Thank you for your explanations. They are clear and I now understand the
relationship in this section. This also helps to clear up another thing
I was wondering about. In the same sutta, upanisaa is used throughout,
but with two different prefixes "sa" and "an." These two resulting
terms: "saupanisaa" and "anupanisaa" are used in contrast with eachother
in relation to another term such as "sa"nkharaa", etc.
Here is one example:
sa’nkhaarepaaha.m, bhikkhave, saupanise vadaami, no anupanise.
sa’nkhaare [m-a/acc/pl] kammic formations
pi [indec] also
aha.m [per pro/nom/1^st sg] I
bhikkhave, [m-u/voc/pl] bhikkhus
saupanise [f-aa/acc/pl] has causes
vadaami, [vad I/ind act/1st sg] I say
no [indec] indeed not
anupanise. [f-aa/acc/pl] not-causes
lit: I say, too, bhikkhus, that kammic formations have causes, indeed
not do they not have causes.
Furthermore Bhikkhus, kammic formations have causes; indeed it is not
true that they have no causes
Am I right in thinking that terms that begin with "a/an" or "sa" can
form bahubbiihis with only one other term? I read Warder's comment on
this on page 137, but I didn't know if these prefixes were taken as part
of the compound, or just additions to an already extant compound.
Anyway, the two forms of upanisaa above definitely agree with
"sankhaare" in the above example. If they are indeed bahubbihi
compounds, then this makes it clear why there is agreement.