a phonetic mystery

[Koontz John E]

On Fri, 20 Apr 2001, Bruce Ingham wrote:
> I'm glad that everyone else finds this confusing too.  Actually I think the word
> for 'steal' is written manun or manon in everything I've ever seen.
> Not man[u] (?)

Perhaps the enclitic boundary is relevant?  It might in some sense block
the interactions between manu(N) and pi, whereas there is no such boundary
between nuN and pa in 'two'.  In that case =ktA should behave similarly,
and =pi with other stems.  Obviously the ablaut vowel conditioned by =ktA
make that something of a special case.  This might clarify whether iN
behaved as if it were in the stem or the enclitic.

Is there any boundary in 'swim'?