Mismatch 1: morphosyntax: subject ~ object
Mismatch 2: morphosyntax: direct object ~ indirect object
Mismatch 1
Inversion in Georgian is one of the most widely cited instances of a morphological mismatch. Inversion refers to the situation where verbal arguments which appear to be subjects (on the basis of various syntactic criteria) are morphologically encoded as indirect objects. The interpretation of inversion is complicated by the fact that it affects not just verb morphology, but also the case marking of the verb's arguments. Because of this, there is some disagreement about whether the mismatch is one between syntax and morphology (e.g. Anderson 1984) or between two different levels of syntax (Harris 1981).
Georgian verbal morphology is highly complex, and what follows is a sererely abbreviated account. There are three sets of person-number markers found on verbs (the v-set, m-set and h-set), which correspond to the three grammatical roles of subject, direct object and indirect object. The full set of forms is shown in the appendix. Conventially, the v-set markers are glossed as 'subject', the m-set as 'direct object' and the h-set as 'indirect object', as seen here:1
v-set
|
m-set
|
h-set
|
||
v-k’lav | m-k’lav-s | mi-h-q’vebi | ||
1.SBJ-kill | 1.OBJ-kill-3SG.SBJ | PRVB-3.IOBJ-follow | ||
'I kill X/them.' | 'X kills me.' | 'You'll follow X/them. |
However, the relationship of argument marking on the verb and grammatical role is more involved than that, and case government is not straightforward. Three patterns are found:
subject | direct object | indirect object | ||
pattern A | argument marking on verb | v-set | m-set | h-set |
case marking | nominative | dative | dative | |
pattern B | argument marking on verb | v-set | m-set | h-set |
case marking | ergative | nominative | dative | |
pattern C | argument marking on verb | h-set | v-set | none |
case marking | dative | nominative | (periphrastic) |
verb class
|
||||
class I | class II | class III | class IV | |
series I tenses (present etc.)2 |
A | A | A | C |
series II tenses (aorist etc.) |
B | A | B | C |
series III tenses (perfect etc.) |
C | A | C | C |
class IV verb 'love'
|
|||
mama-s | tav-is-i | švil-eb-i | u-qvar-s |
father-DAT | self-GEN-NOM | child-PL-NOM | corn-NOM |
'Father loves his children.' |
class I verb 'paint'
|
||||
v-u-xat’av | deda-s | surat-s | ||
1.SBJ-3.IOBJ-paint | mother-DAT | picture-DAT | series I | |
'I am painting a picture for mother.' | (no inversion) | |||
da-v-u-xat’e | deda-s | surat-i | ||
PRVB-1.SBJ-3.IOBJ-paint | mother-DAT | picture-NOM | series II | |
'I painted a picture for mother.' | (no inversion) | |||
ar | da-mi-xat’av-s | ded-is-tvis | surat-i | |
not | PRVB-1SG.IOBJ-paint-3SG.SBJ | mother-GEN-for | picture-NOM | series III |
'I didn't paint a picture for mother.' | (with inversion) | |||
(Anderson 1984: 165, 168, 175) |
Superficially, then, a construction with a verb under inversion looks just like a non-inversion verb with an indirect object: there is an h-series (indirect object) marker on the verb linked to a nominal in the dative case, and a v-series (subject) marker on the verb linked to a nominal in the nominative case. Justification for the claim that, under inversion, the argument encoded as a indirect object is really the subject comes primarily from two phenomena, reflexivization and number agreement.
a) Refelixivzation
Where there is no inversion, it is the nominal encoded as subject which is the antecedent to reflexivization. Under inversion, it is the nominal encoded as an indirect object:
Gela | i-rc’muneb-s | tav-is | tav-s | ||
G.[NOM] | 3.IOBJ-convince-3SG.SBJ | self-GEN | self-DAT | series I | |
'Gela is convincing himself.' | (no inversion) | ||||
Gela-s | turme | da-u-rc’munebi-a | tav-is-i | tav-i | |
G.-DAT | apparenlty | PRVB-3.IOBJ-convince-3SG.SBJ | self-GEN-NOM | self-NOM | series III |
'Apparently Gela has convinced himself.' | (with inversion) | ||||
(Harris 1981: 125) |
b) Number agreement
3rd person arguments mark plural number on the verb only if they are subjects. In non-inversion constructions, a 3rd plural subject is marked by the suffix -(a/e)n/-es/-nen (typically, only animates induce number agreement), while number is not distinguished for 3rd plural objects:
P’ropesor-eb-i | st’udent’-eb-s | c’ign-eb-s | ga-u-gzavni-an |
professor-PL-NOM | student-PL-DAT | book-PL-DAT | PRVB-3.IOBJ-send-3SG.SBJ |
The professors will send the students books.' | |||
(Aronson 1991: 273) |
Under inversion, a 3rd plural subject is marked by the default plural marker -t (found in various contexts; see appendix), while 3rd person objects are marked by 3rd person singular v-set (subject) markers; 3rd plural -(a/e)n/-es/-nen is not found under inversion.
Švil-eb-s | tav-is-i | mama | u-qvar-t |
children-PL-DAT | self-GEN-NOM | father-NOM | 3.IOBJ-love-PL |
'The children love their father.' | |||
(Anderson 1984: 188) |
Mama-s | tav-is-i | švil-eb-i | u-qvar-s/*u-kvar-en | |
father-DAT | self-GEN-nom | child-PL--NOM | 3.IOBJ-love-3SG.SBJ/*3.IOBJ-love-3PL.SBJ | |
'Father loves his children.' | ||||
(Anderson 1984: 188) |
Summary
An account of inversion in Georgian requires recognition of three layers of representation:
The question remains where syntax ends and where morphology begins. In Harris's (1981) analysis, inversion in layer #2 is a syntactic operation. On this view, syntax is multilayered syntactic representation, and morphology is a direct realization of this; there is no syntax-morphology mismatch. In Anderson's (1984) account, inversion in layer #2 is morphological operation. Syntactic representation is simple, and the complexity arises in its morphological realization; thus, on this view, there is a syntax-morphoogy mismatch.
Mismatch 2
The verb 'let go' is inflected for an indirect object. However, the nominal that the indirect object marker agrees with is inflected as direct object. This is evident in series II tenses, where the case marking of direct objects (nominative) and indirect objects (dative) is distinct.
Q’araul-ma | t’usag-i | ga-u-šv-a |
guard-ERG | prisoner-nom | PRVB-3.IOBJ-let.go-3SG.SBJ |
'The children love their father.' | ||
(Hewitt 1995: 551) |
Agreement markers on verbs.
v-set | m-set | h-set | |
1SG | v- | m- | m-, mi-, me- |
2SG | g- | g-, gi-, ge- | |
3SG | -s, -a, -o | Ø | h-, s-, u-, Ø |
1PL | v-...-t | gv- | gv-, gvi-, gve- |
2PL | ...-t | g-...-t | g-...-t, gi-...-t, ge-...-t |
3PL | -(a/e)n, -es, -nen | Ø | h-, s-, u-, Ø |
Notes
1 Both subject and object may be marked on the verb, but only one object at a time (direct or indirect) ever appears, since (i) a 3rd person direct object is not overtly marked, and (ii) in a sentence where one might expect an indirect object + 1st or 2nd person direct object, a construction is used in which the 1st/2nd person are not treated as arguments of the verb. They are treated as possessive reflexives, a phenomenon that Harris 1981 terms Object Camouflage (p. 31).
2 The basic tenses are the non-past (series I), aorist (series II) and perfect/evidential (series III). (Series I also consists of the imperfect/conditional and non-past subjunctive, series II also includes the aorist subjunctive, and series III includes the pluperfect.)
References
Anderson, Stephen R. 1984. On representations in morphology: case marking, agreement and inversion in Georgian. Natural Language and Linguistic Theory 2. 157-218.
Aronson, Howard. 1991. Modern Georgian. In: The indigenous languages of the Caucasus ; Volume 1: Kartvelian languages . Delmar: Caravan Books. 219-312.
Harris, Alice C. 1981. Georgian syntax: a study in relational grammar. Cambridge: Cambridge University Press.
Hewitt, B. George. 1995. Georgian: a structural reference grammar . Amsterdam/Philadelphia: John Benjamins.