The work under review includes seven essays edited by Barry Cunliffe and John T. Koch that address the debated issue of the origins of the Celts. They belong to the Celtic Studies Publications series dedicated to renewing the vision of the Atlantic Celts, a collaboration between the Centre for Advanced Welsh and Celtic Studies at Aberystwyth and the School of Archaeology, University of Oxford.

Cunliffe outlines his thesis on the origins of the Celts, stating that ‘the Atlantic lingua franca developing in the fifth and fourth millennia was Proto-Celtic and that the vector for its extension into middle Europe was the Beaker phenomenon. It was during the Beaker period that the mature Celtic language developed’ (p. 9, see also pp. 182–83). This thesis, aired in Facing the Ocean (2001), assumes that Italic and Celtic would ultimately come from Anatolia via the Balkans around 6500–5500 bc, in association with Impressed Ware decorated with cardial shells (p. 7ff), but Impressed Ware should be considered a ‘substrate’ of Iberian and Sardinian, two non-Indo-European languages, posing a serious difficulty for this thesis.

Chapters 2 and 3, ‘Celtic from the West Meets Linguistics and Genetics’ by Koch, and ‘A Case of Identity Theft? Archaeogenetics, Beaker People, and Celtic Origins’ by Koch and Fernández Palacios, explain research so far. Koch associates the origin of the Atlantic Celts with the Beaker phenomenon. New peoples, probably ‘warriors’ with R1b haplogroup Y chromosome ancestry, arrived from the Eurasian steppes in the Atlantic world, where they replaced the existing male lineages (p. 48). But Koch and Cunliffe maintain that the megalithic, pre-Beaker populations were already Indo-European (p. 40), even though they were a mixture of Mesolithic and Neolithic communities of non-Indo-European ancestry, unless a Palaeolithic origin for the Indo-European languages is argued.

The currently available data contradict the existence of a ‘Proto-Celtic’ lingua franca since the Neolithic (p. 202–03), while recent research seems to indicate that ‘there clearly is a link in burial practice between Yamnaya and the Corded Ware culture’ (Kristiansen et al., Reference Kristiansen, Allentoft, Frei, Iversen, Johannsen and Kroonen2017: 336), and probably also with Bell Beaker users: ‘[…] these fundamental changes have a huge geographic reach, from the Altai to the Atlantic’ (Heyd, Reference Heyd2017: 349). In addition, the vision of language offered, which appears disconnected from the broader ethnocultural system (see Clarke, Reference Clarke and Chapman1978: 101ff., 363ff., figs 15 and 76), makes it difficult to understand the crafting of languages as part of processes of ethnogenesis. Language is related to all other elements of a cultural system, including beliefs and rites, which surprisingly are not mentioned in this book, since religion is an essential component of any ethno-culture, together with material culture, technology, economic and social structure, or ancestry. Not mentioned either is the longue durée in the emergence of Celtic culture since the third millennium bc, as documented by literary sources and the oral traditions of Atlantic Europe, from Ireland to Iberia (Almagro-Gorbea, Reference Almagro-Gorbea2018).

Some details are questionable. For example, the anthroponym Arquius (‘archer’), is associated with the toponym Arcobriga: this is a falsely perceived resemblance between hill and bow (pp. 55–59). Arcobriga means ‘fortified city of Arco’, an anthroponym like Segobriga, which means the ‘fortified city of Sego’ (Abascal, Reference Abascal2002). It is also argued that the Tartessian inscriptions were in ‘Phoenician script’ (p. 12), which can be confusing as the same could be said about the Greek, Etruscan, or Roman script, which all derive from the Phoenician script. It would also be helpful to discuss the first Celtic theonym recorded on a Greek graffiti from Huelva, Niethos ‘bright’, ‘hero’, related to the god Néit in Ireland (Almagro-Gorbea Reference Almagro-Gorbea, Bradley, Coles, Grogan and Raftery2004), as it reveals close Atlantic connections.

‘Connectivity in Atlantic Europe during the Bronze Age (2800–800 bc)’, by K. Cleary and C. Gibson, summarises the changes linked to the Beaker phenomenon in the third millennium bc. Corded Ware had an influence in the making of the Maritime Bell Beaker and its regional, locally developed variants. From the twelfth century bc onwards, during what is labelled the Atlantic Bronze Age, contacts proliferate, metal hoards increase, funerary rites change, and elite feasting equipment (i.e. cauldrons, spits, carving forks) appears (pp. 108ff). It would have been equally relevant to comment on the lyres, as they reveal the existence of epic poetry (Almagro-Gorbea Reference Almagro-Gorbea2018), an aspect that is barely mentioned in the book (p. 73). Chapter 5, ‘Chemistry and Bronze Age Metals in Atlantic Europe: Flows of Ideas and Material’ by P. Bray, clearly explains the changes in mining and metallurgy that took place in the different Atlantic regions, although these are still poorly documented in France and Iberia.

An essential chapter is ‘The Archaeogenetics of Celtic Origins’, by M. Silva and colleagues, which discusses the archaeogenetics of the Celts, albeit centred on the British and Irish Islands (pp. 151–67). DNA studies have shown genetic flows, due to population movements, from the Eurasian steppes westwards that explain the expansion of Indo-European languages, a process that coincides with the molecular dating of haplogroup R1b of the Y chromosome. The variant R1b-M269 is dated 6500 bp, and the subvariants R1b-L11 and Rib-S116 c. 4900 and 4600 bp. These dates coincide with the expansion of Corded Ware and situate the beginning of a ‘Celtization’ process from 2800 bc onwards, not in the Neolithic, as Cunliffe and Koch maintain. As Silva et al. point out, Celtic languages seem to have arrived in Western Europe only about 4500–4000 years ago. The data are compelling and more convincing than the two phases of Celtization proposed by Cunliffe and Koch, a Chalcolithic phase and a Mediterranean Neolithic phase associated with Impressed Ware decorated with cardial shells (pp. 196ff., 202). This I find unconvincing because, as indicated above, such Impressed Ware is associated with a substrate of Iberian and Sardinian populations speaking non-Indo-European languages.

It would be appropriate to rectify the false idea that the prehistoric populations of the Basque Country spoke the Basque language, an error repeated in the work (pp. 24, 70, 156, 165, 204, figs. 3.1, 3.2, 3.5 and 3.10), as well as in works by other linguists and geneticists (Olalde et al., Reference Olalde2019: 1230). The toponymy, anthroponymy, and theonymy of the Basque Country indicate that it was inhabited by Celtic communities in Antiquity (Villar, Reference Villar2014), descendants from populations that arrived during the third millennium bc according to the archaeological evidence (López de Uralde, Reference López de Uralde2010), while the Basque language penetrates into the Basque Country in the Early Middle Ages from Aquitaine (Gorrochategui, Reference Gorrochategui1984). This is supported by genetics, since the Basque Country, together with Ireland, shows the highest frequency of hemochromatosis (p. 177) and the haplogroup R1b-M269 of the Y chromosome, dated c. 5000 bp, reaches in the Basque Country almost one hundred per cent, as in Ireland (p. 70, 156), compared to sixty-five to seventy per cent in Portugal and north-eastern Iberia.

In order to understand the ethnogenesis of the Celts in Iberia, it is necessary to take into account the two main environmental areas of Iberia, the Atlantic and the Mediterranean, which coincide with the two wide linguistic and ethnic areas distinguished by Untermann (Reference Untermann1961): Celtic Hispania and Iberian Hispania (figs 3.2, 3.5, 3.10). The southern and Levantine areas were inhabited by Tartessian and Iberian Mediterranean ethnic groups, in contrast to the Celtic populations found on the Iberian Central Plateau and the Atlantic areas, which included the Lusitanians (pp. 27–30, 203). But, by the end of the second millennium bc, Urnfield people from Central Europe appeared in north-eastern Iberia. Iberian Urnfield communities gave rise to the Celtiberian language and culture (Ruiz Zapatero & Lorrio, Reference Ruiz Zapatero, Lorrio, Karl and Stifner2007), the most important Celtic culture in Iberia, barely mentioned in Cunliffe and Koch's book (pp. 28–29).

To achieve a broad perspective, some essential works should be mentioned (e.g. Villar, Reference Villar2014, not cited). The work presented by Cunliffe and Koch constitutes an important contribution to the understanding of the Atlantic Celts, but it disregards the continental Celts, and without the latter it is not possible to achieve an integral vision of the ethnogenesis of the Celts; they cannot be understood if they are analysed solely from the British side of the English Channel. In continental Europe, the Celts or Gauls go back to La Tène (500–50 bc), Hallstatt (800–500 bc), the Urnfield period (1300–750 bc), the Tumulus tradition (2000–1300 bc) and, ultimately, the Beaker culture (2500–2000 bc) in contact with the Corded Ware culture. Gauls, Celtiberians, Ligurian and Lepontic people do not come from the Atlantic Celts, but from a diverse third millennium bc substrate in continental Europe.

To conclude, the work presents the most recent scientific advances on the origins of the Celts, which makes it attractive, although perhaps controversial. Added to its scientific interest, the book combines a didactic perspective for non-specialists with a bright and clear scientific discourse, encouraging us to delve into the fascinating origins of the Celts. I unreservedly recommend reading this book and congratulate the editors and authors for offering the public such an interesting work.