FOCUS AND NOMINAL ELLIPSIS IN THE BANGLA DP* AMBALIKA GUHA The English and Foreign Languages University, Hyderabad 1xxIntroduction The paper addresses the leftward movement of adjectives, with or without the raising of NP, in the Bangla DP. This leads to the proposal that there is a second focus position, below demonstrative and above numeral-classifier, in the Bangla DP. I provide two empirical evidences that support the claim for a second focus phrase: one is exhaustive identification and the other is nominal ellipsis. The paper is organized into two broad sections. The first section summarizes Cinque’s (2005) DP-internal ‘parameters of movement.’ Then, I discuss the various semantic interpretations of phrasal movement inside the Bangla DP, as noted in the Bangla DP literature (Bhattacharya 1999, Chacón 2011, Dayal 2012, Biswas 2012, Syed 2016). In that section, I also show that the Bangla DP-internal word orders are derived following Cinque’s DP-internal ‘parameters of movement.’ The second section of the paper introduces certain instances of the raising of adjective, leaving the NP in its merge position, in the Bangla DP. This raising of adjective shows that there are two focus positions: one above demonstrative (as shown by Syed 2014), and the other below demonstrative and above numeral-classifier. The focus positions in the Bangla DP are being argued in the background of cross-linguistic evidence of focus positions in DP languages, like Gungbe, Albanian, Russian, and Greek. 2xxDP-internal ‘Parameters of Movement’: Cinque (2005) Revisiting Greenberg’s (1963) Universal 20, Cinque (2005) proposes that the various ‘attested’ orders of the four elements Dem(onstrative), Num(eral), A(djective), and N(oun) are derived * I would like to thank Prof. R. Amritavalli for her insightful comments and input. I would also like to thank the participants of the conference GLOW in ASIA XI. Special thanks to Prof. Roberto Zamparelli, Prof. Caroline Heycock, and Prof. Edith Aldridge for the productive discussions. Also, thanks to the four anonymous reviewers for their helpful comments. Copyright 2017 the author(s). In Proceedings of GLOW in Asia XI, volume 2, edited by Michael Yoshitaka Erlewine. MIT Working Papers in Linguistics #85. Cambridge, MA: MIT Working Papers in Linguistics. 73 74 Guha from the universal order of merge Dem>Num>A>N. He suggests that when none of the four elements (Dem, Num, A, N) moves, the merge order surfaces and the varied orders of these elements are derived only by the raising of the NP or of an XP containing the NP. In order to account for the ‘attested’ and ‘unattested’ orders of the above mentioned four elements, Cinque (2005: 321) set down the following ‘parameters of movement.’ (1) Merge order: [WP Dem [XP Num [YP A [NP N ]]]] (2) Parameters of movement: a. No movement (unmarked) b. Movement of the NP without pied-piping (marked) c. Movement of the NP plus pied-piping of the whose picture type (unmarked) d. Movement of the NP plus pied-piping of the picture of who type (more marked) e. Total unmarked versus partial marked movement of the NP with or without piedpiping (in other words, NP raises all the way up, or just partially around its modifiers) f. Neither head movement nor movement of a phrase not containing the (overt) NP is possible (except perhaps for focus related movements of phrases to a DP initial position). If NP raises alone (option (2b)) from specifier to specifier of each agreement projection found right above the functional projections hosting adjective, numeral, and demonstrative, the order that is derived is N>Dem>Num>A (as in (3)). (3) Movement of the NP without pied-piping In option (2c), the NP first moves to AgrYP, ‘vacuously’ pied-piping the adjective in a whose picture type way. Then, the AgrYP moves to AgrXP. Finally the AgrXP (containing the moved AgrYP and XP) moves to AgrWP, that gives rise to the order N>A>Num>Dem (as in (4)). (4) Movement of the NP with whose picture type pied-piping The option (2d) involves the “partial (marked) raising of NP plus pied-piping of the picture of who type of [A N] (marked) around Num” (Cinque 2005: 323; (6n)) and then the [A N Num] raises around Dem to give the order A>N>Num>Dem (as can be seen in (5)). Here, I assume that NP first moves to AgrYP, pied-piping the adjective in a picture of who way and that this movement is string vacuous. (5) Movement of the NP with picture of who type pied-piping In addition to this, languages vary in whether NP moves all the way up, or only partially around its modifiers (either without pied-piping or with pied-piping of picture of who type), as mentioned in (2e). Focus and Nominal Ellipsis in the Bangla DP 75 2.1xxThe Bangla DP Bangla is a numeral-classifier language of the Eastern Indo-Aryan language family. Its DP merge order (cf. (6)) conforms to the universal order of merge, i.e., Dem>Num>A>N. (6) Dem Num -Cla A ei du -To lal this two -Cla red ‘these two red dresses’ N jama dress As discussed in the Bangla DP literature (Bhattacharya 1999, Chacón 2011, Dayal 2012, Biswas 2012, Syed 2016), the merge order in (6) gives an indefinite interpretation. There is an alternative order (cf. (7)) in the Bangla DP, i.e., Dem>A>N>Num, which Cinque (2005: 320; n.13) also notes. The order in (7) is reported to have a specific reading (Bhattacharya 1999), or a definite reading (Chacón 2011, Dayal 2012, Biswas 2012), or an inclusive reading (Syed 2016). (7) Dem A N Num -Cla ei lal jama du -To this red dress two -Cla ‘these two red dresses’ To account for the order in (7), Bhattacharya proposes that there is a specificity feature in the Num0 which gets realized in the morphologically present classifier. He suggests that the specificity feature on the Cla0 attracts the [A N] sequence to the Spec, NumP in order to give a specific reading. Thus, Bhattacharya (1999: 95, 96) shows that the morphological absence of classifier and the absence of the raising of NP are correlated (cf. (8) and (9)). (8) a. b. (9) a. b. tin bOchor three year ‘three years’ * bOchor tin year three Lit. ‘years three’ car paS four side ‘four sides’ * paS car side four Lit. ‘sides four’ However, Chacón (2011) argues that the occurrence of classifier in an indefinite DP (as in (6)) suggests that the classifier cannot be the attractor of NP in definite DPs (as in (7)). Thus, contra Bhattacharya’s analysis for the order in (7), Chacón proposes that [A N] moves above Num-Cla in (7) in order to check definiteness and not specificity. He claims that a nominalizer n merges with the lexical nominal root to form nP and suggests that the nP (and not the nominal root) moves to Spec, DP to check definiteness. His analysis for the ungrammaticality of (8b) and 76 Guha (9b) is that nominalizer n does not merge with certain nominal roots (like measure nouns), and thus there is no nP which can move to check definiteness. Biswas (2012) states that the Dem>A>N>Num-Cla order gives a definite reading and she shows that the definite reading arises from anaphoricity. Dayal (2012), also, shows that the raised [A N] order above Num-Cla gives a definite interpretation and not the specific one. In fact, she states that the Num-Cla>A>N order is ambiguous between regular indefinite and specific indefinite reading. Dayal (2012: 23) notes that there are three possibilities in which the definiteness in the Bangla DP is achieved: “One, through the lexical meaning of the demonstrative taking the predicative cardinality/classifier phrase as its argument, or through NP raising to spec of DP to value the +def feature on D and undergoing iota type shift, or through a combination of the two.” Syed (2016) addresses a puzzle that shows that definiteness can be achieved without the [A N] movement in the overt presence of demonstrative. Thus, according to his argument both the orders in (6) and (7) should have a definite reading. This has already been observed in Dayal (2012: 14), where she states that for definiteness the non-preposed version of [A N] is possible when the demonstrative is present, but not when there is no demonstrative. As per her analysis it is maximality (Dayal associates maximality with definite descriptions) that drives the movement of [A N] around Num-Cla in the presence of demonstrative. Dayal (2012: 14) points out “The raised version is only possible when the NP refers to the full set of entities that the description applies to. In contrast, the base structure can be used to pick out a subset of a larger group of entities to which the description applies.” Syed, following Lyons (1999), splits definiteness into two concepts: identifiability and inclusiveness. Syed (2016: 391) explains inclusiveness in reference to Hawkins (1978) “the referent of definite noun phrase must be a part of a shared set, where shared set means entities known by speaker and hearer to constitute either the previous discourse, the immediate or the larger situation, or an association set.” He suggests that it is the inclusiveness (Syed (2016: 391; fn.2) relates inclusiveness to maximality) that attracts the [A N] above Num-Cla in the presence of demonstrative (as in (7)). He posits an Inclu(siveness) Phrase above NumP where the raised [A N] lands in order to check the inclusive feature of the Inclu0. Based on the above set of arguments (by Dayal 2012 and Syed 2016), it can be suggested that it is the maximality which drives the [A N] above Num-Cla and below Dem. It is important to note that the various arguments for the order in (7) are ruled by a common factor, i.e., the raised [A N] refers to the entity that has a prior discourse reference. I suggest that the raising of [A N] above Num-Cla and below Dem is a topicalised movement. Definiteness, inclusiveness and topicalisation all of these signify the shared set of knowledge between the discourse participants. It has also been claimed in É. Kiss (2007: 70-71) that “names, definite noun phrases, and specific indefinite noun phrases (or PP subsuming such a noun phrase) are all possible topics, irrespective of their subject, object, or prepositional object status.” Thus, I argue that the raised [A N] in the Dem>A>N>Num-Cla order moves to a topic position below demonstrative and above numeral-classifier, and not to the Spec, NumP, or Spec, DP, or Spec, IncluP. I suppose that the mechanism behind the raising of [A N] around Num-Cla follows from Cinque’s (2005) analysis for the derivation of the order Dem>A>N>Num. He suggests that the order Dem>A>N>Num is derived by the partial (marked) movement of NP plus pied-piping the AP in a picture of who type way (marked) around Num-Cla. Here, I will mention that in the Bangla DP the only way the NP can move, when the adjective is present, is by pied-piping the Focus and Nominal Ellipsis in the Bangla DP 77 adjective. When there is no overt adjective, the NP can raise to the left of Num-Cla (cf. (10)). But in the presence of adjective, the NP cannot move, leaving the AP stranded (cf. (11)). (10) (11) ei jamai du this dress two ‘these two dresses’ * ei this Lit. -To Cla ti du -To jamai Sundor two Cla dress beautiful ‘these two dresses beautiful’ ti The ungrammaticality of (11) suggests that the Bangla DP does not allow the raising of the NP without pied-piping. Hence, the whose picture type pied-piping is blocked in the Bangla DP as for this type of pied-piping, the NP has to first move to the left of AP, which is not possible (as shown in (11)). In the Bangla DP the only way NP can raise is by obligatorily pied-piping the AP in a picture of who way (as in (7)). Bhattacharya (1999), on the other hand, argues that the raising of [A N] is an instance of only NP movement. He shows that [A N] sequence behaves as an ‘independent unit’ and cannot be separated, as can be understood from the ungrammaticality of (12) and (13). (12) (13) Num * tin three Lit. Dem * ei this Lit. -Cla A Dem N -Te Sobuj ei boi -Cla green this book ‘three green these books’ A Num -Cla N lal du -To boi red two -Cla book ‘these red two books’ (Bhattacharya 1999: 56) (Bhattacharya 1999: 10) Following (12) and (13), Bhattacharya claims that in the Bangla DP the adjective is base generated at the Spec, NP and any leftward movement of the adjective, i.e., the extraction of the specifier of NP, is not allowed. I suggest that there is an alternative explanation for the ungrammaticality of (12) and (13). As per Cinque’s (2005: 323; 6q, 6n) analysis the orders in (12) and (13) cannot be derived as the NP has not moved and the modifiers to its left are in the wrong merge order. However, in this paper we will witness certain instances which will show the raising of the AP without the raising of the NP in the Bangla DP. In a way, that will support the condition (2f) of Cinque’s ‘parameters of movement.’ 3xxFocus Movement in the Bangla DP As Cinque (2005) observes, an exception to the condition for the DP-internal movement indicates that the phrasal movement, without the raising of NP, to the DP initial position (for focus purpose) is possible. Syed (2014: 5), also, shows a similar exception in the Bangla DP. He proposes a word order (cf. (14)) where the [A N] raises to the left of Num-Cla and below Dem, and then the adjective moves above Dem from the raised [A N]. Syed argues that the movement 78 Guha of the extracted adjective above demonstrative is an instance of focus movement, as evident from the ungrammaticality of (15). (14) (15) [LALj ei [tj boi]i -Ta red this book -Cla ‘This red book is of my liking.’ *[lalj ei [tj boi]i -Ta red this book -Cla ti] amar my pochondo liking ti] amar my pochondo liking It is not the case that the adjective can only be extracted when the [A N] has raised to the left of numeral-classifier and below demonstrative. Syed (2015: 337) shows that the adjective can also move all the way up, i.e., above demonstrative, leaving the NP in its merge position (cf. (16)). (16) [joghonyok oi du -To tk biskut] kheye, disgusting those two -Cla biscuit eat-PART sorir kharap lag.te laglo body bad feel.Inf start.Pst.3 ‘I started feeling sick eating those disgusting two biscuits.’ ama.r I.GEN Thus, what we can follow from Syed’s argument is that in the Bangla DP the adjective can move to the focus position above demonstrative in either of the two ways: first, by the raising of the NP, pied-piping the adjective in a picture of who type manner above numeral-classifier, following which the adjective moves above demonstrative out of the raised [A N]; second, by the raising of the adjective from its merge position without the movement of the NP. 3.1xxA Post-demonstrative and Pre-numeral Position of Adjectives In this paper I propose that there exists a second focus position in the Bangla DP, which is below demonstrative and above numeral-classifier. I suggest that the adjective, when bears focus, can also move to the second focus position, leaving the NP in its merge position (as shown in (17). (17) Dem A Num-Cla ei NEELi du -to this BLUE two-Cla ‘these two blue pens’ ti N kalam pen Cinque (2005: 315; fn.2) states that in some languages where the alternative order Dem>A>Num>N is possible, the adjective is not an attributive one (which occurs below numeral), but the source of it is from relative clause (which is above numeral). However, without getting into detail of the two sources of adjectives, I suggest that the occurrence of adjective in a non-merge position (in (17)) is focus-driven as it is evident from (18), which shows that the adjective in a non-merge position is not allowed if it does not bear stress. (18) * ei this neeli blue du -to two-Cla ti kalam pen Focus and Nominal Ellipsis in the Bangla DP 79 I should mention here that Syed (2015: 334) also seems to have noticed a “pre-numeral” position of a focused adjective, as in (19). (19) [LAL ek -Ta bari] dekhlam red one –Cl house saw ‘I saw a RED house.’ We can observe that in (19) there is no demonstrative. In fact a demonstrative cannot cooccur with ‘ek’ (one), as shown in (20), because ‘ek’ behaves as an indefinite determiner. (20) * [ei this LAL red ek-ta one-Cla bari] dekhlam house saw Thus, it is obvious from (20) that Syed’s ‘pre-numeral’ occurrence of adjective in (19) is not below demonstrative. However, in the Bangla DP the adjective can move above numeralclassifier and below demonstrative only with numerals which are higher than ‘ek’ (one). This has already been shown in (17) and repeated below in (21). (21) ei NEELi du -to this BLUE two-Cla ‘these two blue pens’ ti kalam pen Let us call the post-demonstrative and pre-numeral position of adjective Focus2 and the one above demonstrative Focus1. (22) Focus1 position (23) Focus2 position 3.1.1xxCross-linguistic Evidence of the DP-internal Focus Positions Drawing parallelism to Rizzi’s (1997) clausal left periphery, Aboh (2004) proposes a nominal left periphery in Gungbe. He suggests that the Gungbe DP involves both topic and focus phrases, whose heads are morphologically realized. Interestingly, in light of my proposal, i.e., a focus position below demonstrative and above numeral-classifier, Aboh (2004: 4; (6)) posits the topic phrase and the focus phrase between the DP and the NumP in Gungbe (as shown in (24)). (24) [DP… [D…topic… focus [NumP …[Num… [FP…N…]]]]] Guisti (1996) also argues for DP-internal focus and topic phrases in Albanian and Russian. In Albanian the adjectives are postnominal (25), and Guisti shows that the order of postnominal adjectives in Albanian is fixed. Thus, the inverted order of adjectives in (26) is not acceptable. 80 Guha The A(djective) O(rdering) R(estriction) in Albanian follows from Cinque’s (1994)1 hierarchy of adjectives. (25) (26) Njё grua tjetёr e a woman other the ‘another nice woman’ * Njё a grua woman e the bukur nice (Guisti 1996: 111; (14a)) bukur tjetёr nice other (Guisti 1996: 112; (16a)) However, it is possible to have a prenominal adjective in Albanian, where the adjective bears stress but not necessarily contrastive (as in (27)). Guisti (1996: 113, 114) shows that the prenominal position is available for both the adjectives, irrespective of their hierarchy (as in (28) and (29)). (27) (28) (29) Njё e bukur grua a the nice woman ‘a nice woman’ tjetra grua e other-the woman the ‘another nice woman’ e bukura grua the nice-the woman (Guisti 1996: 113; (20b)) bukur nice tjetёr other Guisti (1996: 115, 116), also, shows that in Russian it is possible to find different orders of adjectives only if the preposed adjectives are emphasized (as shown in (30), (31), and (32)). (30) (31) (32) eta staraja amerikanskaja knjiga o linguistike this old american book on linguistics ‘this old American book on linguistics’ * eta amerikanskaja staraja knjiga o linguistike this american old book on linguistics Lit. ‘this American old book on linguistics’ eta amerikanskaja staraja knjiga o linguistike this american old book on linguistics Lit. ‘this American old book on linguistics’ A similar kind of AOR violation can be observed in Bangla, which I will show in the subsection (3.1.1.1). Guisti (1996: 114, 116), further, notices that in the Albanian and Russian DPs, the focused adjectives must follow and cannot precede the demonstrative (as in (33) and (34)). (33) 1 *e the bukur(a) nice (-the) kejo this grua woman Cinque (1994) suggests that every adjective occupies the specifier position of different functional projections and the hierarchical organisation of the functional projections follows from the universal adjective ordering restriction. He suggests the following ordering restriction: Possessive >Speaker-oriented >Subject-oriented >Manner/Thematic. Focus and Nominal Ellipsis in the Bangla DP (34) * amerikanskaja eta american this 81 staraja knjiga o linguistike old book on linguistics The above discussion shows that in languages like Russian, Gungbe and Albanian, there is a DP-internal focus position, which is below DP. Correspondingly, we have observed that in the Bangla DP there is a second focus position which is below demonstrative. Also, we have noticed that there is a focus position above demonstrative in the Bangla DP, which is not possible in other DP languages, like Albanian and Russian. 3.1.1.1xxAOR (Adjective Ordering Restriction) Violation and Exhaustive Identification In Bangla the ordering restriction of adjectives can be disturbed only when the adjectives are contrastively focused (CF) and not when they bear information focus (IF). I suggest that the adjective, when bears IF, stays in situ and when it bears CF it necessarily moves either to the Focus2 or Focus1 position. This can be understood in respect to the difference between two kinds of focus: IF and CF. There is a considerable literature (Rooth 1992, É. Kiss 1998, Horvath 2010, among others) that discusses the difference between the two types of focus, stating that IF and CF occur in different kinds of discourse contexts and exhibit distinct semantic and syntactic properties. Horvath (2010: 1350) states that “(F)ocus is taken to be the non-presupposed, new information part of the sentence, i.e., information not shared by the speaker and the hearer at the point in the discourse where the sentence is uttered. This notion of focus is often referred to as information focus.” Whereas contrastive focus “represents a subset of the set of contextually or situationally given elements for which the predicate phrase can potentially hold; it is identified as the exhaustive subset of this set for which the predicate phrase actually holds” (É. Kiss 1998: 249). É. Kiss, based on Hungarian and English data, shows that the element bearing CF moves to a designated A’-position and the element expressing IF always remains in situ. É. Kiss, further, defines CF as expressing exhaustive identification (EI) and IF as non-exhaustive identification. EI means that by selecting some members of the set, all other members are excluded. Bangla also shows the distinction between the two types of foci. Here I will claim that the adjective in the Bangla DP obligatorily violates the AOR in order to express exhaustive identification. Thus, the data in (35a) and (36a) where the adjective has moved to the Focus2 and Focus1 position respectively, and not the one in (37), give CF reading. In order to verify that the moved adjective in (35a) and (36a) bear CF or express EI, we can apply a diagnosis (earlier proposed by Donka Farkas) that is mentioned in É. Kiss (1998). Let us have a situation, where only red tables (and not other colored ones) of small size belong to X. Now, if we negate the sentences (35a) and (36a) in (35b) and (36b) respectively, we will see that the negation holds. But the negation of the sentence (37a), where the adjective stays in situ, as in (37b), is not possible. (35) a. X: [ei LALj [choto tj tebil]i -gulo ti] this red small table -Cla ‘These small red tables are mine.’ b. amar mine na, [ei NEELj [choto tj tebil]i -gulo ti no, this blue small table -Cla ‘No, these small blue tables are also yours.’ o] also tomar your 82 Guha (36) (37) a. X: [LALj ei [choto tj tebil]i -gulo red this small table -Cla ‘These small red tables are mine.’ b. na, [NEELj ei [choto tj no blue this small ‘No, these small blue tables are also yours.’ a. X: [ei [choto lal tebil]i -gulo ti] this small red table -Cla ‘These small red tables are mine.’ b. # na, [ei [choto neel tebil]i -gulo no, this small blue table -Cla ‘No, these small blue tables are also yours.’ ti] amar mine tebil]i -gulo ti o] tomar table -Cla also your amar mine ti o] also tomar your In (35a) and (36a) the moved adjective ‘lal’ (red) represents that it is only the red colored (and not any other colored) tables that belong to X. In (37a) where ‘lal’ is focused in-situ, indicates that there is a possibility that apart from red tables other colored tables also belong to X. Thus, the negation of (37a) in (37b) is not possible. Whereas, the negation of (35a) in (35b) and (36a) in (36b) is felicitous as they follow the sentences that assert that it is only the red tables that belong to X. The discussion so far shows that there are two focus positions in the Bangla DP, where the adjectives move to when they bear contrastive focus or exhaustively identified. The last subsection of this paper provides further evidence for the existence of second focus phrase in the Bangla nominal domain. 3.1.1.1.1xxNominal Ellipsis in the Bangla DP Let us, first, observe the following set of data and gradually I will lay out my argument which is supported by these data. (38) (39) (40) (41) (42) (43) [kon which tin -Te chele] three -Cla boy ‘Which three boys came?’[ei this three -Cla tall boy ‘These three tall boys came.’ * [ei tin -Te lamba chele] this three -Cla tall boys Lit. ‘These three tall came’ [ei lamba chele tin -Te] this tall boy three -Cla Lit. ‘These tall boys three came.’ [ei LAMBA chele tin this tall boy three Lit. ‘These tall three came.’ * [ei lamba chele tin -Te] this tall boy three -Cla eSechilo? came tin -Te came lamba chele] eSechilo eSechilo came eSechilo came -Te] -Cla eSechilo came eSechilo came Focus and Nominal Ellipsis in the Bangla DP 83 There are three things that we can notice from the above set of data. First, the NP cannot be elided in its merge position (as shown in (40)). Second, the NP can and has been elided (in (42)). Third, there are two conditions for the elision of NP that can be seen in (42). One, the NP has to move from its merge position and second, the adjective has to be focused. If we compare (41), where the NP has not been elided, and (42), we can observe that in (42) the elided NP has moved below demonstrative and above numeral-classifier, pied-piping the AP in a picture of who type way. Again, if we compare (42) and (43) we can see that the AP has to be focused for the NP to be elided. To explain the pattern of the above set of data, I will follow Ntelitheos’ (2004) and Cinque’s (2012) argument for nominal ellipsis. Ntelitheos, based on Greek data, proposes that DP-internal nominal ellipsis involves two movement operations: NP topicalisation and modifier focalisation in the nominal left periphery. Following Johnson (2001) who proposes that VP ellipsis is licensed by VP topicalisation, Ntelitheos suggests that nominal ellipsis is also preceded by nominal topicalisation in the Greek DP. He states that this NP topicalisation is a discourse driven syntactic process that involves movement of the NP to the nominal left-peripheral position, which he calls TopP. He further mentions that NP topicalisation is followed by the movement of the remnant modifier into a focus position above the topic position where the NP has moved to. He argues that this focus movement of the remnant modifier licenses the phonological deletion of the NP. Ntelitheos (2004: 14) illustrates his argument through the following set of data (44a– 44e). (44) a. o Giannis agorase tria vivlia kai o Petros the Giannis bought-3SG three books and the Petros agorase ena vivlio bought one book ‘John bought three books and Petros bought one book.’ b. [XP [TopP [FocP [TopP [DefP… [FP ena … [NP vivlio]]]]]]] c. [XP [TopP [FocP [TopP [NP vivlio] [DefP… [FP ena …tNP]]]]]] d. [XP [TopP [FocP [FP ena …tNP] [TopP [NP vivlio] [DefP… tFP]]]]] e. [XP [TopP [FocP [FP ena …tNP] [TopP [NP vivlio] [DefP… tFP]]]]] In (44c) the NP ‘vivlio’ has moved to the topic position as it has already been mentioned in the discourse in (44a). In (44d) the modifier ‘ena’ which occupies the specifier position of some functional projection FP, pied-pipes the trace of the NP and moves to the focus projection above the topic position where the NP has moved to. The movement of the remnant modifier to the focus position licenses the phonological deletion of the NP (as can be seen in (44e)). Following Ntelitheos, I suggest that nominal ellipsis in the Bangla DP also involves two movement operations. First, the NP moves, pied-piping the AP in a picture of who type way to a topic position above Num-Cla and below Dem (as in (45)). I have already mentioned in Section (2.1) that I consider NP movement to the left of Num-Cla a topicalised movement as the entity the moved NP refers to has a prior discourse reference. Second, the adjective moves out of the raised [A N] to the Focus2 position which is below demonstrative and thus licenses the deletion of the NP (as shown in (46)). (45) [DemP ei[TopP [lamba chele]i [NumP tin this tall boy three -te -Cla [ti]]]] eSechilo came 84 Guha ‘These three tall boys came.’ (46) [DemP ei [FocP LAMBAj [Topic [tj chele]i [NumP tin -Te [ti]]]]] eSechilo this tall boy three -Cla came ‘These three tall ones came.’ From the set of data in (40)-(42) and their structural analysis in (45) and (46), it can be suggested that nominal ellipsis in the Bangla DP follows Ntelitheos’ (2004: 15; (9)) generalization on ellipsis: “Phonological deletion targets elements that have moved to some sort of discourse-related projection, usually a topic phrase.” Our claim that NP must move in order to be elided is consistent with a well known set of observations in the literature. Kayne (2006) suggests that “silent elements can never be in the same position as their pronounced counterparts.” Cinque (2012), referring to the works of Ross (1982) and Koopman (2000), shows that in German and Dutch the phonological deletion of elements is necessarily preceded by the movement of those elements from their merge position. He begins by observing that “(i)n a number of works, deletion of a constituent (or its non-pronunciation) has been taken to depend on the prior movement of that constituent to a left-peripheral position (references omitted).” Cinque (2012: 1) proposes two notions on DP-internal ellipsis. First, “nominal modifiers can be silent (present but unpronounced) only if the NP and the extended projection of the NP below them are also.” Second, the first notion “follows from a condition on DP-internal movement proposed in Cinque (2005), to the effect that only constituents containing the (unmoved) NP can licitly move (and in the case at hand be unpronounced as a consequence of that).” Though the nominal modifier elision is beyond the scope of this paper, the second principle of Cinque’s argument for DP-internal ellipsis is relevant to our discussion of nominal ellipsis. The Bangla DP-internal nominal ellipsis follows the raising of the NP pied-piping the AP. Thus, it can be suggested that nominal ellipsis in the Bangla DP follows Cinque’s (2005) condition on DPinternal movement. Further, I should mention that the focused adjective has to be adjacent to the elided nominal in the Bangla DP, otherwise the NP cannot be elided even in the moved position. This can be seen in (47), where the adjective moves to the Focus1 position which is above demonstrative and thus does not license the NP deletion. (47) * [LAMBAj ei [tj chele]i tin tall this boy three Lit. ‘Tall these three came.’ -Te -Cla ti] eSechilo came 4xxConclusion Restating the main arguments of the paper, we have observed that there are two focus positions in the Bangla DP. One is above the demonstrative (Focus1 position), and the other is below the demonstrative and above numeral-classifier (Focus2 position). Both these focus positions are identified as the contrastive focus positions, as evident from the AOR violation. Syed (2014) draws parallelism of the Focus1 position with Rizzi’s (1997) focus position in the clausal left periphery. Here, I suggest that the Focus2 position is analogous to Jayaseelan’s (2001) IPinternal focus position in Malayalam. I also argue for a topic position right above numeralclassifier and below demonstrative in the Bangla DP. The NP pied-piping the AP moves to that Focus and Nominal Ellipsis in the Bangla DP 85 topic position, which, further, shows that the raising of NP is a discourse related notion. Thus, I propose the following structure of the nominal left periphery in the Bangla DP. (48) [FocP1 [DemP [FocP2 [TopP [NumP [XP AP [NP]]]]]]] References Aboh, Enoch. 2004. Topic and Focus within D. Linguistics in the Netherlands 21:1-12. Bhattacharya, Tanmoy. 1999. Specificity in the Bangla DP. In Yearbook of South Asian Languages and Linguistics, ed. by R, Singh, 2:71–99. New Delhi/London: Sage Publications. Biswas, Priyanka. 2012. Reanalyzing definiteness in Bangla. In Proceedings of the Berkeley Linguistics Society 38, ed. by F, Lionnet, 19-30. Chacón, Dustin. 2011. Head Movement in the Bangla DP. Journal of South Asian Linguistics 4:3-24. Cinque, Guglielmo. 1994. On the Evidence for Partial N-Movement in the Romance DP. In Paths Towards Universal Grammar. Studies in Honor of Richard S. Kayne, ed. by Guglielmo Cinque, Jan Koster, Jean-Yves Pollock, Luigi Rizzi, and Raffaella Zanuttini, Georgetown Studies in Romance Linguistics, 85-110. Washington, DC: Georgetown University Press. Cinque, Guglielmo. 2005. Deriving Greenberg’s Universal 20 and its Exceptions. Linguistic Inquiry 36:315-332. Cinque, Guglielmo. 2012. A generalisation concerning DP-internal ellipsis. IBERIA: An International Journal of Theoretical Linguistics 4, 1:174-193. Dayal, Veneeta. 2012. Bangla Classifiers: Mediating between Kinds and Objects. Italian Journal of Linguistics 24, 2:1-30. Web. É. Kiss, Katalin. 1998. Identificational focus and information focus. Language 74:245-273. É. Kiss, Katalin. 2007. Topic and Focus: Two Structural Positions Associated with Logical Functions in the Left Periphery of the Hungarian Sentence. In Interdisciplinary Studies on Information Structure. Vol. 6, The Notions of Information Structure, ed. by Féry, C., G. Fanselow, and M. Krifka, 69-81. Giusti, Giuliana. 1996. Is there a focus P and a topic P in the noun phrase structure? University of Venice Working Papers in Linguistics 6, 2:105-128. Greenberg, Joseph. 1963. Some universals of grammar with particular reference to the order of meaningful elements. In Universals of language, ed. by Joseph Greenberg, 73-113. Cambridge, MA: MIT Press. Hawkins, John A. 1978. Definiteness and Indefiniteness: A study in reference and grammaticality prediction. London: Croom Helm. Horvath, Julia. 2010. “Discourse features”, syntactic displacement and status of contrast. Lingua 120:1346-1369. Jayaseelan, K. A. 2001. IP-internal Topic and Focus Phrases. Studia Linguistica 55.1:39-75. Johnson, Kyle. 2001. What VP ellipsis can do, what it can’t, but not why. The Handbook of Contemporary Syntactic Theory, ed. by M, Baltin. and C, Collins, 439-479. Oxford: Blackwell. 86 Guha Kayne, Richard S. 2006. On Parameters and on Principles of Pronunciation. Organizing Grammar. Linguistic Studies in Honor of Henk van Riemsdijk (Reprinted in Kayne 2010), ed, by H, Broekhuis et al., 289-299. Berlin: Mouton de Gruyter. Koopman, Hilda. 2000. The Syntax of Specifiers and Heads. London: Routledge. Lyons, Christopher. 1999. Definetess. Cambridge University Press. Ntelitheos, Dimitrios. 2004. Syntax of Elliptical and Discontinuous Nominals, M.A. Thesis, UCLA, Los Angeles, CA. Rizzi, L. 1997. The fine structure of the left periphery. Elements of Grammar, ed. by L, Haegeman, 281-337. Dordrecht: Kluwer. Rooth, M. 1992. A theory of focus interpretation. Natural Language Semantics 1:75–116. Ross, John R. 1982. Pronoun Deleting Processes in German. Paper presented at the annual meeting of the Linguistic Society of America, San Diego, California. Sproat, Richard, and Chilin Shih. 1991. The cross-linguistic distribution of adjective ordering restrictions. In Interdisciplinary approaches to language: Essays in honor of s.-y. kuroda, ed. by C. Georgopoulos and R. Ishihara, 565-593. Dordrecht: Kluwer Academic Publishers. Syed, Saurov. 2014. Adjectives, adjective-fronting, and evidence for focus and topic within the Bangla nominal domain. In The Lexicon–Syntax Interface: Perspectives from South Asian languages, ed. by Pritha Chandra and Richa Srishti, 53-70. Amsterdam: John Benjamins. Syed, Saurov. 2015. Focus-movement within the DP: Bangla as a novel case. In Ulrike Steindl, et al (eds.). Proceedings of the 32nd West Coast Conference on Formal Linguistics: 332-341. Somerville, MA: Cascadilla Proceedings Project. Syed, Saurov. 2016. Decomposing Definiteness: Argument for a split D-domain in Bangla. Kyeong-min Kim et al (eds.). Proceedings of the 33rd West Coast Conference on Formal Linguistics: 390-397. Somerville, MA: Cascadilla Proceedings Project. Ambalika Guha Ph.D research scholar Department of Linguistics and Contemporary English The English and Foreign Languages University, Hyderabad guhaambalika64@gmail.com