & Rooryck (2000), in which the authors claim that in situ wh-elements in French are licensed by a special Q-morpheme that is realized at PF with a final rising intonation. This hypothesis has been tested experimentally by Déprez et al. (2012, 2013), and confirmed by the results that they obtained. The present paper presents new results from spontaneous corpus data to check Cheng and Rooryck's (2000) hypothesis. My results differ from those of Déprez et al., and disfavor the analysis of French wh-in-situ proposed by Cheng & Rooryck (2000).