Case suffixes with trapped pronouns/person affixes.

[The Hanged Man]

There are (at leas) 23 languages with person inflection on adpositions used with nouns (according to this article on WALS). Abhaz is an example of such a language, as shown in the article:

Code: Select all

a-jə̀yas a-q’nə̀
def-river 3sg-at
‘at the river’ 

It is a general rule that adpositions arise from nouns or verbs, so one could think that person inflection of adpositions could come from person inflection applied to ancestral verbs or nouns, and that possibility is also stated in the article, with pointing out common (occuring in about 77% of languages) similarity of person affixes on adpositions either to personal possessive affixes, or verbal person inflection, or both.

Anyway, what I'm interesed in now is what happens next. Another step in grammaticalization cline could be turning person inflected adposition into a case suffix (if it would be originally a postposition) - provided it would preserve person inflection. The result would be former person inflection trapped between noun root and former postposition. Since probably the most frequent person inflections on adpositions used with nouns (or I should rather say: the only person inflections on adpositions used with nouns) would be those for 3rd person, resulting case forms would be additionaly inflected for number. In other words, the minimal scheme for noun inflection would look like this:

Noun-SG
Noun-SG-3SG-Case

Noun-PL
Noun-PL-3PL-Case

The question is: do you know any attested examples of such outcomes? I have read somewhere that some papuan languages underwent evolution of person-inflected postpositions I am talking about, and perhaps another, non-papuan language did this, but I was unable to recall or find anything specific about source of this information and its details, so I'm hoping someone here knows something more...

Thank you in advance!

[Pole, the]

Maybe it's not what you are asking for, but in the Slavic inflected adjectives have absorbed inflected pronouns leading to:

Psl. *dobrŭ jĭ
good.SGM.NOM 3SGM.NOM

Psl. *dobra jego
good.SGM.GEN 3SGM.GEN

… becoming:

Pl. dobry
good-3SGM.NOM

Pl. dobrego
good-3SGM.GEN

[hwhatting]

It is a general rule that adpositions arise from nouns or verbs, so one could think that person inflection of adpositions could come from person inflection applied to ancestral verbs or nouns, and that possibility is also stated in the article, with pointing out common (occuring in about 77% of languages) similarity of person affixes on adpositions either to personal possessive affixes, or verbal person inflection, or both.

But that's not the only possible road. In the example I know best, Old Irish "inflected prepositions", the development was like this:
1) PIE originally didn't have prepositions, only adverbs of place / time, some of which clearly go back to case forms of nouns
2) In late PIE or in the early daughter languages, some of these adverbs came to be used as prepositions, governing specific case forms of nouns and pronouns
3) In some proto-stage of Old Irish, the prepositions fused with clitic forms of the pronouns, forming the "inflectional" paradigms attested in Old Irish.

So, this is a third possibility, where the person inflection is neither caused by a nominal origin of the prepositions (noun + possessive), nor by a a verbal origin (verb + object pronoun).

[The Hanged Man]

Recently, I've been revisiting the idea again, so I'm bumping this thread to ask: does anyone know or recall any specific information on the topic?

[Salmoneus]

Well, Indo-European has these very odd endings that show case, person and number all at once, as in your example. Iirc some IE languages do seem to double-mark categories (i.e. have endings of the kind -pl-3pl or -3-3pl), but this was probably in response to mergers in the system introducing ambiguities that were resolved by analogising the most marked suffixes into other parts of the paradigm.