(There is) the word DHAATU according to the view ( i.e., the usage)
of the Buddha. According to my view it is (a word) of the feminine
gender. In the science (of grammar , i.e., as a grammatical term) (it
is) in the masculine gender. According to Kaccaayana (it is) in both
(genders).
[...]
The second sentence is a terse statement in stanza form. I venture to
think that the text is corrupt here. It is likely to be
"itthiling(o)'attano mate". The word 'itthilingo' would go with
'dhaatusaddo' , the initial word of the stanza. >>
Jim: I have to disagree here. The reading 'itthili"ngattane' is
perfectly fine. The suffix '-ttana' can be found at Sd 780
(.nya-tta-ttana-vya-taa bhaave) and is equivalent to Skt. -tvana.
Since Aggava.msa uses the masculine form of dhaatu in the sense of
verbal root we can conclude that his opinion is the same as that of
Kaccaayana. I would translate the first line as:
The word 'dhaatu', in the opinion of the Buddha, is understood (to be)
in the feminine gender.
Notice my insertion of 'to be'. So it does work here also with the PPP
'mata'.