- Chapter Two -
Gaulish is sometimes made to embrace both Cisalpine and Transalpine Gaulish,
although it is now fashionable (but somewhat misleading or unfruitful) to separate
out, for northern Italy, a Lepontic or Luganian sector and (less commonly) for
Transalpine Gaul sectors or variants such as Narbonensian and Aquitanian. For
Hispano-Celtic the term 'Celtiberian' finds favour for an area of Celtic linguistic
prominence or concentration in central Spain, especially along the upper and middle
reaches of the rivers Douro and Tagus, and the name 'Lusitanian' is used for a more
westerly area (this is applied to a range of sources concerning the Celticity of which
there has been a lot of controversy). 'Galatian' is used for the Celtic of Asia Minor.
The identification of Celtic in central and eastern Europe and in the Balkans is prob-
lematic, not least because it is scanty and limited to proper names; the manipulation
of Roman imperial province names here for the identification of Celtic and other
contact languages is probably unavoidable, but it is by and large unrewarding.lo
Leo Weisgerber's claimll that we should exercise caution in many cases before we
allocate epigraphic sources to particular languages is still valid. We should also heed
Michel Lejeune's warningl2 that in speaking of Celtiberian, Gaulish and Lepontic
(these are the three dominant identifying tags) we can only speak very generally of
the languages or dialects of particular groups of people who used in various sites or
areas various alphabets (forms of the Iberian, Etruscan, Greek and Latin alphabets in
continental Europe) in their inscriptions and that we really cannot speak (even for
the Celts of the Iberian peninsula or of ancient Gaul) of people constituting 'groupes
d'une complete unite linguistique'. This is the case in spite of the discovery during
the last twenty-five years or so of a number of relatively long inscriptions, especially
in Gaul and Celtiberia, which are demonstrably Celtic and (for all their difficulty)
valuable because of the quite considerable amount of additional information they
provide.
At best the earliest linguistic evidence concerning the Celts reflects fragmentary
languages.^13 But it would be wrong nowadays to accept Zeuss's statement that we
cannot hope to find linguistic variation in this early evidence.14 That there were
divergences or cleavages in a linguistic group attested over such a wide area is not
surprising, however partial or unsatisfactory our demonstration of the differences
observed up to now continues to be.
The broad division of the Celtic family of languages into Continental and Insular
sectors is still, in my opinion, essentially a convenient geographical one, despite all
the interconnections bridging that general divide which have been viewed by scholars
in various ways. Here we should note that the Breton language is usually regarded
as being an Insular Brittonic language, although the case has been argued in various
ways for recognizing in it the partial survival of a northern or north-western form
of Gaulish.ls The ways in which scholars view the status of Insular Celtic as distinct
from Continental Celtic should have an important bearing on our understanding
of the emergence and interrelationship of Celtic languages, of the proper approach
to the interpretation of early Celtic linguistic evidence overall, and indeed our
perception of the relationship of Celtic to other Indo-European and non-Indo-
European or pre-Celtic languages.^16
There is disagreement on the nature and significance of the Continental/Insular
Celtic linguistic divide. This stems in part from the different range of evidence
10