Armorican-type socketed axes, particularly numerous in Brittany and Normandy, have been the subject of debate on both their purpose and their date since the 18th century. At the end of the 19th century, G. and A. de Mortillet claimed an Iron Age attribution for some medium-sized and miniature non-functional versions, whereas the Brandivy-type functional axes were still dated to the Late Bronze Age. It was however J. Déchelette’s point of view, confirmed by J. Briard in his famous book Les dépôts bretons et l’âge du Bronze atlantique published in Rennes in 1965, which prevailed until the end of the 20th century: these axes were attributed to the carp’s tongue sword horizon from the Late Atlantic Bronze Age 3 (i.e. BF IIIb = British LBA 3 = HaB2-3, around 950-800 BC). Even though J. Briard favoured a production of these axes mostly during the end of the Bronze Age, he also demonstrated that this production continued during the First Iron Age. This traditional point of view of a mostly Late Bronze Age production began to be questioned at the end of the 20th century, even by J. Briard himself, but nevertheless it still persists among some authors. The purpose of this article is thus to produce as comprehensive a file as possible on this issue. The elements testifying to a production of these axes exclusively during the Early Iron Age are indisputable: on mainland Europe, no axe can today be dated for certain prior to HaD. The rare axes currently appearing in some Late Bronze Age 3 hoards can be explained by the ‘pollution’ of these hoards with objects from other origins, as a consequence of hazardous conservation conditions in collections or museums (for instance Plounéour-Lanvern and l’Île Verte, Finistère; Durtal, Maine-et-Loire; Longy-Common, Aurigny). Sometimes, objects from non- or poorly-documented origins have been artificially brought together and presented as hoards, although the latter never existed (for instance Césarin, Corrèze). In other cases, there may be a misinterpretation of the documentary sources (e.g. Louvigné-du-Désert, Ille-et- Vilaine). It is a crucial established fact that no hoard from the carp’s tongue sword horizon among those formerly discovered and indisputably documented, even the largest ones (for instance Vénat, Jardin des Plantes, Prairie de Mauve), has ever yielded an Armorican-type socketed axe. The same applies to all the hoards discovered during the second half of the 20th century and the beginning of the 21st (for instance Gouesnac’h, Finistère; Belle-Ile-en-Mer, Morbihan; Saint-Père-en-Retz, Loire-Atlantique; Challans, Vendée; Meschers, Charente-Maritime; Triou, Deux-Sèvres). Consequently, if these axes had already been produced in their thousands by this period, it seems absurd they should never appear in hoards, either complete or as fragments. Prototypes of Armorican-type socketed axes were abundant in hoards from the Late Atlantic Bronze Age 3: a few short axes decorated with vertical ribs, of British origin or related to them, could herald the Brandivy-type axes, or more generally the other patterns of Armorican-type socketed axes decorated with vertical ribs. Above all, axes of the Plainseau type would seem to prefigure those of the Tréhou type and types close to the Dahouët and Plurien variants. At the beginning of the Early Iron Age, some axes do indeed appear as intermediate forms between the Plainseau type and the Tréhou/Dahouët/Plurien Armorican type (Fossé-Creuzette hoard in Verberie). On the contrary, the small axes from Maure, Saint-James and Couville, as well as those made of almost pure lead, have no prototypes in Late Bronze Age hoards. They could represent a recent production of Armorican-type socketed axes, in other words the end of an evolutionary process leading to the least functional forms, which, due to our ignorance of their real purpose, we might consider the most aberrant of all. Undoubtedly, the production of Armorican-type socketed axes had not yet begun during the Late Bronze Age. Nevertheless, the question of their actual beginning remains uncertain: in Gaul, they all date from HaD. Whether their production could have begun as early as the Early Iron Age (HaC) remains a mere hypothesis given our current state of knowledge, despite a few associations with axes of the Sompting type in the British Isles, or in Brittany in the Hengoat hoard (Côtes-d’Armor): the production of Sompting-type axes goes beyond the HaC Lynn Fawr horizon. The increasing number of hoards of Armorican-type socketed axes is above all one manifestation of a global phenomenon occurring in Gaul during the middle and late phases of the Early Iron Age (HaD): a revival of the burying of metal hoards, similar for instance to the hoards from the Launac horizon in the South of France, or the ornament hoards in the Centre-West and Middle Loire areas (translation E. Thauvin-Boulestin).