LISSIM 6
June 1-15, 2012@ Kangra
Selected Essays
Ditransitive Structures
Sonali Raj
MA, 1st Year, DU
This essay aims to determine a tree diagram for ditransitive verbs in Hindi.
Some popular ditransitive verbs in Hindi include: sikhaana, dikhaana, dena, daalna, rakhna, ghataana, baantna, bichhaana etc.
Consider:
(1)
|
Subject Actor |
DO Theme |
IO Goal |
Verb Action |
|
Maine |
kitaab |
mez par |
rakhi |
|
Munna ne |
choklet |
daadi ko |
khilaayi |
|
1 |
2 |
3 |
4 |
In general,
1. In Hindi, such sentences can be scrambled anyhow. There are 24 (4x3x2) ways of scrambling sentences such as the examples above and all the ways result in grammatical usage, by default.
However, in sentences such as:
(2) maine ghoRe ko uski Sakl dikhaayi
I-erg horse Dat its face show.prf.fem.1.s
I showed the horse(i) its(i) face.
The possessive pronoun uski here requires an antecedent and therefore it is ungrammatical to have it preceding the NP it refers to, as in:
(3) *maine uski Sakl ghoRe ko dikhaayi
2.
+tra verbs in Hindi may select either only inanimate arguments, or they may select one or more animate arguments and this is semantically determined. For instance, the verbs ‘samjhaana’, ‘dikhaana’ etc. require an animate IO; unless a variation has been justified in context.
3.
When the IO is a lexical item (such as didi, Raju, paudha, etc.) it must precede a preposition (such as ‘se’, ‘ko’ etc.):
(4) a. Sanju ne baccon ko daant sunaayi
Sanj erg children Dat scolding hear-cause.prf.3.f.s
Sanju made the children listen to the scoldings
b. Sanju ne unhen daant sunaayi
c. *Sanju ne baccon daant sunaayi
A possible reason for this could be that an Accusative/Dative pronoun such as ‘unhen’ is already marked with case; while ‘bachchon’ is caseless per se and is assigned case by a head such as ‘ko’.
If this is true, then (4) should have a structure in which the IO is under a barrier and cannot take case from the other elements of the sentence. Further, our structure should be applicable not just to verbs like sunaana, khilaanaa, denaa, bichhaanaa; but also to the semantic equivalents of typical ditransitives in other languages – such as ‘to forgive’; ‘to blame’ in English.
Consider the sentences:
(5) Usne mujhe der se aane ke liye maaf kiyaa
s/he-erg I-Dat late Inst come.inf Gen take.CP forgive do.pst.3.m.s
S/he forgave me for coming late.
(6) Polis ne hame katl ka doSi thehraayaa
Police erg we.dat murder gen accused stand.prf.3.m.s
The police accused us of the murder.
Here, a helping verb and a main verb are used for the bold words in: ‘He forgave me for coming late’ and ‘The police blamed the murder on us’. In these Hindi sentences, the heads ‘maaf’ and ‘doSi’ require an actor, a benefactor, a theme and a helping verb:
(7) Usne mujhe der se aane ke liye maaf
kiya
Actor Benefactor Theme
MV HV
However, this criterion may be bent for semantic reasons:
(8) Usne hamaari bhul
maaf kii
s/he.erg we.gen
mistake forgive/ pardon do.prf.3.f.s
S/he pardoned
out mistake.
(It is understood that it is us who will be forgiven for our mistakes, i.e. nobody else will, unless it is explicitly stated, such as in: Unhone Sanju ki ma ko uske ladke ki bhul maaf ki ('S/he forgave Sanju's mother for her son's mistakes'). Therefore, sometimes the benefactor needs not be stated.)
Even English permits this: ‘They forgave his incessant bickering’.
Meanwhile, both the following are also grammatical:
(9) They forgave him his incessant bickering
(10) Usne hame hamaari bhul maaf kii
s/he.erg we.dat we.gen mistake pardon do.prf.3.f.s
S/he pardoned us for our mistakes.
Constituents like ‘maaf karna’ therefore select fixed types of arguments (mostly NPs, APs, CPs or PPs; and ‘maaf’, ‘doSi’ etc. select a VP as well: ‘calo maafi maango!’; ‘Wo nirdoS hai!’
Now consider the sentences:
(11) Didi ne/Sanju ne uski bevakufi maaf kii
Elder sister erg/ Sanju-erg s/he.gen foolishness pardon do.prf.3.f.s
Elder sister/ Sanju pardoned his/her follishness.
(12) *Didi ne/Sanju ne uski bevakufi maaf kiyaa
(13) Didi ne/Sanju ne uski bevakufi ko maaf kiyaa
The gender of the helping verb ‘karna’ in such sentences does not seem to be affected very much by the gender of the subject of the sentence. Moreover, if we want to use ‘kiya’, we must also say ‘uski bevakufi ko’ instead of just ‘uski bevakufi’. One reason for this could be that without ‘ko’ the helping word gets gender from ‘uski bevakufi’; and when ‘ko’ is appended to ‘uski bevakufi’, the argument does not assign gender to ‘kiya’. We could conclude that the DO gives gender to the helping verb in such cases.
This seems to be true of +tra verbs like ‘denaa’:
(14) Usne mujhe kancaa diyaa
s/he.erg I.Dat marbles give.prf.3.m.s
S/he gave me marbles
(15) *Usne mujhe kancha dii
(16) Usne mujh ko kancha diyaa
(17) Usne mujhe paisa diyaa
(18) *Usne mujhe paisa dii
And,
(19) Usne mujhe paise diye
(20) *Usne mujhe paise diyaa/dii
Thus, the verb must agree with the DO.
In the presence of ‘ko’, arguments cannot assign AGR to the verb and it is free to take gender from contextual factors; but when ‘ko’ is absent, then arguments assign AGR to the verb.
We have also observed that when the DO includes ‘ko’, the verb invariably takes masculine gender. This means that masculine is the default form of verbs, and when there is no ‘ko’ barrier, verbs may take the feminine form.
We do not know whether the ‘ko’ form is present in the default structure for ditransitives, though. If it is, then irrespective of whether it is overt or not, its restrictions (barrier character) will be evident; whereas if it isn’t present by default in the structure, it will never display barrier characteristics because its occurrence will be a surface phenomenon. One solution would be to posit two structures for ditransitives – one with ‘ko’; and one without. But this would be an inelegant solution to the problem.
We know that in the folllowing:
(21)
Maine kitaab (ko) mez par rakhi (rakha)
I-erg book (Dat) table on put.prf.3.f.s (3.f.s)
I book put the book on the table.
Both the IO and DO are complements of the verb. When the sentences below are said in isolation, they both lack some information:
(22) a. Maine Sanju ko dikhaaya
I-erg Sanju Dat show.cause.3.m.s
b. Maine uskaa huliya dikhaaya
I-erg his/her face show-case.3.m.s
Also, the second of these makes less sense when said in isolation than the first does. This means that in (22b), both arguments are necessary and the DO is more important than the IO. For this type of construction, with multiple selected arguments in a hierarchy of importance, the most viable structure that has been proposed takes the DO as SPEC of the VP[1], as in:
For ‘ko’ constructions this structure could be:
However, this structure is dissatisfying because we are
left with the following questions unanswered:
- Whether it can accommodate other types of sentences in which a word selects multiple arguments.
- Whether it is applicable across languages and,
- Whether it is valid in the current syntactic framework
I regret I am unable to provide a conclusive solution with my current knowledge (and within this word limit!) but were I to be part of LISSIM 6, I’d find the solution with the faculty’s help for they have worked on this very problem.
Bibliography:
- Introduction to Government and Binding Theory, Liliane Haegeman, Basil Blackwell, 1991
- Asymmetry of Objects and the Larsonian VP-shell for Hindi/ Urdu, Shiti Malhotra, University of Delhi, 2005
[1] Asymmetry of Objects and the Larsonian VP-shell for Hindi/ Urdu, Shiti Malhotra, University of Delhi, 2005
For enquiries, write to secretary@fosssil.in
