By means of a phraseological identification method, this study provides a general description of the use of conventional sequences (CSs) in interviews at four different levels of spoken L2 French as well as in interviews with native speakers. Use of conventional sequences is studied with regard to overall quantity, category distribution and type frequencies. The most predictive measure is overall quantity, which yields significant differences between several learner levels. It is also found that Lexical CSs are the most difficult to acquire for second language speakers: only the most advanced group use them to the same extent as native speakers. No significant differences are found between the most advanced group of L2 speakers (LOR in France greater than 5 yrs) and native speakers, probably due to the measures and the task investigated. The results are then related to Ellis et al. (TESOL Quarterly 42: 375-396, 2008), suggesting that the sequences' frequencies of occurrence vs. their MI score in a larger corpus might influence their acquisition and use.