Из "Encyclopedia of Indo-European Culture", о келтским језицима:
Generally, the archaeological evidence for continuity can be traced further to the late Bronze Age Urnfield culture of the thirteenth century BC and even earlier but with a decreasing sense of linguistic utility. It is clear from the Celtic languages that they shared common words for 'iron' and other technological items (“shield', 'caldron', etc.) that they are unlikely to have encountered anywhere prior to the Urnfield or late Bronze Age. Moreover, the Golasecca culture of north-west Italy is part of the general Urnfield phenomenon and there is some evidence of the spread of the urnfields into northwest Iberia as well. Hence, the initial spread of the Celtic languages may have begun during the late Bronze Age and involved some population movements and later migrations are suspected for the Iron Age. But major shifts in populations are not envisaged: analysis of skeletal remains from the Hallstatt and La Tene, for example, point to broad homogeneity among western and central European populations with more marked differences between them and those of the British Isles. Similarly, it has not been possible to discem any particularly genetic features which are shared by all of the main Celtic-Speaking populations, even in the British Isles. Here the physical and genetic composition of the Celtic populations has generally been regarded as merely “residual" or “peripheral" European rather than particularly derived from some continental Celtic “homeland". For this reason
archaeologists in western Europe have also emphasized other social processes that may have led to the spread of the Celtic languages. The establishment of Bronze Age hillforts and later centers of Iron Age chieftains has been viewed as providing
an arena for language change and diffusion with varieties of Celtic expanding along marriage networks between the social elites (vaguely like French among the Russian nobility of the early nineteenth century) or via travelling craftsmen who received the patronage of such elites. lt has even been suggested that Celtic may have served as something of a pidgin or lingua franca among the trade-routes of western Europe. This latter theory, however, seems most unlikely as pidgins are characterized by brusque simplification of grammar, a feature that is hardly supported by both the conservatism of reconstructed Common Celtic and the complicated evolution that some of the Celtic languages took, such as Old Irish with its augmentation of the existing verbal forms.
Finally, there are those who hold to a theory of Indo-European origins that would seek the roots of all IE Stocks in Europe in the spread of the agricultural economy from the Near East. Such a model would have the (Proto-) Celtic stock emerging out of the languages of the Neolithic inhabitants of western Europe during the period с 5000–4000 BC. This model, however, seems most unlikely given the general similarity of all the Celtic languages with one another that we find with the first inscriptional evidence, e.g., the Old Irish expression 'the women' would be rendered inna mná which, were we to find it on an ogham inscription of the fourth- seventh century AD, would have been written *indás mnás, the precise form that we do find it on a Gaulish inscription of c 100 AD. It is most improbable that the (Proto-) Celts were able to maintain parallel linguistic development from Ireland across western continental Europe from the beginning of the Neolithic to the historical period, a time-span on the order of four thousand years. For this reason, linguists have generally confined the search for the Proto-Celts to the later Bronze Age (c 1200 BC onwards) or the lron Age."