Pronominal presence of the researcher in written scientific discourse is a controversial subject. The analysis of statistical data confirms the hierarchy of frequency between je, nous and on, but reveals considerable variations in one and the same functional category of one and the same author and thus suggests a survey on the interiorised norm. Confirming the persistence of the traditional taboo of I, this study shows that the majority of the interrogated persons prefers the nousexclusif and the hortatif to the pronoun je, but rejects the pronoun on ‘je’ as well as the indefinite on in passive constructions, which is also not considered the genus verbi scientific par excellence.