LISSIM 6

June 1-15, 2012@ Kangra

Selected Essays

Nominal Modification

Vineet Chaturvedi
PhD/ EFLU

Hindi, modulo scrambling, is a head-final language. is holds true in the domain of nominals as well. (1) gives a few examples.

 (1) a.    raam.kiiRam.gaaRii car

‘Ram’s car’

      b.   ũũcaa pahaaR

high mountain

‘A high mountain’

      c.   rElgaaRi

rail-vehicle

‘Railway(s)’

                  d.  [DP[raam.ke jaane.mẼ] kuch ] samay hE

Ram.gen going.loc some time is

            ‘There is some time in Ram’s leaving’

 The head in all three is final, (a) being a genitive DP, (b) a nominal modified by an adjective and (c) a N-N compound. Of course, (c) is not the only way compounds are formed in Hindi, as with other languages, and therefore is not a good or a particularly illustrative instance of compounding, but in relation to the notion of modification, the head is clearly identifiable here.[1] The last case is of a gerund, where too the head is final.

 The effect of embedding of nouns within a functional projection, the ‘DP hypothesis' of Abney (1987), results in any modification to a nominal to be eventually located within a DP.[2] The desired parallelism between CP and DP gives the DP a structure as rich and finely articulated as CP. The vaalaa construction in Hindi therefore, should also be an addition to the structural modification in DPs.

 2 The interesting data

 The vaalaa construction also appears to be a case of typical nominal modification like (1). The head indeed is always final in the straight forward case of a vaalaa DP, but atypically, it enables the modification of the head by either adjectives or nouns.

             (2)       baRii / raam vaalii gaaRii

                        big    / Ram  V       car

                        ‘Big / Ram’s car’

 raam vaali gaaRii would suggest that \textit{vaalaa} makes possible the modification of a nominal by another nominal, thus allowing the [N + vaalaa] complex to modify a nominal head. This function is described as one of the strategies of making an adjective out of nouns in Hindi in \cite{verma-1971}. The status of [A + vaalaa] is not clear in Verma's generalization.

 A verb modifying a noun  is made possible by vaalaa in (3). As the glosses indicate in (a), there are two possible readings: future time and descriptive marking. However, the structure is not as clear as with nominal or adjectival modifiers. The syntactic ‘head' for instance seems to be missing here.

             (3)a.    ye kapRe dhulne vaale h~E

             these clothes wash.inf V are

                         ‘These clothes are going to be washed' OR

                         ‘These clothes are for washing'

                 b.    yah sher gaajar khaane vaalaa hE

             this lion  carrot  eating    V         is

                         ‘This lion is [about to eat carrots]/[one who eats carrots]/[a      carrot eater]’

 In (a) the non-finite verb jaanaa, ‘to go' along with vaaalaa gives a future time reading. The (b) case is analysed  in Mohanan (1994) as three-way ambiguous –  an immediate future marker, a subject relative clause and an N + V compound.

 3 Why it is interesting

 There are two types of issues here that can be of interest. The first is the issue of an adequate empirical generalization and second is the theoretical implications of the issues of the first type.

 Emperical Issues Of the many strategies and associated structures of (nominal) modification, how do vaaalaa constructions compare? Nominal modification has been extensively explored and placing vaalaa amidst all of those is one of the issues to work on.

 But a preliminary brief description is that the morphology of vaalaa is consistent with the agreement facts of other functional elements in a DP. Further, unlike adjectives, only attributive modification seems to be made possible by vaalaa. The curious part of the modification mechanism comes from the different types of modifiers that can be c-selected: a full DP, certain types of adjectives and denominal verbs. Informally, the generalization seems to be that  vaalaa allows a lexical element to classify and thus restrict the reference of the head noun.  Why vaalaa is sensitive to a class of adjectives and thus disallowing them is not clear. When verbal elements are the modifiers, they are always nominalized. The fact that internal structure of verbs is preserved in the process allows for a temporal reading.  Lastly, the facts of vaalaa modification point to an analysis which does not involve a derivation similar to predicative adjectives — namely derivation from an equivalent relative clause. vaalaa modifier in a DP seem to be base generated always.

 In some recent work, Fredrike Moltmann pursues another path to nominal modifcation - that of tropes.  On a cursory examination, a characteristic of the vaalaa construction appears to be the trope (like) quality. A standard example in English of a tropes is the wisdom of Socrates. In Hindi, vaalaa and genitive DPs appear similar to tropes. The relevancy of examining Hindi DPs from the perspective of tropes is still unclear though.

 Theoretical Issues There are two explanations that can be given for a phenomenon like the two distinctive syntactic variations of the vaalaa construction.[3] One explanation is to place or locate the difference in the lexicon. The argument against doing so is that arbitrariness leaves little room for ‘explanation’ or generalization.

 The other explanation is to place the difference in syntax. The argument for locating variation in syntax broadly revolves around two concepts: the first is, what counts as a convincing explanation; the second is the notion of economy. These two arguments together presumably provide better predictability.

 Now a problem like vaalaa breaks some standard assumptions. When empirical data contradicts or does not match relevant licensing conditions, the abstractions associated with the explanation require amendment. That there is a mismatch with the vaalaa case is obvious. A question to explore is how the window provided by the mismatch yields insight (if any) into the language faculty, CHL.  Some issues that come up are described below.

 Current research on the syntax of DP advances the view that the structure of DP is finely articulated with different functional elements serving as heads of their own projections. Although there is variation among how certain issues are handled and consequently what projections are considered essential, there seems to be a consensus on how roughly the functional categories are sequenced in relation to each other. The noun is in the complement domain of the DP which serves as the functional category interfacing with external syntactic processes.[4]

 Further, the definitive feature of vaalaa modification is that it enables modification in an attributive position only. Given the fact that relativization process in Hindi always yields correlatives, subsuming vaalaa as a relative does not appear to be too convincing a solution in providing vaalaa some nomenclature. These two facts cast doubts on any explanation which seek to derive, in any significant manner, a vaalaa DP out of some other base structure. On the other hand, cases like (3) can be explained only by assuming some movement process, like raising, operating on a base DP.

 Speaking of semantics, a rough generalization seem to be that none of the modifications with vaalaa operate on the intension. Operating on the extension from the point of view of syntactic processes implies that the distinction in meaning is a consequence of the interaction of semantics with discourse. vaalaa provides discourse focusing of modifiers, in effect highlighting and providing context salience to a property identified by the modifying element. In other words, a formal statement of semantics has to be worked out and the characterization above validated.

 References

 Abney, Steven. 1987. The english noun phrase in its sentential aspect. Doctoral Dissertation,     MIT.

Mohanan, Tara. 1994. Argument structure in hindi. Stanford, CA: CSLI.

Verma, Manindra K. 1971. The structure of the noun phrase in english and hindi. Delhi: Motilal Banarsidass.

 

[1] Some syntactic jugglery is required to make Case marked nominals head final because Case markers always follow nominals. Head-finality here would presumably make (at least) any overtly Case marked nominal an instance of a Case head.

[2] Which makes all structures head-initial in the Generative tradition therefore.

[3] Namely: [[ n/adj + vaalaa] N} ] and [v + vaalaa].

[4] Although not necessarily acting as a ‘phase' that has been spelt-out.



For enquiries, write to secretary@fosssil.in

bullet Skype Interview
bullet Interview List
bullet Latest News
bullet LISSIM 6 Page
bullet Home