“A language reconstructed on
the basis of meager inscriptions does not acquire substance until one proves
that, on the basis of those inscriptions, that language and only that language could
be reconstructed. But in our case, on the basis of so fragile a collection of
data, it might be possible to construct or reconstruct not one but a hundred
languages. Thus one would be confronted by the amusing case of a piece of
poetry which could have been written in any one of a hundred languages, each dissimilar
from the others and from the first…”
