The article "Gaulish" has this paragraph which I find interesting

Can anyone explain what this means?

=============================
But what about Caelum and Cieux? Without knowledge of the singular
form ciel,
the neutral observer would hardly define Caelum and Cieux to be the
same state. The coding
question boils down to: Which cutoff is appropriate for deciding when
to consider two items as
identical? The answer is that the cutoff should be dictated by the
least-known language to
ensure a level playing field for all languages in the phylogenetic
analysis, otherwise the branch
lengths in the phylogenetic output would not reflect language evolution,
but only the researchers
extent of knowledge on the languages involved. This requirement, i.e.,
to consciously ignore
etymological information when coding a mix of well-known and
lesser-known languages, is
probably the most difficult and subjective step in the entire analysis.
The necessary decisions
may be checked using a negative control. In this example, a negative
control might be the text in
a language such as Basque which is effectively unrelated to any of the
languages under
consideration. The negative control would be aligned and coded blindly
along with the other
languages, and the number of identical codings between unrelated
language pairs would give an

--
Mark Hubey
hubeyh@...
http://www.csam.montclair.edu/~hubey