The traumatic experience is characterized by “uns-peakability:” there are no words to truly express, to translate, torepresent the horror. And yet, paradoxically, the main access toour knowledge of traumatic dissociation comes from what patientsmanage to tell us, if only partially, chaotically. . . The lexical, syn-tactic, and pragmatic imprints of the trauma on discourse, whichwe have baptized psycholinguistic traumatic syndrome (SPLIT),are expressed according to an automatico-voluntary dissociationtaking gradual form: the more the trauma appears present, themore the psycholinguistic stigmata manifest themselves; the morethe subject psychically moves away from the trauma and its conse-quences, the more the discourse returns to the nominal state.How can the subject become master of his own life, of his ownspeech, without being constrained by trauma? How do some sub-jects manage, alone or with the help of others, to extract themselvesfrom their flashbacks by narrativizing the traumatic scene? Thesespaces of return to the “I” have never been studied in French in thediscourse of patients suffering from posttraumatic stress disorder.In addition to the linguistic form “I,” what other forms (personal,exclusive, inclusive, generic) do speakers of traumatic narrativesuse in their discourse in order to refer to themselves? Methods. – Based on the testimonies of survivors of the Bataclanattack collected immediately after the event and a few years later,as well as on the narratives produced by military patients sufferingfrom chronic posttraumatic stress disorder, we analyze, through theprecise linguistic study of pronominal markers, how the speakersrefer to themselves.Results. – A detailed analysis of the pronominal forms and thedifferent values they signify led us to four main findings: (i) over-representation of the first person in traumatic narratives, both bythe use of the pronoun “I” and by others pronouns including thespeaker; (ii) predominance of the use of generic pronominal valuesin traumatic narratives; (iii) preferential use of the exclusive valueof the pronoun “on” in the “War in Afghanistan” corpus; and finally(iv), the temporal unfolding of these first three results mirrorsthe habitual chronology that characterizes the clinical evolutionof posttraumatic stress disorder.