-n- adjectival suffix in Latin

[Jens Elmegaard Rasmussen]

On Thu, 22 Jul 1999, Paolo Agostini wrote:

> [Can ...] the Latin adjectival
> suffix -n-us, -n-a, -n-um appearing in words like _mater-nus_
> [...], _pater-nus_ [...], _feli-nus_ [...] be traced back to IE? Does the
> morpheme exist in other IE or non-IE languages? Any idea abt its etymology
> and/or development?

This thread has already provoked replies from Damien Perrotin and Eduard
Selleslagh involving Etruscan and some more promising nasal suffixes of
diverse IE languages. The suffix involved must be *-ino-, a common suffix
of appurtenance, particularly well attested in derivatives from names of
seasons and other time spans, as Gk. earino's, OCS vesnInU 'of spring'. It
was plausibly derived by Chantraine from the locative of the season name,
which was typically an r/n-stem, as *we's-n-i 'in spring', adding the
"thematic vowel" *-o- (the real seat of the expresion of appurtenance),
with a histus-filling ("ephelcystic") *-n- in between, and dissimilation
of the product *-ni-n-o- to *-ri-n-o-. The r/n stem seems to have been
generalized for all seasons, cf. the beautiful pair Lat. hi:bernus = Gk.
kheimerino's equated by Szemere'nyi as IE *g'heimerinos (with dissim. in
Lat.). The part -rnus was repeated with other designations of time in
Lat., as diurnus (from the loc. diu:, the "endingless" variant of Skt.
dya'v-i) and from there noct-urnus (if not from noctu:, itself copied on
diu: at a time when this still meant 'throughout the day, all day long'
and not just 'long').
   The special meaning of the old derivatives in *-ino- makes it sensible
to derive them from the locative: *wesr-ino-s is 'what is _in_ spring'.
This distinguishes *-ino- from the suffix *-io- which has no such obvious
connection with the locative (although it has been argued to have
precisely that origin). A derivative like *ek'wi-o-s 'pertaining to a
horse' does not signify thing on or inside a horse with any preference
over things connected with a horse in a non-local fashion, so the suffix
form *-i-o- is simply the product of the addition of a "thematic vowel"
*-o- to the bare stem normally posited as *ek'wo-, in which we observe the
transformation of two thematic vowels to *-i-o-. I see this as a simple
consequence of the reduction rule of an unaccented thematic vowel to *-i-
applying in very old lexicalizations; since _both_ thematic vowels could
not be accented, the product *-i-o- may simply be from *-o- + *-o-. It
seems that the addition of a syllabic morpheme shifted the accent towards
the end, so that *-o'- yielded *-o-o'- whence *-i-o'-. It should not be
held against the analysis that examples with an independent accent show
*-i'-o- with accent of the -i- part, for that would be the further
development anyway if the form is older than the introduction of initial
accent I claim to have discovered for a prestage of PIE. This analysis
provides an answer to the question why there is no *-n- in *-i-o-: there
was no word boundary here, while in the hypostatic derivatives based on
locatives in *-i as *-ri-n-o- there was.
   A preform like *p at 2teri-n-o-s may indeed have existed in IE, but then
with the specific meaning 'which is at the father'; but it may just as
well reflect a simple Latin (Italic) analogy with the season-based
adjectives. In Balto-Slavic *-ino- enjoyed an enormous productivity (Russ.
vostok 'east', vosto{cv}nyj 'eastern' from *-k-ino-s). This has nothing to
do with the -n- of Germanic n-stems which turns up wherever the stem final
is allowed to surface, not only in the genitive. For 'pertaining to a
father', the conglomerate *-io- of thematic stems was generalized and had
created *p at 2tr-i'o-s in PIE already.
   Thus, the -n- is not in origin a morpheme of appurtenance or of a
genitival relation, and so there is no point in equating it with something
outside of IE which is.

Jens