nasal pres / root aor

[Jens Elmegaard Rasmussen]

On Tue, 3 Aug 1999, petegray wrote:

> On Strunk ... nasal presents & root aorists,

> I have now read Strunk. [...]
> We note that Rix (lexicon I-G Verbs) partially demurs, saying that the
> distinction of full grades I and II in nasal presents is "not sufficiently
> proved".

> Either way, the basic tenet of Strunk's argument at this point is now
> standard stuff.   The point of our discussion arises in his next argument.
> [..] He says (my
> trans.) "There *must* be a strong affinity between nasal present and root
> aorist."

> This is where my problems begin.   What does he mean by "affinity"?    And
> when you mention, Jens, a "paradigmatic companionship", what do you mean?
> Naively, but understandably,  I thought you and Strunk meant within a
> language, but you say:

>> It was never claimed for the individual languages, only for the common
>> reconstructed protolanguage.

> which is a good reflection of what Strunk says.

> That means you are arguing only that if a PIE language has a nasal present,
> there is likely to be a root aorist somewhere else in PIE.   Strunk gives 8
> major examples in his book - and in two of them he has to find the root
> aorist in a different language from the nasal present.

(J aside: Delete the P in PIE, the attested languages are not _P_IE.)

> Now my argument is that this is not really "paradigmatic companionship".
> I do not believe it is anything more than a restatement of what a nasal
> present is.  It seems to me that you and Strunk are merely saying:
> A PIE root can form:
> (a) presents of varying kinds in various LL, incl. perhaps a nasal present
> somewhere;
> (b) aorists of various kinds in various LL, incl often a root aorist
> somewhere.

> So we are not surprised that a good number of roots which show a nasal
> present somewhere also show a root aorist somewhere else.    The
> "paradigmatic companionship" is only a significant claim if it occurs within
> the same language, which - at least here - you deny.

> To be fair, we should look to see if there is a strong PIE correlation, and
> if there is a correlation within individual languages. [Reference to
next posting.]

I hope I'm allowed such extensive a quote from Peter's posting, seeing
that it makes the context very clear. I do indeed believe that the
existence of a nasal present in the IE protolanguage _normally_ implied
the existence of a root aorist in the same protolanguage. But that does
not necessary repeat itself in the daughter languages because of the many
changes that have occurred between PIE and any attested IE language.
Still, the impression is so clear that one is justified to believe that,
in PIE, it was clearer still. It may be like the strong verbs of Germanic:
you don't find many well-structured strong verbs in present-day Gmc. lgg.,
but you do find enough to put you on the track of something that seems to
have had more of a system to it once than it has now.

I have made some checkings into the tenacity of this companionship, and it
does actually look quite good. I think it is a fair statement that,
wherever the nasal is present-forming, there is a non-present form
differing only by lack of the nasal. Now, that is a root aorist (where it
is not the perfect).

There are at least three cases of strong support what were not used by
Strunk:

1. Tocharian: "Zu Ps. VI geho"ren (...) Ko. V und Pt. I", "Neben Ps. B VII
steht regelma"ssig Ko. V und Pt. I." (Krause & Thomas, Tocharisches
Elementarbuch I, 204 and 205). By P(r)s. VI is meant the formation with
-na:- (from -n at -, the weak form of the Indic class IX), while Prs. VII of
the B dialect refers to the nasal-infix structure matching Lat. vinco:,
Ved. yunakti etc. These both regularly take "Preterite I" which is the IE
root aorist. Here we have a language that synchronically presents the
paradigmatic companionship assumed by Strunk for PIE. Note that this is
not the IE perfect which is continued in two other Toch. categories, viz.
(i) in the subjunctive (!), (ii) in the Preterite III which also reflects
the s-aorist (due to the special Tocharian merger of the vowels *-o- of
the perfect and *-e:- of the s-aorist.

2. Baltic: Lithuanian (3rd person) prs. bunda, prt. budo (busti 'wake
up'), rinka riko (rikti 'make a mistake'), minga migo (migti 'fall
asleep'), tinka tiko (tikti 'touch, fit') and many others have an -n-
inserted in the present, while the prt. lacks it. The meaning is
practically always situation-changing, i.e. the natural domain of an
aorist, so we expect a root aorist to be formed without further ado. And
what could the preterite stems bud-, rik-, mig-, tik- etc. be if not the
weak form of root aorists? The meaning is mostly intransitive-reflexive
which makes it even better, for the added preterite morpheme *-a:- (whence
Lith. -o-) is opposed to a different preterite morpheme *-e:- (Lith. -e:-,
spelled with the dotted e) which is typically active and transitive. Now,
it is a widespread doctrine that thematic verbs had *-e-/-o- in the
active, but constant *-o- in the middle voice. Therefore, if the basis is
the 3sg in act. *-e-t, mid. *-o (or *-o-t), these would come out as Balt.
*-e and *-a resp.; and when _they_ go inflected by the addition of
thematic endings, the result would be act. *-e-e and mid. *-a-a which
would yield exactly Lith. -e: and -o as we find them to be. For the
addition of thematic "endings" cf. the fate of the root aorist in Slavic
3sg nes-e (Ved. a:nat. /-:nas'/ from *Hnek^-t 'reached').
   This means that the constant weak-grade form of the Baltic nasal infix
presents simply reflects the IE middle-voice stem which of course had weak
form all through. This is decisively supported by the semantics of nasal
verbs derived from adjectives which are, not factitive as in Hittite, but
ingressive: bukas 'blunt' forms bunka buko (bukti 'become blunt'), dubus
'deep' forms dumba dubo (dubti 'become hollow, concave'), pigus/-as
'cheap' forms pinga pigo (pigti 'get cheaper'). This can only be the
middle voice of a factitive: "be made cheaper" = "become cheaper".

3. Hittite. Though Anatolian does not distinguish different stems within a
given verbal lexeme, the nasal-infix stems are opposed to structures
without the nasal. And in Hittite there is a clearcut opposition of
function, the nasal structure being causative: hark-zi 'vanishes' :
hanik-zi 'destroys'. This is of course not the function of the Lith.
intransitives like minga 'falls asleep', but if we remember and respect
the probable origin of the Lith. structure in the middle voice, it's okay
again: then the middle voice of a causative "makes oneself fall asleep" or
"is being caused to fall asleep" will come full circle and end up meaning
'fall asleep' just as the base verb did in the first place. And, with an
adjectival basis, Hitt. tepu-s 'small, inferior' forms tep-n-u-zzi 'makes
inferior, humiliates' (which was a PIE lexeme, cf. Ved. dabhno'ti
'damages'). A corresponding Baltic verb would have been based on the
midle voice and have meant 'become inferior', as the midle voice
presumably meant already in PIE (or even earlier).

The productive derivative status of the nasal structure in Hittite is not
the whole story, however: There are also remains of lexemes that had
passed through the whole history reflected by th other languages, thus
notably tamekzi, tamenkanzi 'adhere, stick (vel.sim.)' from *tm.-ne-k-ti =
Skt. tana'kti 'run thick, coagulate' (root *temk- of Eng. tight) and
hamekzi, hamenkanzi/hamankanzi 'bind' from *H2m.-ne-g^h-ti (root of Lat.
ango:, Gk. a'nkho: 'tighten, narrow').

The full story of the nasal present must be something like this: The
formation was in origin _factitive_ "make (into) -", "cause to be -". But
since the verbal root was also an agent noun (vr.tra-ha'n- is a 'killer of
Vr.tra-', Lat. re:g- is a 'ruler'), the nasal structure made from root
nouns of agent-noun semantic created simple causatives: 'make a binder' =
'make bind'. Then, the middle voice of that 'be made a binder, be made
bind, be caused to bind' was simply an elaborate way of saying 'bind'.
Thereby the structure widely lost its specifically middle-voice semantics,
and so it is no great wonder if it turns up with active endings. The whole
scenario must - at least in large part - have been completed before the
working of the ablaut, for the new active forms have escaped the accent
shifts seen in the middle voice: *(H)yew-ne'-g-e 'is made join' replaced
the middle endings by active ones, 3sg inj. *(H)yew-ne'-g-t, while the
middle was restructured to *(H)yew-ne-g-to' with accent shift onto the
syllabic ending, and only then did the ablaut reductions make act.
*(H)yune'g-t, mid. *(H)yung-to' out of these forms. - Note that the
presumed earlier middle-voice preform of a nasal prs. like Vedic
s'r.n.o'ti 'hears' is indeed found in Old Irish ro-cluinethar 'hears'
which is a deponent verb.

Despite the retention of productivity in Hittite, the nasal present type
has plainly become grammaticalized as "just a present" with individual
verbs in PIE, and it is hard to escape the impression that the
corresponding aorist (when there was one) was the root-aorist type.

Jens