Re: [tied] number of cases in PIE

From: Jens Elmegaard Rasmussen
Message: 17940
Date: 2003-01-22

On Mon, 20 Jan 2003, Andrew Howey wrote:

> Hello, all: I was wondering how many cases there really were in PIE, or
> more specifically, Late PIE. I specify Late PIE because according to
> most information I've read, PIE went through an earlier "ergative" stage
> first. In any case, most sources seem to agree that there were eight
> cases: nominative, genitive, dative, accusative, instrumental, ablative,
> locative, and vocative.

Anybody got an idea as to what underlies Latin local adverbs in -inc 'from
...' and in -u:c 'to ...' (hinc, illinc 'form here, there' : hu:c, illu:c
'to here, there')? They look like older cases, but which ones? The forms
in -inc resemble in-de 'from there' which has un-de 'from where' with
coloration from the (lost) labiovelar beside it matching OCS koNde^, koNdu
'from where'. The vocalism is as with ibi : ubi and Skt. iha : kuha. Much
of this must be of PIE date, but with what status?

As for the classical eight cases, it looks as if the "strong cases"
without (original) vowels in their endings make up a basic core, being
simply the nominative and accusative (and the unmarked vocative which may
not be called a case at all since it enters into no sentence structure).
Other cases are formed by the addition of (underlyingly) syllabic
morphemes to the stem. However, none of the case morphemes has any obvious
internal etymology, so they are not well-defined postposition collocations
(although they may well have such an origin in a longer perspective). The
locative singular is a special matter: Always accented on the final vowel
of the stem and without desinential morpheme (not counting the optional -i
which probably just meant something like 'there' and was a separate word),
the locative must have been marked by an enclitic vowel: Enclitics are
preaccented, and unstressed short vowels are lost, so "+E" would cause
just that. Still, that opens no interesting further insight to me.

With the personal pronouns, it is surprising to find that the "weak cases"
are all founded on the accusatives, quite in contrast with the general
system of IE case inflection with nouns. The possessive adjectives are
formed by derivation by means of the thematic vowel forming adjectives, as
*twe 'thee' => *twoï-s 'thy'. As genitive of the personal pronoun was used
an uninflected substantivized form made by "vrddhi", a derivative
procedure which is in origin simply the "contrastive accent" causing
switch from adjective to substantive and vice versa, which had the
side-effect that a vowel which was otherwise lost was retained if it
received the accent in time. Thus gen. *teïwe 'of thee', some uninflected
form of *twoï- with shifted accent. If this is the whole story, the
widespread variant type *tew¢-s 'thy' is most easily explained by

Another surprising thing with the personal pronouns is to find that the
accusative ends, not simply in *-m, but in *-meï (to which the endings are
added in the weak cases). I have no idea what the final accented /-e/ is
meant to express. It may be a simple prop-vowel, for all otheer forms of
the personal pronouns are vocalized before the final consonant and so end
in only a single consonant; then, if the acc. was formed by adding /-m/ to
the nom., the result would be a word ending in two consonants. However, I
am not familiar with any other instance where an accented vowel is being
added simply to avoid a final cluster.