Bengali Constructions
Edited by Gautam Sengupta and Lars Hellan
This is a discussion page around annotations for Bengali. It is accompanied by a Log page for the annotated example collection Bangla labeled constructions. The log page is Bangla constructions log, containing 26 annotated and commented examples using the Construction Labeling system (see Verbconstructions cross-linguistically - Introduction). It also relates to the page Multiverbs and Complex Predicates, which is an attempt at assembling and classifying cross-linguistically a variety of types falling into these categories; references made below to Ga and languages of the Volta Basin will be understood from that page.
Multiverb constructions (mvc) come in two varieties in Bangla:
- SVCs, with one important difference from SVCs in the Volta Basin: where in the latter all the verbs share tense and aspect (and count as ‘finite’), in the Bangla SVCs there is just one finite verb (the last one), and the others have a form called ‘conjunctive participle’.
- VVs – 2-verb sequences with one ‘pole’verb and one ‘vector’ verb, where the pole verb (the first verb) can be any transitive or intransitive verb, and the vector verb is one from a set of 15-20 verbs, in the VV connection determining aspect or orientation. The pole verb is in conjunctive participle form, the vector verb is finite.
Inside an SVC, a VV can take part counting as one ‘verb’.
In this respect, VVs relate to SVCs in a similar fashion as EVCs relate to SVCs in Ga.
So, schematically speaking, one can distinguish a ‘flush mvc’ from a ‘participle-finite mvc’, and an ‘indef-numbered mvc’ from a ‘restr-numbered mvc’:
Volta-BasinSVC: flush and indef-numbered
BanglaSVC: participle-finite and indef-numbered
EVC/ ‘essential SVC’ (Ga): flush and restr-numbered
VV (Bangla): participle-finite and restr-numbered
What is called ‘multinumbered’ is the temporally consecutive type of SVC. But in the SVC typology, there are also, e.g., benefactive SVCs and instrumental SVCs, which are restr-numbered, and in Ga, these are not EVCs – one may use Christaller’s term ‘essential’ SVCs for these. Such SVCs occur in Bangla too, but it may be debated whether they constitute a class distinct from VVs.
For ‘first slot’ labeling, vv, ev and sv are candidates, but ‘sv’ should probably be split into two, for the ‘flush’ and the ‘non-flush’ variant. We suggest using sv for the Volta-Basin type, and cv for the Bangla type, read as ‘conjunctive participle serial verb’, alternatively ‘converb sv’ (reflecting Ethiopian tradition).
Among issues concerning 'v-v-sequences' in the normal glossing are:
- how to gloss each of the verbs;
- how to gloss the 'participial' morpheme.
Consider this annotated token:
| āmi |
| āmi |
| I1SG |
| PN |
| paṛe | |
| poṛ | e |
| fall | CMPL |
| V | |
| gelām | |
| ge | lām |
| go | 1P |
| V2 | |
Here 'e' is glossed 'CMPL', for 'completed', which is motivated by the fact that this so-called 'participial' form expresses completedness.
As for the verbs as wholes, one tradition calls the 'free' verb (the first one) the POLE verb and the other verb the VECTOR verb. Should these notions be used here?
Less telling is using 'V1' and 'V2'. For the time being, they are labeled 'V' and 'V2'.
--Lars Hellan 10:38, 22 December 2013 (UTC)
Bangla Postpositions
The label "postposition"/"PPOST" in Bangla is used as a cover term subsuming three diferent kinds of words that share the functional load of prepositions in English-like languages. These are:
- words like পাশে /paSe/ that take nominal suffixes and govern the genitive case, as in
| nOYon |
| nOYon |
| Nayan |
| Np |
| amar | |
| ama | r |
| 1SG | GEN |
| PNposs | |
| paSe | |
| paS | e |
| side | LOC |
| N | |
| boSbe | |
| boS | be |
| sit | 3PFUT |
| V | |
| nOYon |
| nOYon |
| Nayan |
| Np |
| amar | |
| ama | r |
| 1SG | GEN |
| PNposs | |
| paSer | |
| paS | er |
| side | GEN |
| N | |
| ghOre | |
| ghOr | e |
| room | LOC |
| N | |
| boSbe | |
| boS | be |
| sit | 3PFUT |
| V | |
- words like দিয়ে /die/ "having given" that take the completive/perfective verbal suffix and govern the object case, as in
| ami |
| ami |
| 1SG |
| PN |
| oke | |
| o | ke |
| 3SG | OBJ |
| PN | |
| die | |
| di | e |
| give | CMPL |
| V | |
| kaj |
| kaj |
| work |
| N |
| kOracchi | |||
| kOr | a | cch | i |
| do | CAUS | CONT | 1PPRES |
| V | |||
- words like জন্য /jonno/ "for" which take neither nominal nor verbal suffixes but assign genitive case, as in
| ami |
| ami |
| 1SGNOM |
| PN |
| tomar | |
| toma | r |
| 2SG | GEN |
| PN | |
| jonno |
| jonno |
| for |
| PPOST |
| Opekkha |
| Opekkha |
| wait |
| N |
| korbo | |
| kor | bo |
| do | 1PFUT |
| V | |
--Gautam Sengupta 11:01, 11 December 2009 (UTC)
Multi-/Serial Verb Constructions
What are multi-/serial verb constructions? Are sentences like the following serial verb constructions?
Sally wants John to leave.
John wants PRO to leave.
Susan is easy to please.
Presumably, these are not mvc/svc's, at least they are not so-called in the literature. What is the justification hen of labeling the following as a serial verb construction
| ami |
| ami |
| 1SG |
| PN |
| baRi |
| baRi |
| home |
| N |
| gie | |
| gi | e |
| go | CMPL |
| V | |
| bhat |
| bhat |
| rice |
| N |
| khee | |
| khe | e |
| eat | CMPL |
| V | |
| ghumabo | ||
| ghuma | bo | |
| sleep | ||
| V | ||
which may be structurally identical to the second sentence in the previous paragraph?
ami [PRO baRi gie] [PRO bhat khee] ghumabo