Another Proto-Mississippi Valley *py Set

[Koontz John E]

Note that this was a follow up on the discussion of *py and *ky in
conneciton with 'pot' and 'ice'.

On Sun, 5 Oct 2003, Jimm GoodTracks wrote:
> "pyúbraN" (mint; Indian perfume plant) is from:
> pi    =    good
> ubráN    =    to smell; emit an odor
>
> The "y" is the contracted sound resulting from  "pi + ubráN".

Aha!  After I got done kicking myself for not seeing this it occurred to
me that next question was what does OP nubdhaN mean?  If it has some
fairly straightforward analysis that explains nu from another source, then
this set is just a figment of my imagination.  Otherwise it offers a nice
example of how *py clusters can arise.  The obvious possibilities for nu
that occur to me are niN 'water' + ubdhaN or compounds with nu 'male' or
nu 'potato'.  None of these seem especially plausible, but I may well be
missing something.

If this is a valid example, then note that *py here is from *hpi + u.

> I am aware that many of the entries on the English side do not appear in
> the IOM side. In the enlarged unabridged edition of the dictionary,
> these lapses will be added. Jimm

I wouldn't worry too much about this - in the sense of being embarrassed
about it, anyway.  Essentially all hand-compiled bilingual dictionaries
have this problem, I think, e.g., the LaFlesche Osage Dictionary.  My
comment was more in the sense of warning people how to find the reference,
and that the problem could occur, since your dictionary is actually
pretty reliable about having words on both sides.

One way to be fairly certain this doesn't happen, though it is not an
absolute guarantee, is to not maintain the two halves of the dictionary as
separate databases, but produce one or the other from the other by
mechanical (computer-programmed) inversion.  For example, Bob Hsu's
seminal dictionary software would allow you to precede any word in a
definition (or other field) with an asterisk, like this:

hw phyubraN
ma phi good
ma ubraN have smell
def *mint
def Indian *perfume
...

When you ran the inversion module it would produce an English variant of
the database with entries like this:

hw mint
gl phyubraN

hw perfume, Indian
gl phyubraN

There's nothing magic about the use of asterisk as the marker, but this is
definitely an extremely important idea.

Bob Hsu's programs were called Lexware, as I recall, though I've tended to
think of them as Hsubox since Bob Rankin came up with that version.

I couldn't resist throwing in those ma fields, too.  Of course what the
formatter produces from this entry is something like:

<bold>phyubraN</bold> 1. mint, 2. Indian perfume, = <bold>phi</bold>
<bold>ubraN</bold> 'good' + 'have smell'.

And you can hav a tool that generates temporary supplementary entries for
the IO side from the ma
fields, e.g.:

hw phi
xref phyubraN

hw ubraN
xref phyubraN

leading to formatter output like

<bold>phi</bold> see <bold>phyubraN</bold>

Of course, with appropriate modifications you can get the definition in
here, too:

hw phi
xref phyubraN
def mint
def Indian perfume

<bold>phi</bold> <italics>cf.</italics> <bold>phyubraN</bold> mint, Indian
perfume

The more you do mechanically, the easier it is to avoid simple editorial
gaffes.  Of course, you have to be alert from programming errors ...