Classical texts have long proved influential arbiters of views on barbarian peoples around the Empire's edges. Post-colonial perspectives have reacted against these, seeing motives of domination and control in both ancient and modern uses of the sources. Woolf's important book, which he describes as ‘a contribution to this process of decolonization’ (2), finds a much more complicated picture. His subject is this ancient literary tradition, if tradition it can be called — for these were not ethnographies in our sense, but elements built into histories or geographies, and most often compiled in Mediterranean libraries rather than from active fieldwork. He notes in particular the way in which ethnographic information was used within a literary construct — for instance, being placed in narratives to space and slow the story between moments of high action (90). This literary context, and its intended audience, is key — W. notes that many apparent contradictions within the texts make sense when we assume an informed, educated audience. For instance, what he terms the ‘truth gap’, where Britain remained a ‘permanent barbarian theme park’ and fourth-century Gauls could still be described as ferocious and drunken (94, 106–8), makes sense for a knowing audience who would recognise the resonances of these stereotypes from older literature. There was also a clear selection of perspective according to situation: Caesar's Gaul is fierce, Sallust's Africa is wild, because it suits their narrative needs.

Ancient literature provides diverse and often contradictory views of barbarians, but W. notes the surprising lack of curiosity in resolving tensions between different explanations for barbarian groups: was the height of the Gauls due to breeding or climate, why were Asian groups less warlike than western ones, and so forth? The development of critique and argument seen in fields such as mathematics and medicine was not found in these views of barbarians — there was no general treatment of the human condition because, in W.'s view, ethnographic information had specific purposes. There was no real intellectual interest in barbarians. So far, the literary construct dominates.

Yet other important elements lie behind this material. W. notes divergent views in, for instance, explaining the origins and spread of the Celts, or the difference between Celts and Gauls, and sees ancient authors struggling to explain patterns they can observe in their data. Critically, he also finds evidence of tales with different origins — some follow long-running stereotypes of ferocity and barbarity, but the interesting ones are divergent, less negative stories, where druids can be philosophers and Hercules can be founder of the Galatians. W. introduces the ‘middle ground’, a concept taken from the contact zone in the Great Lakes of North America in the period after European contact but before its formal incorporation in colonial structures (17–19). Here, tales were told, changed and retold, serving the needs of both locals and colonisers. In these tales the locals could find resonances and connections with their new worlds — and for the Romans, drawing figures like Hercules or Ulysses into local origin myths provided a way to unite this diversity. Finding this ‘middle ground’ is not easy — the voices are faint, and only a little of this contact phenomenon survives into the literature, but W. identifies it in some versions of tales of the plundered Toulouse gold (74–6), while behind the details of Tacitus' Germania he sees Germanic auxiliaries, bringing tales of areas far away (103–4).

This is a powerful concept and an important way of looking at these texts — aspects of local voices and adaptations begin to emerge from the pages. For W. these tales are essentially new creations, with no evidence of pre-existing local traditions coming through, in contrast to the eastern Empire where local heroes and tales persisted. This is an area meriting debate, and literature is not the only place to look — it lies also in the Roman towns founded on Iron Age oppida, themselves often tied in to older places of myth and history, reusing and recreating them; in the art styles which reform and modify older habits, the provincial sculpture adapted from Classical forms, the burial sites running on from ancestral grounds. W. provides new insights into ancient texts, and stimulating new ways of looking at these ancient views of barbarians — chasing his ‘middle ground’ provides an exciting challenge for Romanists working with other fields of evidence.