Buck on the WWWeb?

[ECOLING at aol.com]

As I read messages on the IE list,
and piecemeal discussion of particular word families,
it seems to me that one of the most useful things we could possibly
have is a fully-updated version of something like Buck,
but organized by etymon not by semantic meaning.

This could be something like a dissertation project
and more, not simply a scanning of already existent data.

Pokorny is not easily usable in the fashion I have in mind,
though I know its computerization has been mentioned.
It uses a visual format in which one must read through a paragraph
to find info, not as visually accessible as it could be, perhaps,
it does not even make use of the vertical vs. horizontal dimensions
of the page.  That to me is now too old-fashioned, a waste
through non-use of potential information-carrying display devices.

Buck is much more accessible.  Can the two be merged in some way?
A recent dictionary of verb stems (citation not handy) is another source,
and its publication is recent enough that a CD could perhaps be
made available?

Just consider the amount of evidence available to correspondents
on this list.  What if we had a simple listing of cognate
roots in each of the languages, with some cross-reference notes
such as that both /durgn/ and /borgn/ in Armenian share the
non-satemization of the final /g/?  (pardon if my memory for
the spellings is not right; I'm referring to a message received today).
Or what the meaning contrast might be in particular languages
where two roots apear with related meanings, such as between
reflexes of the /hurkis/ Hittite for 'wheel' and the *kwel- root?
When such info is available (more of that kind is in Buck,
sometimes).

Attestations in the left column, for fast searching,
comments to the right.  It takes more paper, in printing,
but is infinitely more usable, and does not take more storage
on CD or on the web.

Etc.

We need some new tools, and creating them requires work.
***It also requires a very balanced individual to supervise, one who will
neither try to prove how rigorous they are by excluding almost
everything interesting or on the edges of knowledge,
nor one who will include every possibility without distinguishing
the kinds of support available, or not, for particular analyses.***

If there are two word-families which within themselves cohere
quite clearly, but which pose some problems for linking the two
together, then that should be overt and public, not hidden by either
lumping them or splitting them.  There *are* ways of noting
possible links without having to take a final position on their
validity or not, while still giving reasoning both for and against hypotheses.

Lloyd Anderson
Ecological Linguistics