[tied] Morphology (6/20)

From: Miguel Carrasquer
Message: 14434
Date: 2002-08-21

6. The optative

The optative is represented far better in the attested IE languages than the
conjunctive. We find it preserved as such in Tocharian, Greek, Germanic (there
called the "conjunctive") and Indo-Iranian. In Latin, the conjunctive of the
verb "to be" is an old optative. In Balto-Slavic, the old optative survices as
the imperative. In Tocharian B, Armenian and Celtic, the optative is disguised
as an imperfect tense, and in Albanian it's at least the category of the
optative that survives (in what is etymologically an aorist conjunctive). Only
Anatolian seems to lack any evidence for the optative, but perhaps an
etymological connection exists with the Anatolian 1sg. imperative.

The optative comes in athematic and thematic variants. The optative suffix can
be reconstructed as *-yéh1-, giving the following reconstructed paradigms (root
always zero grade in the athematic forms):

athematic thematic
*-yéh1-m *-ih1-mé *-o-ih1-m *-o-ih1-me
*-yéh1-s *-ih1-té *-o-ih1-s *-o-ih1-te
*-yéh1-t *-ih1-é:r *-o-ih1-t *-o-ih1-ent

The forms:

Tocharian B:

-im -im(o)
-it -icer
-i -im.

Here the plural athematic form of the suffix *-i:- (< *-ih1-) was generalized at
the expense of sg. *-ye:- and thematic *-oi(h1)-. Verbs with root ending in -a:
have the suffix -oy-, which does not represent *-oih1-, but a Tocharian B
development of *-a:-i:-.
The Tocharian B optative is a true optative when added to the conjunctive stem,
but it serves as an imperfect when added to the present stem.


athematic thematic
-jau -eima -- -au -aima -aiwa
-eis -eiþ -eits -ais -aiþ -aits
-i -eina -ai -aina

In the athematic conjugation, the suffx -i:- has been generalized.

Armenian (imperfect)

"to be" thematic
ei eak` ber-ei ber-eak`
eir eik` ber-eir ber-eik`
êr ein ber-êr ber-ein

In the imperfect thematic, the forms of the verb "to be" are suffixed to the
verbal root. The forms of the verb "to be" itself can be explained as follows:

*h1es-yeh1-m > *ehye:n > *e(y)i:n > eí
*h1es-yeh1-s > *ehye:z > *e(y)i:r > eír
*h1es-yeh1-t, with metathesis:
*h1es-ye-th1 > *ehyeth, weakened to:
> *ehyedh > *ey&r > éyr > êr
*h1(e)s-ih1-tesw > *ehi:tesW > *ei:(y)@kW > eík`
*h1(e)s-ih1-(e)nt > *ehi:end > *ei:(y)@n > eín

The 1pl. form has the suffix -ak`, like all Armenian past tense forms (derived
from the middle preterite in -a:-). In the imperfect, we would have expected


Old Irish has an imperfect tense with the following endings:

-in(n) -^mis
-^tha -^the
-ed, -ad -^tis

In Brythonic, the imperfect endings are:

Mid.W. Mid.Br.
-wn -em -enn -emp
-ut -ewch -es -ech
-(e)i -ynt -e -ent

Now Welsh /u/ comes from *oi, which is an important clue. The Old Irish forms
are always conjunct, and are the same for the active and the deponent (middle),
which might suggest an origin in the PIE middle. Combining the two clues, we
arrive at a middle optative (thematic). The Old Irish forms are then
reconstructable as:

*-oih1-mh2ai > -(i)n -oih1-mosdho(i) > -(i)mis
*-oih1-th2a:s > -(i)tha -oih1-teso(i) (for *-dhwesoi) > -(i)the
*-oih1-to > -(i)d -oih1-nto > -(i)t, analog. -(i)tis after 1pl. -(i)mis

For the endings, see the chapter on the middle.


Only the conjunctive of the verb "to be" reflects the old optative:

sim < siem si:mus
si:s < sies si:tis
sit < siet sint

Baltic (imperative)

2sg. OPr. -ais
3sg. Lith -ai/-ie~
2pl. Latv. -ait/-iet

Lithuanian has innovated in the 2nd person sg. and pl. (-k[i], -kite[:]) and
Latvian uses the 2sg indicative as a 2sg imperative. Otherwise the Baltic
imperative forms, as in Slavic, are descended from the PIE thematic optative in

OCS (imperative):
1 -- -êmU -êvê
2 -i -ête -êta
3 -i -- --

This represents the thematic optative (with development -oih1 > -i:, -oih1- >
-e:-, as is regular in Slavic).


athematic thematic
-ie:n -îmen -oi-mi -oi-men
-ie:s -îte -oi-s -oi-te
-ie: -îen -oi -oi-en


athematic thematic
-yá:m -yá:ma -yá:va -eyam -ema -eva
-yá:s -yá:ta -yá:tam -es -eta -etam
-yá:t -yúr -yá:ta:m -et -eyur -eta:m

Here the sg. suffix -yéh1- has spread into the plural and dual, replacing
original -i:-.

Hittite has a 1st person imperative ("let me ...", "may I ..."), with endings
-lu(t), them. -allu(t), characterized by the imperative affix -u (as in 3sg./pl.
-tu, -antu). The verb "to be", however, has an irregular (and presumably older)
form <es-lit> "may I be", where *-lit may stand for *-lét. Considering that
outside Anatolian such a form would correspond to **-léh1-/*-lh1-, the only
thing that prevents us from equating this 1sg. imperative ending with the
optative is the problem of initial *l- vs. *y-. But alternation between *l- and
*y- is not unknown from a number of IE lexical items (*ye(:)kWrt ~ *le:pWrt
"liver", for instance), and may easily be explained (like the labialization we
have seen in a number of grammatical forms) as the result of the loss in pre-PIE
of **i and **u as full vowels (replaced by *^& and *w&, respectively). We can
perhaps think of a relation between the optative suffix *yéh1- (< *lyét < **lít)
and the verbal root *leh1(i)- "to want".

Miguel Carrasquer Vidal