Psych Verbs in Ergative Languages

Amiridze, Nino Nino.Amiridze at LET.UU.NL
Fri May 27 13:08:18 UTC 2005

The distribution of cases that Stephen Hewitt gives is correct for Georgian.
However, the dative-nominative marking is not the only pattern for psych
verbs in Georgian. There are two classes of experiencer verbs to
distinguish. While the dative-nominative marking is for the
subject-experiencer verbs (1), there is the so-called ergative-nominative
marking for object experiencer verbs (2). In that object-experiencer verbs
behave like transitive verbs (3).

    a. uq.vars ghvino          (TMA Series I)
       man-DAT wine.NOM
       "The man loves wine"

    b. uq.varda ghvino         (TMA Series II)
       man-DAT wine.NOM
       "The man loved wine"

    c. hq.varebia ghvino        (TMA Series III)
       man-DAT wine.NOM
      "The man has loved wine"

    a. abrazebs up.asuxismgebloba      (TMA Series I)
       man-DAT it.angers.him irresponsibility.NOM
       "[Someone else's] Being irresponsible angers the man"

    b. gaabraza up.asuxismgebloba-m      (TMA Series II)
       man-NOM it.angered.him irresponsibility-ERG
       "[Someone else's] Being irresponsible angered the man"

    c. gaubrazebia up.asuxismgebloba-s      (TMA Series III)
       man-NOM it.has.angered.him irresponsibility-DAT
       "[Someone else's] Being irresponsible has angered the man"

    a. k.lavs mezobel-i      (TMA Series I)
       man-DAT he.kills.him neighbour-NOM
       "The neighbour kills THE MAN"

    b. mezobel-ma      (TMA Series II)
       man-NOM he.killed.him neighbour-ERG
       "The neighbour killed THE MAN"

    c. mouk.lavs mezobel-s      (TMA Series III)
       man-NOM he.has.killed.him neighbour-DAT
       "The neighbour has killed THE MAN"

I think the term "ergative" as a case label is misleading for the Georgian
marker ma-/m- (3b/2b). The very marker in Georgian is used not only to mark
the subject of transitives (3b) and thus distinguish it from the rest of the
roles (as the object of transitives, the subject of unergatives and
unaccusatives), as in ergative languages


the subject of transitives (3b) plus the subject of unergatives (4). Note
that the subject of unaccusatives (5) and the object of transitives (
in 3b) are both marked alike, by NOM. This is the kind of marking that is
identified as active. I am afraid, the term "split-ergative" is vague for
Georgian and does not really help to know how much ergative (if at all) it
is or it is not.

(4) iqvira      (TMA Series II)
    man-ERG he.shouted
    "The man shouted"

(5) mokvda      (TMA Series II)
    man-NOM he.dies
    "The man died"

By the way, in the Georgian linguistic literature (as well as in some other
works like G.Klimov's Typology of Active Type, Moscow, Nauka, 1977,
A.C.Harris' Diachronic Syntax: The Kartvelian Case (Syntax and Semantics,
18). New York: Academic Press, 1985 among others) the case is not called
ergative (but narrative) and concequently there is no bias towards the
particular alignment model.

Nino Amiridze

-----Original Message-----
From: Hewitt, Stephen
Sent: 5/26/2005 7:55 AM
Subject: Re: Psych Verbs in Ergative Languages

Georgian has:

(1) split ergativity with transitives and unergative intransitives -
subject(-object) as follows:
        (a) nominative-dative[=accusative] marking in present-future
        (b) ergative-nominative marking in preterite tenses;
        (c) "inverted" dative-nominative marking in perfect[=evidential]
(2) no ergativity, nominative marking in all tenses, with passives and
unaccusative intransitives;
(3) "inverted" dative-nominative marking in all tenses, with experiencer
verbs of the sort you mention.


Steve Hewitt
s.hewitt at


Carol Rosen wrote :

> I have a question about psych verbs in languages with ergative
> morphology.
> English psych verbs, of course, vary a lot in how they treat the
> experiencers. Verbs like remember, forget, fear take the experiencer
> subject, while such verbs as annoy, bother, frighten seem to take the
> experiencer as direct object. In other languages the experiencer often

> appears as a dative.
> I hope to discover whether any one of these patterns tends to be
> preferred
> in languages with ergative morphology.
> I'm grateful not only for data, but also for references to appropriate

> sources. -- With thanks, Carol Rosen

More information about the Lingtyp mailing list