In the early eighth century, language was fundamental to the identity of the Irish as a single, unified people. So, at least, was the opinion of the author of Auraicept na n-Éces (‘the primer of the poets’) who claimed that the Irish language was invented at the school of the grammarian Fénius a few years after the division of the world into different languages at the Tower of Babel. The Irish, then, came to be a people through the creation of a shared language. Nor was this any ordinary language: the Irish language was not the result of divine punishment but was rather invented through combining the best elements of other languages. This shared language was the best language of all, and the Irish, by extension, the best people. This association between a people and their language was far from unique in the early middle ages. We see something similar in the Historia ecclesiastica, for example, where Bede links four of the five languages of Britain to specific gentes, the final language, Latin, being used by all.

Such medieval texts clearly viewed linguistic difference as worthy of comment – a shared language could contribute to the definition of a gens and distinguish it from other gentes. But where was language located in the constellation of characteristics that might be drawn upon to construct group identities? Rees Davies did not deny that language was important, but also observed that medieval writers were more obviously preoccupied with other characteristics, such as shared history and law. And although linguistic difference is acknowledged in medieval texts, it is not often presented as a source of contention between peoples. In the context of early medieval Europe, the importance of language is in further doubt. Walter Pohl has dismissed language as an indicator of ethnic identity, arguing that most people were bilingual, and that medieval writers rarely commented upon language. This view is widespread among the proponents of the ethnogenesis model put forward by the Vienna school. The overarching consensus, then, is to caution against overestimating the importance of language to identity construction in the middle ages.