Neuter (n.a.) = abl.?

A few more quotes from Abondolo: The Uralic Languages:
"
Case
In proto-Uralic, the noun had at least two grammatical cases: an
accusative *-m, which probably was used chiefly to mark the definite
direct objects of finite verbs (i.e. verbs inflected for person), and
a subordinative suffix *-n which functioned as a
genitive/prenominalizer with nouns and as an adverb-formant with
verbs. There were also at least three local cases, including a
locative *-nA, a separative *tA ~ *-tI, and perhaps the latives *-k
(and/or *-n,) and *-cJ (and/or *-nJ).
"

"
Number
Number in the protolanguage was marked with two different suffixes,
but in all likelihood plurality was the salient semantic component of
neither. One suffix, *-t, seems to have performed duties analogous to
the absolutive (nominative) -0 in the singular, but with additional
meanings of definiteness and non-singularity. The other, *-j,
functioned as a non-singular analogue both to the singular accusative
*-m, marking certain direct objects, and to the singular genitive *-n,
marking nouns in prenominal position (qualifiers, possessors). This
*-j suffix also preceded further inflectional suffixes; whether case
suffixes were so used is unclear, but the person suffixes must have
occurred in such strings (Janhunen 1981a: 29).

The plural marker *-t is preserved as a suffix of nominal inflection
only at or near the periphery. Note the nominative plural of 'tree':
Taz Selkup poo-t, Sosva Mansi jiw-&^t, Finnish puu-t, Moksha Mordva
s^uft-t. There is no reason to doubt that the glottal stop of the
Nenets and Nganasan nominatives plural, e.g. Forest Nenets pJa-?
'trees' and Nganasan mun,ku-? 'forest' are also from this *-t; the
Kamassian pluralizer -?je ?- ~ -?ji?- ~ - ?i seems to have contained
both *-t and *-j. New plural suffixes have replaced *-t in the noun
paradigms of Mari (pu-wlak), Permian (e.g. Komi pu-jas), and Hungarian
(faa-k), although Mari has preserved *-t in collectives formed from
kinship terms and names, e.g. awa-m-&^t MOTHER-sl-plur 'my mother and
her associates'.

The other plural marker, *-j, now marks plurality of possession in
Hungarian, e.g. faa-i-m 'my trees', cf. faa-m 'my tree'. In many
Fennic languages it is used in all non-nominative cases, e.g. Finnish
pu-i-ssa TREE-plur-ine 'in trees'; Estonian uses either *-t- or *-j-,
e.g. puudes/puis. Plural genitive forms such as Finnish kalojen (<
*kalo-j-D-en < *kala-j-t-en) 'fishes pG' show that the two plural
suffixes could co-occur in the west, as well; contrast the parallel
(stylistically marked) kala-i-n 'fishes pG' < *kala-D-en, Suhonen
1988: 309. In Nganasan and Nenets, *-j combined with stem-final vowels
to produce complex vowel alternations.
"

"
The separative *-tI (~ *-tA) is reflected in ablative forms such as
Moksha Mordva oj-dJ& 'butter sAbl' and Fennic partitives such as
Estonian või-d 'butter sP'. In Samoyedic, it occurs with the same
co-affix as the locative, e.g. Tundra Nenets mya-kø-d° 'tent' sAbl,
and in Fennic and Mordva it occurs with the same *-s- coaffix, e.g.
the elative forms Moksha Mordva kud&^-st&^ (< *kud&^-s-t&^) and
Finnish talo-sta (< *talo-s-ta), both '(coming) out of the house'.
In Mari, a trace of the separative remains in the c^J at the ends of
the postpositions g&^c^J and dec^J, as in pört g&^c^ '(coming) out
of the house', jes^-&^z^ dec^J FAMILY-s3 FROM 'from his/her family'.
"

Now, the situation in Mordva, where in the plural the only case
that is different from the corresponding singular one is nom. in
*-t, the situation in the Baltic Fennic languages, where similarly
the nom.pl. suffix is *-t, but where all the other plural case are
constructed with completely different suffix -i- < *-j (but Estonian
has later changed that) might make one suspect that the plural
nom. t-suffix is somehow identical to the ablative -> partitive
suffix *tI ~*tA (if so, the situation in Mordva, where the nom.
pl. is 'just another case' and not a separate category (ie not
really a nom.pl. but a kind of special partitive), is the
original one, and not a case of syncretism and case loss).

Similarly in PIE, there was an ablative in *-od (but only in the
thematic inflection), and an ending *-d in the nom.acc. of the
neuter. Now suppose the neuter *-d is the 'missing' athematic
ablative, here used as a partitive (as in the Mordva pl.)?

If it is, it would remove the only true difference between
PIE m. and n. (the use of n. thematic -om in the nominative
could be seen as a problem of use and considered unoriginal)
and PIE would have had only one single genderless paradigm.

Also it would explain why Fennic(?)-speakers learning Slavic
would use *-od for n.nom.acc.: they saw it as a partitive.
Or perhaps that form is original PIE?

Now how could a case have been borrowed? But we don't have to
assume that, cases were originally postpositions, which were
originally particles, in this case *tV or the like. It may also
have survived in Russian ot "to", WGermanic *to, Gothic do.
The alternation d/t is probably the result of cliticization and
decliticization. The fact that the PIE ablative mostly is *-d,
but the Skt. ablative is -t and the Hittite is -az could be
explained similarly.


Torsten