11. The causative
The typical causative formation is a root in o-grade, followed by -ei-e/o- (e.g.
Skt. pa:tayati "he causes to fly" < *potéyeti, root *petH- "to fly"). Some
Sanskrit causatives (most of them with a root historically ending in a
laryngeal, all of them synchronically ending in a vowel) use the suffix -paya
(*-pey-e/o-) instead (dha:- -> dha:paya- "cause to put", r. -> arpáya- "cause to
go"). Since the *-p- is not readily explicable as a feature of the preceding
laryngeal, it stands to reason to see it as a prefix to the element -eye-. The
only such prefix I'm aware of is the preverb *pe(:)-/*po-. If so, this must
mean that the element *ey-e/o- is itself a verb (cf. the optative), which can
then perhaps be identified with Hittite iyami "I make, do" < *ei-e/o-, a verb
which in Hittite verb does indeed occur with preverb p(e)- as piyami "Ich
schicke hin". That a form like *potH-éieti "he makes fly" should become a
causative is not surprising . One Sanskrit verb, bhi: "fear", makes its
causative with -sáya-, perhaps another preverb s-?
As to the o-grade of the root, J.E. Rasmussen has argued for an ancient
consonantal element *O (which I will refer to as **R) appearing in the causative
as well as in other formations (especially thematic nominal derivations like
*bhór-os ~ *bhorós or *tog-áh2). This **R-infix has some peculiar
characteristics: (1) it vanishes in certain complex (heavy) root structures
(e.g. *bhewg-, causative *bhug-éie-, not **bhoug-éie-); (2) it causes a
laryngeal in the same root to disappear (*perh2- -> **pRr(h2)-(m)nós (Grk.
pornós)), as already noted by de Saussure; (3) when preserved (in light roots),
it usually vocalizes as *o (but *o: when stressed, e.g. *swó:peye- "causes to
sleep"); (4) In the thematic formations, the accent falls on the o-grade root
syllable if the root was originally heavy, on the thematic vowel if the root is
light (*terh1- -> tórmos, *ker- -> kormós); (5) the infix *o appears as a
_prefix_ in roots beginning with r- and perhaps *H- (*reug- "to vomit" ->
*orugah2 "act of vomiting").
The reduplicated aorist in Sanskrit is closely associated with the causative.
The reduplication syllable is long, either by length of the reduplicative vowel
in roots beginning with a single consonant (C1i:C1-), or by position in roots
beginning with two consonants (C1iC1C2-). The root is in zero grade, or, in
verbs ending in a single consonant, e-grade. Roots ending in a laryngeal show
the element p- of the causative present in the reduplicated aorist as well
(R-aorist jijñipát, from the root *genh3-, causative jña:páya-), so we must
assume that the verbal root *ei- (unstressed *i) was originally also present,
but that it was completely absorbed here before the stressed thematic vowel for
some reason. Given the parallels between reduplicated aorist and causative
present, it seems reasonable to suppose that the length of the reduplicative
syllable must have the same origin as the o-grade of the causative present. The
lack of lengthening of the reduplicative vowel before a complex initial CC-
cluster matches the vanishing of the **R in causative/thematic roots ending in
-CC.
Given the various reflexes of **R, we can conclude that we are dealing with a
consonantal element not unlike *h3, except that **R always vocalizes to *o (not
only in Greek):
**R. > *o (kR.m(n)ós > kormós)
**éR > *ó: (swéRp-eye- > swó:peye-)
**Ré > *ó (pRr(m)néh2 > pRér(m)neh2 (anaptyxis) > pórna:)
**iR > *i: (giR-genh1- > gi:genh1- (Skt. ji:jan-))
As to the semantic content of the *R, something which covers both its use in the
causative, in combination with a verb meaning "to do, make", and its use in the
formation of thematic derivatives might be found in an ancient prefixed copula:
*R(bher)-os "which IS carried / which IS carrying", *R(men) éieti "he makes him
BE thinking". At least as regards the causative, it is not uncommon for
causative affixes to include forms of the verb "to be" (e.g. the Georgian-Zan
causative affix *ren-, *rin-, containing *r- "to be", or the Basque causative
prefix e-ra-, which probably contains the root *da (> -ra-) "to be").
=======================
Miguel Carrasquer Vidal
mcv@...