This PhD thesis explores the way multiple and tenuous links connect collective memory and space in the contemporary city. It does so through a field enquiry carried out in Rome, Italy. This palimpsest-city has a particular relationship to the past, which confers it an exceptional symbolic status in Western imaginary. However, the choice has been made to approach it from its margins and to interrogate a dissenting Roman memory: that of the working-class. The Roman industrialization project of the end of the 19th century – strongly contested and brutally interrupted – has deeply marked the landscape of Ostiense, situated south of the historic center, in which most of the empirical research was conducted. Through a multidimensional analysis of this territory, this study interrogates the meaning and forms taken by the working-class memory in a city mainly considered as a non-industrial city. In this perspective, the thesis proposes a reading of the territory from three complementary angles: the one of the historiographical controversies which developed around Ostiense’s trajectory; the one of the everyday life, as we have been able to observe it through the field enquiry; the one of the public policies which were implemented in the zone (urban planning and heritage conservation). The research shows that apprehending the working-class memory of Ostiense requires to revisit the categories elaborated with regard to emblematic industrialization cases in Italy, Europe or the United-States. Indeed, Ostiense has represented a sui generis industrial zone, be it from the point of view of the activities covered or in regards to the socio-professional categories and the class relationships implied. Based on this reframing process, the study highlight a plurality of signs – material traces and individual narratives – which testify of the existence of a memory linked to the industrial past. Simultaneously, it reveals two types of obstacles to the existence of a memory lived as collective. The first is temporal and relates to a series of ruptures and discontinuities which have characterized the zone and submitted the former production sites to a process of extended abandonment. The second is spatial and linked to the way in which the territory tends to inhibit social dynamics that are favorable to the sharing of individual memories and that because of the sensory characteristics inherited from its past. While recognizing the weight of these obstacles in the access to a satisfying memory experience for our interviewees, this study calls for their recognition as constitutive components of the industrial legacy.