- Chapter Two -
the reconstruction of particular features of the Ursprache should not be discouraged,
because even an idealized model reconstruction should not be deprived of the
features present in any living language'.22
Schmidt's quest for criteria to enable us to recognize the order of emergence
of the historically attested Celtic languages has attracted considerable attention,
interesting with regard to methodology and important for our subject. Professor
Oswald Szemen!nyi remarked on the fragility of views concerning the incidence of
a vocalic element a and e in the development of the syllabic nasals.^23 This criticism
has been embroidered of late by Professor Kim McCone.^24 He has been obliged to
concede that the testimony of the termination of the Gaulish form OeKaVtell (/-ev)
(acc. sing.) 'tithe'25 may still be anomalous, not to say decisive, in his detailed review
of the development of the syllabic nasals in Celtic. The development of *kw to ;~p
(altogether characteristic of both British and Gaulish) is dismissed by him a little
too summarily as being 'not at all reliable as a criterion for the genetic classification
of Celtic languages'. He plausibly assumes that Celtic (recte British?) Britain
received its p-isogloss (structurally motivated, as he thinks) from northern Gaul,
possibly through Belgic incursions in the second or third century Be. For McCone
the occurrence of the change '~kw > '~p 'in both Gaulish and British falls well short
of necessarily implying a Gallo-British subfamily, from which Irish must then have
already separated at some earlier stage'.26 He has consistently tended to favour the
view that there was at one time a Common Insular Celtic (he has not declared any
good reason for arguing that there was a Common Continental Celtic, and this is
understandable). For me McCone's argument is inconclusive^27 and seems to strain
overmuch to see patterns in Insular Celtic evidence inherited from an underlying
commonality pre-dating the separating out of Goidelic and British (e.g. in the
augment morphology of the verb, in the absolutelconjunct contrast in verb forms
and in the Old Irish and Brittonic s-preterite forms - all concentrating on morph-
ology), with patterns allegedly distinct from those reflected in the Continental Celtic
evidence hitherto available.^28 He then claims that this indicates that the notional
Common Celtic from which this alleged dialectal split, which he claims we have here,
arose must be assumed to have been in place a long time before the historically
attested 'Gaulish' migrations to the south and west in the fifth century Be. This kind
of allure, favouring an early InsularlContinentallinguistic divergence or split, should
he approached with all due caution. The dating of the divergence is bound to be
uncertain (it must precede the earliest attested linguistic evidence) and the overriding
notion of a Common Celtic subsumed by it all is, for me, vague and unsatisfactory.29
McCone's hypothesis that 'within the Celtic family Irish and British are particularly
closely related through a shared Insular Celtic intermediary'30 is in no way surpris-
ing and should not be discounted. But it is uncertain. There is no way of proving or
disproving it on the basis of the linguistic evidence available to us.
Moreover, McCone sees no difficulty in claiming that 'the "La Tene" migrations
into North Italy and along the Danube as far as Asia Minor from the fifth to the third
centuries Be are more correctly considered Gaulish [than Celtic]' and that this phase
was preceded by 'Celtic migrations westwards and northwards to Spain, France,
Belgium and Britain' entailing 'the spread of a Gaulish variety of Celtic certainly
characterized by p for kw from a centre not far north of the Alps'. After positing the
12