NEGATIVE INVERSION IN MODERN HEBREW By JODI RESNICK A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORID A IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS UNIVERSITY OF FLORIDA 2003
Copyright 2003 by Jodi Resnick
This thesis is dedicated to Sam, who has always supported me unconditionally, continually reminding me that as long as I try I will find success and happiness.
iv ACKNOWLEDGMENTS I would like to thank my friends, family, and professors, without whom I could not have reached this level of success. Dr. Galia Hatav, my advisor and professor, has been a constant source of thoughtful guidance. Beyond countless gramma ticality judgements, Dr. Hatav has been an excellent mentor in that she made th e goal of learning through experience number one and has helped me to learn from every turn my thesis took. Dr. Eric Potsdam also deserves recognition for being a dedicated committee member and professor. His Socratic style of teaching helped prepare me to question the literature in the field as well as my own id eas. I appreciate all of the assistance and advice that he has given me pert aining to my thesis as well as my future graduate studies. I would also like to thank the professors who believed in me and helped me to believe in myself. Dr. Anne Wyatt-Brown has showed me care and concern throughout my time at the University of Florida, list ening to my ideas about teaching, my studies, and my future. Working with her has help ed me to become a better-prepared and more confident instructor. Dr. Roberta Golinkoff at the University of Delaware first showed me that I had the ability to be a competent li nguist. She trusted me to work in her lab and set her expectations of me high, leav ing me only the option of success. In addition, I would like to show my appr eciation for the native speakers of Modern Hebrew who have enthusiastically given me their time and attention. I would like to thank Malka Dagan, my instructor of Hebrew, who has allowed me to attend her Hebrew
v classes and has entertained my many questions . Furthermore, I would like to recognize Rabbi Andy Koren and Rabbi Berl Goldman for their determination to find me native speakers. They introduced me to many native speakers who became an amazing source of information. Furthermore, the Israeli St udent Organization welcomed me into their group and invited me to come to their func tions. They were eager to learn about my study and participate as consultants. I am grateful for their generosity. I would like to also thank my friends and colleagues at the University of Florida. Without the support they have given me by unde rstanding the trials and tribulations of being a graduate student, I wonde r if I would have made it to this point. Specifically Robert Blue has continually reminded me that I too will survive graduate school and that I should strive to please only myself. In addition, Phil Monahan has answered many questions about syntax and the thesis process, allowing me to learn from his experiences. Furthermore, I would like to acknow ledge my friends around the world, from Semester at Sea, the University of Delaware, the American University of Paris, Choate Rosemary Hall, and the city of brotherly love. Even when I could not talk about or think about anything other than my thesis and my gr aduate studies, they stood by me as friends. They have been an amazing source of comic relief through my two years at Florida and I am looking forward to spending more time with them in the near future. Finally, I cannot conclude th is section without recongiz ing the love and support of my family. My parents, Ilyse, Bubbe, Pop, Sa m, and the Reichs have all been a source of constant support throughout my undergraduate a nd graduate studies. They have been my cheerleaders, sending e-mails and packages at just the right time to make me smile. From Passover macaroons and sweets to alligator sh aped bath toys and peeps paraphernalia,
vi from vegetable peelers to rose bushes, th eir thoughtfulness has constantly reminded me that although I am far away in miles, I am not in spirit. Furthermore, without understanding fully what I do or why I do it, they have listened to me and have continually expressed their pride in my accomplis hments. I am grateful for their love and support as I recognize that it has been the dr iving force in my determination to continue on no matter how difficult my studies become.
vii TABLE OF CONTENTS Page ACKNOWLEDGMENTS.................................................................................................iv ABSTRACT....................................................................................................................... ix CHAPTER 1INTRODUCTION......................................................................................................ix 1.1 Proposal..................................................................................................................2 1.2 Basic Word Order of Modern Hebrew...................................................................3 1.3 Inversion in Modern Hebrew..................................................................................5 1.3.1 Free Inversion...............................................................................................7 1.3.2 Wh-Interrogative Inversion..........................................................................8 220.127.116.11 Wh-phrases as separate from other triggers.......................................8 18.104.22.168 Structural Analysis of Wh-Interrogative Inversion..........................10 2TRIGGERED INVERSION.......................................................................................12 2.1 Triggers.................................................................................................................12 2.2 Topics and Triggers..............................................................................................14 2.3 Verb Second Phenomena......................................................................................15 2.4 Descriptive Grammars and Grammar Books........................................................16 3FORMAL ANALYSES..............................................................................................19 3.1 Borer (1995)..........................................................................................................19 3.2 Doron and Shlonsky (1990), Shlonsky (1997).....................................................21 3.2.1 Doron and Shlonsky (1990)........................................................................22 3.2.2 Shlonsky (1997)..........................................................................................23 3.2.3 Negative Inversion in Shlonsky (1997)......................................................26 4LEVON (2001)...........................................................................................................28 4.1 Triggered Inversion as a Pseudocleft Construction..............................................28 4.1.1 Glosses........................................................................................................28 4.1.2 Evidence.....................................................................................................30 4.2 LevonÂ’s (2001) Structural Analysis......................................................................32 4.3 Equatives and the Copula in Modern Hebrew......................................................35
viii 5NEGATIVE INVERSION..........................................................................................37 5.1 The Realization of Negation in Modern Hebrew.................................................37 5.2 Sentential Negation as a Separate Ca tegory of Trigger in Modern Hebrew........38 5.2.1 C-command and Scope...............................................................................38 5.2.2 Negation as a Trigger.................................................................................40 5.3 The obligatoriness of Negative I nversion and Coexisting Grammars..................41 5.4 A Structural Analysis for Negative Inversion......................................................42 5.4.1 Sentential Negative Triggers in the CP layer.............................................44 5.4.2 Sentential Negative Triggers in the Focus Projection................................45 5.4.3 Clarification of the Motivation for Inversion.............................................47 6CONCLUSION...........................................................................................................49 6.1 Recapitulation of Negative Inversion...................................................................49 6.2 Modern Hebrew as a language in Transition........................................................50 LIST OF REFERENCES...................................................................................................51 BIOGRAPHICAL SKETCH.............................................................................................53
ix Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Arts NEGATIVE INVERSION IN MODERN HEBREW By Jodi Resnick August 2003 Chair: Galia Hatav Major Department: Linguistics This research explores the differences between negative triggers and other triggers of inversion in Modern Hebrew. It provide s evidence that constitue nt negation is not a separate trigger in itself, but instead a part of potential tr iggers of Triggered Inversion, while sentential negation is a separate trigger and warrants an analysis other than that of Triggered Inversion. It revi ews the previous analyses of Triggered Inversion and discusses their strengths and weaknesses as well as why they do not account for Negative Inversion, inversion as a result of sentential negation. It al so develops an analysis for Negative Inversion, which posits spec, FocP for the landing site of negative triggers. Finally, this research project collected data that questions the oblig atoriness of inversion with negative triggers and hypothesizes th at the variation in grammaticality judgments regarding sentences with pre posed negative triggers is a result of coexisting grammars. Below are example sentences with co rresponding trees, illustrating the two grammatical orders with sentential negative tri ggers. Both have the interpretation: Never does Dina drink coffee.
x XPSVO (No Inversion) af-paÂ’amdinaloSotakafe once-everDinanotdrinkscoffee XPVSO (Inversion) af-paÂ’amloSotadinakafe once-evernotdrinksDinacoffee Preposed sentential negative triggers are in the CP layer, but are not in spec, CP at the surface structure. They are in spec, FocP. Movement of the verb, or lack thereof, is motivated by the Neg-Criterion and the deep structure position of the feature [Neg]. When inversion occurs, the verb is in Foc at surface structure. The feature [Neg] is base generated on lo Â‘notÂ’. The Neg-Criteri on forces movement of both lo and the verb into the CP layer to form a relationship between th e head, Foc, and its specifier. When there is no inversion the verb remains in I because the [Neg] feature is base-generated in Foc and the Neg-Criterion is already satisf ied without additional movement of lo or the verb. t r i gger ver b Foc I P Foc' FocP CP t r i gger Foc I P Foc' FocP CP
1 CHAPTER 1 INTRODUCTION This paper investigates the word order pos sibilities in Modern Hebrew sentences in which a negative phrase comes before both the subject and the verb. It details the cases in which SVO and/or VSO word orders are f ound as the result of this negative trigger, and when applicable, suggests reasons for why there is a variation in word order, as well as a variation in the grammaticality judgments of native speakers. The following sentences illustrate the di fference between the SVO and VSO orders in Modern Hebrew sentences with negative triggers. 1 Each sentence uses the negative expression af-paÂ’am Â‘never (once ever).Â’ This is a negative trigger in itself; however, Modern Hebrew requires the additional word lo Â‘notÂ’ to follow af-paÂ’am in order for the sentence to be grammatical. 1. af-paÂ’amloSotamiryamkafe ever-once (trigger)notdrinkMiriamcoffee Â‘Never does Miriam drink coffee.Â’ 2. af-paÂ’ammiryamloSotakafe ever-once (trigger)Miriamnotdrinkcoffee Â‘Never does Miriam drink coffee.Â’ 3. miryamaf-pamloSotakafe Miriamever-oncenotdrinkcoffee Â‘Miriam never drinks coffee.Â’ 1 I use the following abbreviations in examples: ACC-accusative, COP-copu la, PCPL-participle, ffeminine, m-masculine, s-singular, pl-plural, p-person , pass-passive, past-past, fut-future, pres-present, unacc-unaccusative, ben-Benoni, which is a category of words that is often used as a verb in the present tense.
2 Sentence (1) shows the order that has be en previously claimed by Shlonsky (1997) to be the only grammati cal possibility when the trigger is fronted, XVSO.2 Sentence (2) is an example of the additional order show n to be grammatical in this paper, XSVO. Sentence (3) is another grammatical possibil ity; however, since the negative phrase is not acting as a fronted trigger in this sentence, its grammaticality is not contested and will not be discussed in great detail in this paper. 1.1 Proposal Beyond discussing the grammaticality of the above sentences, in this paper I will do the following: present the previous studies related to inversion in Modern Hebrew, highlighting their strengths and weaknesses; select a system for categorizing the vari ous triggers of Modern Hebrew based on their syntactic and semantic properties a nd discuss why certain types of negative phrases are different and de serve a separate analysis; highlight the differences between the type s of negative triggers and discuss to which analyses they belong; explain my analysis of Negative Invers ion, inversion with sentential negation, which follows Haegeman and GuÃ©ronÂ’s ( 1999) analysis of Negative Preposing in English; discuss the evidence this study provides fo r the existence of co existing grammars in Modern Hebrew. The remainder of Chapter 1 introduces the basic word order of Modern Hebrew as well as inversion in Modern Hebrew, incl uding information about Free Inversion and Wh-Interrogative Inversion. Chapter 2 di scusses Triggered Inversion in relation to topics, Verb Second Phenomena and its desc ription in Modern Hebrew descriptive 2 Shlonsky (1997) uses le-olam lo Â‘neverÂ’ instead of af-paÂ’am lo . Since le-olam lo is of a more formal register, it is possible that ShlonskyÂ’s (1997) claim that inversion is obligatory with negative triggers is a result of using data only from a formal register.
3 grammars and grammar books. Chapters 3 and 4 introduce three formal analyses related to inversion in Modern Hebrew, BorerÂ’ s (1995), Doron and ShlonskyÂ’s (1990), and LevonÂ’s (2001). In Chapter 5 I present my analysis of Negative Inversion, inversion triggered by sentential negation, as a separate process from Triggered Inversion. Chapter 6 is the paperÂ’s conclusion and it highlights the main points of my research as well as some of its implications. 1.2 Basic Word Order of Modern Hebrew Before getting into the details of word or der specific to negation, it is necessary to first review the basic word order of Modern Hebrew. In Modern Hebrew, the following simple word order patterns are considered to be grammatical. The example sentences listed below each pattern are adapted from Borer (1995). 4. S-V-O ha-xatulxisel etha-gvina the-catfinishedACCthe-cheese Â‘The cat finished the cheese.Â’ 5. XP-S-V-O lifneySe-higanuha-xatulxisel etha-gvina before that-arrive[past/1p/pl ]the-catfinishedACCthe-cheese Â‘Before we arrived, the cat finished the cheese.Â’ 6. XP-V-S-O lifneySe-higanu xiselha-xatul etha-gvina beforethat-arrive[past/1p/pl]finishedthe-catACCthe-cheese Â‘Before we arrived, the cat finished the cheese.Â’ In sentences (5) and (6) partic ular phrases, triggers, are present. These phrases are called triggers in the literature because they can trigger inversion of the subject and the verb when they are preposed. Additional, somewhat more complex, gramma tical word order patterns exist. They include sentences that have more than one inflected item, resulting in constructions with
4 compound tenses. The following sentences are adapted from Goldberg (2001) to illustrate these additional patterns. 7. S-COP-PCPL-O rivkahaytakotevetmixtav Rivkabe[past/3p/s/f]write[ben/s/f]letter Â‘Rivka wrote a letter [distant past]Â’3 8. XP-COP-S-PCPL-O yomyomhaytarivkakotevetmixtav daydaybe[past/3p/s/f]Rivkawrite[ben/s/f]letter Â‘Everyday Rivka wrote a le tter [distant past].Â’ 9. S-PCPL-COP-O rivkakotevethaytamixtav Rivkawrite[ben/s/f]be[past/3p/s/f]letter Â‘Rivka wrote a letter [distant past].Â’ 10. XP-PCPL-COP-S-O yomyomkotevethaytarivkamixtav daydaywrite[ben/s/f]be[past/3p/s/f]Rivkaletter Â‘Everyday Rivka wrote a le tter [distant past].Â’ The above sentences show word order pattern s that all start with some element that is not a verb or verb phrase. This does not mean that it would be ungrammatical to start a sentence with a verb in Modern Hebrew; however, in Modern Hebrew, sentences that start with verbs tend to be marked and need specific contexts to be grammatical. One situation, in which VSO word order occurs in Modern Hebrew, without the presence of another phrase preceding the verb, is that of Free Inversion (Doron and Shlonsky 1990). Sentences that have undergone Free Inversion contain a verb that is either unaccusative or passive, and a subject that is indefinite. Free Inversi on is not a major topic in this paper; however, it is discussed in sections 1.3 and 1.4. 3 There is some dispute over the use of the words Â‘distant pastÂ’ here. Dr. Galia Hatav, a native speaker of Modern Hebrew, stated th at Modern Hebrew does not differentiate between the distant and recent pasts. However, since I am not a native speaker myself, an d since these example sentences were taken directly from Goldberg (2001), I am hesitant to change them.
5 All of the above data show that sentences with subject-verb word order patterns are, for the most part, grammatical, that is as l ong as there are not othe r problems with the sentences unrelated to word order. Furthermor e, SVO word order is considered to be the unmarked word order in Modern Hebrew (Goldberg 2001). Schwarzwald (2001:48) goes as far as to say, Â“Modern Hebrew overwhelm ingly prefers the SVO order with the use of simple tenses.Â” This is not the case fo r earlier forms of Hebrew. According to Schwarzwald (2001:48), although the unmarked or der in Modern Hebrew is SVO, in both Biblical Hebrew and Mishnaic Hebrew VSO was the Â“dominant word order.Â” 1.3 Inversion in Modern Hebrew In addition, the data given in section 1.2 show that clauses with VSO word order may, in certain cases, still be grammatical. This marked word order is collectively referred to as inversion. Diffe rent lines have been drawn to syntactically categorize the different types of grammatical inversion in Modern Hebrew. According to Shlonsky (1997) there are tw o general types of inversion in Modern Hebrew, Free Inversion and Triggered I nversion. A previous study by Doron and Shlonsky (1990) also discusses inversion in terms of these same two main types; however, they do not include negatives or wh-int errogatives in their li st of triggers. They do discuss wh-movement, but give it a nearly identical analysis to Triggered Inversion. This type of syntactic classification accounts for the fact that one type of inversion requires a trigger while the other does not; however, it does not account for all of the structural differences between th e inversion types with triggers. More recent research on the inversion pr ocesses of Modern Hebrew categorizes inversion in a different manner. Most re search agrees that Free Inversion should be analyzed separately from other types of i nversion in Modern Hebrew. However, there
6 are varying ways to divide what Doron and Shlonsky (1990) refe r to as Triggered Inversion. Based on the Â“syntactic and stylisti cÂ” characteristics of the inversion process with the varying triggers, the following categ ories have also been hypothesized by Levon (2001). Free Inversion Triggered Inversion Wh-Inversion Negative Inversion This is different in that previous studies about invers ion in Modern Hebrew include Wh-inversion and Negative Inversion with in Triggered Inversion (Shlonsky 1997). Levon (2002:2) bases this diffe rent system of classifying inversion on cross-linguistic facts and the varying Â“stylistic marginalit yÂ” among some of the triggers. Although much of Doron and ShlonskyÂ’s approach to Triggere d Inversion appears to present a plausible analysis for an inversion process, it begins to fail when it uses the same analysis for several different triggers, re gardless of their semantic properties or interpretations. Like Levon, I propose that inversion in Modern Hebrew can be split into four categories. These categories are based on th eir structural analyses. The difference; however between LevonÂ’s definition of the categories and mine, is that I believe constituent negation along with the constituent it is negati ng is a trigger of Triggered Inversion and that the only pote ntial trigger of Negative I nversion is sent ential negation. These four syntactic categories as I define them are shown below and will be discussed further in this paper along with their analyses.
7 Free Inversion o Section 1.3.1 Wh-Interrogative Inversion o Section 1.3.2 Triggered Inversion: including constituen t negation + constituent being negated o Chapter 2 introduces Triggered Inversion o Chapters 3-4 present previous formal analyses o Chapter 5 includes modifications for constituent negation Negative Inversion: inversion due to preposed sentential negation o Chapter 5 1.3.1 Free Inversion Free Inversion, as defined by Doron and Sh lonsky (1990), is a type of inversion in which VSO word order results without the pr esence of any triggering element before the subject and the verb in a given sentence. Free Inversion can only occur when the verb does not assign an external theta role, wh en the verb is unaccusative or passive. Furthermore, definite subject DPs cannot be a part of sentences with Free Inversion. The following is a grammatical set of senten ces with Free Inve rsion (Shlonsky 1990). 11. noladmanhigxadaS was-born[unacc]leadernew Â‘A new leader has been born.Â’ hunxumisparhaxlatot alSulxanha-memSala placed[pass]severalresolutionsontablethe-government Â‘Several resolutions were placed on the governmentÂ’s table.Â’ The following sentence shows that a defin ite DP cannot grammatically occur with the VSO word order being a resu lt of Free Inversion (Shlonsky 1997). 12. */?hunxaha-haxlata alSulxanha-memSala placed[pass]the-resolutionontablethe-government Â‘The resolution was placed on the governmentÂ’s table.Â’ Some native speakers say that sentence (12) is marked, but not completely ungrammatical. Shlonsky (1997) states that sentences like (12) are ungrammatical. Either way it is not as acceptable as th e sentences in (11) or sentence (13).
8 The following sentence shows that VSO word order is not ungrammatical with definite DPs, that is as long as VSO word order is not the resu lt of Free Inversion, but instead is the result of Tri ggered Inversion (Shlonsky 1997). 13. alSulxanha-memSalahunxaha-haxlata ontablethe-governmentplaced[pass]the-resolution Â‘The resolution was placed on the governmentÂ’s table..Â’ The fronted position of the phrase al Sulxan ha-memSala Â‘on the governmentÂ’s tableÂ’ is what forces the inversion here, making this sentence grammatical even though the word order is still VSO. 1.3.2 Wh-Interrogative Inversion In section 1.3 above I state that Triggere d Inversion is not an all-encompassing process that includes all invers ions as a result of a trigger. Based on this conclusion, this section is devoted to a brief su mmary of Wh-Interrogative Inversion. 22.214.171.124 Wh-phrases as separate from other triggers There are a few qualities of Wh-Interrogative Inversion that set it apart from Triggered Inversion. The first is one that is mentioned in Levon (2001) is that WhInterrogative Inversion occurs of ten cross-linguistically. Belo w is an example of subjectverb inversion in French adapted from Ta raldsen (2001) and an example of subjectauxiliary inversion in English. 14. oÃ¹crois-tuquevontsecacherleschats wherethink-youthatwillthemselveshidethecats Â‘Where do you think the cats will hide themselves?Â’ 15. What does Mindy want to do this weekend?
9 In each of these examples, a wh-phrase appear s clause-initially, and as a result, subject inversion occurs. In these tw o cases, that of the matrix verb in example (14) and the auxiliary in example (15) , inversion is obligatory. Another piece of evidence offered by Levon (2001) is that when a trigger of Triggered Inversion is preposed, with or without inversion, the resulting sentence is marked. This is not the case with Wh-Inte rrogative Inversion. There appears to be no semantic difference between the SVO and VSO word orders with a preposed wh-phrase. This is shown below. 16. matailamadSlomo whenstudiedShlomo 17. mataiSlomolamad whenShlomostudied Â‘When did Shlomo study?Â’ Shlonsky (1997) places Wh-interrogatives within his list of triggers, thus being grouped into his observation that the optionality of all triggers is based on register or dialectal differences. More recently, Levon (2001:2) says that Wh-Interrogative Inversion is optional Â“without any discernible semantic effectÂ” unlike other forms of Triggered Inversion. An additional difference between wh-phrases that are serving as triggers and the triggers of Triggered Inversi on is that wh-phrases can trig ger inversion in an embedded clause and then move to a position outsid e of this clause while still maintaining grammaticality and the intended meaning. Tr iggers of Triggered Inversion in Modern Hebrew are clause bound and cannot move in to a higher clause without changing the meaning or grammaticality of a given sentence.
10 126.96.36.199 Structural Analysis of Wh-Interrogative Inversion As mentioned above, since Shlonsky (1997) considers Wh-Interrogative Inversion to be a type of Triggered Inversion he does not give it a separate analysis. Doron and Shlonsky (1990) state varying things about th e relationship between Triggered Inversion and Wh-Interrogative Inversion that leaves it somewhat unclear as to what their exact analysis of the situation is. Regardless of how they classify inversion as the result of a wh-trigger or a non-wh-trigger, they do note on ly one main difference in their analysis. Doron and Shlonsky (1990) hypothesize that Tr iggered Inversion is a case of recursive CP phrases, one in which the trigger lies at surface structure and another higher CP to prevent the trigger from moving to a higher pos ition in a given sentence. The difference they posit is that in Wh-Interrogative Inve rsion an additional CP category above the CP containing the trigger is not created to prevent movement. Finally, they do state that overt and covert movement of I to C account s for the appearance of optional movement. LevonÂ’s (2002) analysis is based mainly on Kayne and Pollock (2001) and their analysis of French Stylistic Inversion. In his analysis, Levon (2002) states the inverted word order is the product of a series of sy stematic movements that are motivated through feature checking. First the wh-interrogative ph rase moves to a projection in the CP layer. Then the subject is topicalized, moving above the wh-interrogative phrase. Next, the IP moves to a position that now dominates the subject. Finally, the wh-phrase moves above the other constituents. The illustrati on below is adapted from Levon (2002). a) Deep structure: [S V Wh] b) Wh-extraction: [Whi[S V ti]] c) Subject topicalization: [Sj[Whi[tj V ti]]] d) IP preposing: [[tj V ti][Sj[Whi]]] e) 2nd Wh-extraction: [Whi[[tj V ti][Sj[ti]]]]
11 In this analysis the only movement that SVO and VSO word order patterns have in common is the first movement of the wh-int errogative to CP. The additional movements: topicalization of the subject, raising of IP over the subject , second movement of the whinterrogative, do not occur when the final linear order is SVO; they only occur with VSO order.
12 CHAPTER 2 TRIGGERED INVERSION Doron and Shlonsky (1990) use the term Tri ggered Inversion to refer to the process in which a phrase appearing at the beginning of a sentence, in some analyses as the result of movement, can cause the verb to raise above the subject, resulting in VSO word order. Doron and Shlonsky (1990) label these phras es that occur at the beginning of a sentenceÂ’s linear order, and that can cause inversion, triggers. Before discussing what can and cannot be a trigger, it is important to note that while not all of the inversion processes are the same, and not all of the triggers are the same; they are not completely unrelated eith er. For instance, in sections 2.2.1 and 2.2 of this chapter the term triggers will be used to include all triggers of all inversion processes in Modern Hebrew that occur as the result of some XP appe aring before the verb and the subject at surface structure. 2.1 Triggers The following is a list of trigge rs as defined by Shlonsky (1997). a) Temporal Adverb b) Prepositional Phrase c) Clausal Adverb d) Direct Object e) Indirect Object f) Clausal Complement g) Wh-expression h) Relative Operator i) Negative Phrase
13 As mentioned, this paper will consider all of the above to be triggers as in the term Triggered Inversion except for wh-expressions and sentential negati on, which are triggers of distinct forms of inversion. Doron and Shlonsky (1990) also note that certain phrases cannot act as triggers in Modern Hebrew. Manner adverbials cannot serve as triggers. The following sentence shows this with its ungrammaticality. 18. *beSeketlomedha-talmid etha-xeshbon quietlystudythe-studentACCarithmetic Â‘Quietly the student studied arithmetic.Â’ Part of the reason manner adverb ials cannot act as triggers in Modern Hebrew lies in the fact that they cannot be clause -initial in any sentences, se ntences with SVO or VSO order (Shlonsky 1997). Furthermore, in order to conclude that a given XP is a trigger it needs to be a part of the clause in which inversion is present (Doron and Shlonsky 1990). 19. le-harbepe ilimhodi aha-miStaraSe-hi to-manyactivistsannouncedthe-policethat-she tagiStvi a pressescharges (a) Â‘The police told many activists that it will press charges.Â’ (b) *Â‘The police announced that it will press charges against many activists.Â’ The prepositional phrase le-harbe pe ilim Â‘to many activistsÂ’ can cause inversion in the given sentence; however, the only grammatical interpretation is that of gloss (a). This is because le-harbe pe ilim belongs to the matrix clause, wh ich in this sentence, is the clause that shows inversi on. In order to get the in terpretation in gloss (b), le-harbe pe ilim would move only to the beginning of the embedded clause and the verb tagiS
14 Â‘pressesÂ’ would show the inve rsion in respect to the pronoun hi Â‘she.Â’ The sentence for gloss (b) is shown below. 20. ha-miStarahodi aShe-le-harbepe ilimtagiS the-policeannouncedthatto-manyactivistspresses hitvi a shecharge Pervious analyses have listed the tri ggers and non-triggers as shown above. A generalization can be made about what can a nd cannot act as a trigge r. A trigger is a phrase that must be able to appear clause initially, is not a topi c, and that, with the exception of wh-phrases, is clause bound. 2.2 Topics and Triggers It is important to note that there is a difference between topics and triggers. This difference is in the structure as well as th e practical use or interpretation of a given sentence. According to Doron and Shlonsky (1990) triggers are st ructurally in the specifier of the complementizer phrase, whereas topics are adjoined with an IP and even though they are in different position, additional problems rule out the possibility of them co-occuring. However, examples used in Shl onsky (1997) as well as data collected for this research appears to show that they can co -occur. As will be returned to in chapter 5, the anaysis of this paper believes that both Topic Phrases and Focus Phrases exist within the CP layer and that they are the landing sites for topics and triggers respectively. Topics are more restricted than triggers in that topics must be definite and are always followed by subjects. Triggers, on the other hand, can be definite or indefinite. Doron and Shlonsky (1990) claim that trigge rs are always followed by verbs; however, Shlonsky (1997) as well as my own research found that triggers do not always trigger
15 overt inversion. Furthermore, another difference between to pics and triggers is that topics are not clause bound as triggers are. The table below summarizes the above obs ervations about topi cs and triggers. topictrigger clause-boundnoyes (except wh-phrases) indefinitenoyes definiteyesyes inversionnoyes/no As this table shows, in most sentences it is possible to distinguis h between topics and triggers using morphological and syntactic observations alone. However, when no inversion occurs and the fronted constituent is definite, it is difficult to tell if the fronted constituent is a trigger or a to pic. In this situation the difference between a topic and a trigger is a matter of a fronted phraseÂ’s interpretation within a given sentence. 2.3 Verb Second Phenomena Previous analyses state that Triggered I nversion in Modern Hebrew is similar to, but not the same as, the verb second phenomena of Germanic languages. However, it does not appear that any of their obser vations completely separate the two. Shlonsky (1997) states that in an embedde d context, inversion can occur regardless of the type of verb. Some Germanic langua ges only allow inversion in embedded clauses with bridge verbs, verbs that take senten tial complements and that allow extraction from their complements. Furthermore, in some Germanic languages inversion cannot occur with overt complementizers. Below is an example from Shlonsky (1997) to illustrate the grammaticality of inversion with an overt complementizer Se Â‘that.Â’
16 21. mipneySeba-pSitaha-leilit acraha-miStara becausethatin-the-raidthe-nightlydetainedthe-police pe ilimrabimhexlatnul argenhafgana activistsmanydecidedto-organizedemonstration Â‘Because the police arrested many activists in the nightly raid, we decided to organize a demonstration.Â’ These observations are not enough to separate Triggered Inversion from verb second because although languages such as Germ an do not allow inversion with overt complementizers and only with bri dge verbs, languages like Yiddish do. Shlonsky (1997) states that the sentence be low is evidence that Triggered Inversion can occur in other positions besides the second position, unlike verb second languages. 22. etha-pa ilha-zematai acraha-mishtara ACCthe-activistthe-thiswhendetainedthe-police Â‘When did the police detain this activist?Â’ In this sentence the verb occurs in the third position with two constituents, ha-pa il ha-ze and matai , coming before it. It is not clear if ha-pa il ha-ze is acting as a trigger. If it is a topic it could be said to be outside of the clause in which inversion is occurring, labeling the verb position as th e second position, not the third. 2.4 Descriptive Grammars and Grammar Books The alternation referred to as Triggered Inversion in this paper is mentioned to varying degrees in Modern Hebrew descri ptive grammars and grammar books. Since these texts are not formal syntactic analyses , they do not attempt to account for Triggered Inversion. However, they are still helpful in that they discuss the authorsÂ’ observations about how native speakers use the rules behind Triggered Inversion. Many reference grammar books, such as Modern Hebrew Structure , by Ruth Aronson Berman, describe a rule that app ears to capture the marked word order
17 associated with Triggered Inversion (1978: 151). In BermanÂ’s text, verb preposing is referred to as being Â“a re-ord ering rule in formal, liter ary Hebrew.Â” According to Berman, this rule is: Â“where [a] comp lement-whether an obligatory object-type complement or an optional modifier type co mplement-is preposed to the beginning of the sentence (generally for purposes of focus) the verb of that complement must precede its subject NP.Â” In other words, the complement appearing at the begi nning of the sentence triggers the verb to come before its ag ent. Berman goes on to say that often in Â“colloquial usageÂ” the above ru le is not applied and the unmarked order, SVO remains. What is particularly interesting about th e above-quoted material and what will be discussed further in this res earch is the use of the word must in the rule as well as the observation that this rule applies in Â“formalÂ” and Â“literaryÂ” settings. Berman (1978:151) concludes her comments on verb preposing by saying that it is Â“very commonÂ” for the subject to come before the verb in colloquial contexts. Previously analyzed sentences lead Shlonsky (1997) to the co nclusion that Triggered Invers ion is obligatory in certain contexts, specifically with negative triggers. However, additional sentences were collected as part of this research project a nd they appear to show that, at least in some more colloquial contexts , that even negative triggers do not always require inversion. In fact, it appears as if Berman (1978) is actua lly describing what I have determined to be the effect of coexisting grammars in Modern Hebrew that are based on register variation. In addition, more descriptive grammars, such as Modern Hebrew by Ora Schwarzwald (2001:66), states a similar rule, th at Â“if an object or an adverbial occurs at the beginning of a verbal sentence, the verb precedes the subject NP.Â” In her grammar,
18 Schwarzwald says simply that this rule is optional. Schwarzwald unfortunately does not go into the details of when she believes the rule to be adhered to and when it is violated. A final note about Triggered Inversion and its description in Modern Hebrew grammar books is that it apparently used to be something only found in the past and future tenses. It was not part of the be noni, or present tense (Berman 1978, Schwarzwald 2001). This makes sense because traditionally in Hebrew the benoni was not considered to be a verbal tense. In present day M odern Hebrew, native speak ers perceive the benoni as a verb tense even though it behaves quite differently from the other tenses. Since people have begun to perceive it as a verbal te nse, it is not surprising that they have also begun to apply verbal rules to it such as Triggered Inversion. Formal studies of subject and verb movement in general, as well as studies specific to Triggered Inversion have been completed. Three distinct analyses will be discussed in detail in the following chapters, that of Bo rer (1995), that of Doron and Shlonsky (1997), and that of Levon (2001).
19 CHAPTER 3 FORMAL ANALYSES Borer (1995), Doron and Shlonsky (1990), a nd Shlonsky (1997) give analyses for inversion as the result of some constituent being clause-initial at surface structure. While there are elements of these analyses that ar e insightful, these analyses do not completely capture the observations that have been made about Triggered Inversion and its interpretation. 3.1 Borer (1995) Borer (1995) does not specifically discuss the triggers of Triggered Inversion as defined by Doron and Shlonsky (1990) in great detail; however , Borer (1995) does include a brief section on this inversion proces s. Borer uses the term Stylistic Inversion to refer to inversion processes including Tri ggered Inversion within her larger work on verb movement in Modern Hebrew. According to Borer (1995:1), Modern Hebr ew is currently a language in transition. In her paper on verb movement, she explains that it is logical for Modern Hebrew to follow the patterns of both VSO and SVO la nguages because Modern Hebrew is in an Â“intermediate stageÂ” between the two. Mo re specifically, Borer accounts for the grammaticality of both word orders by stating that, in a given sentence, both the subject and the verb are in the verb phrase at the deep structure, in SPEC-VP and V respectively, and that both the subject and the verb can optionally raise.
20 Borer (1995) presents several scenarios that may occur as the re sult of the optional movement of the verb and the subject. On e scenario in which SVO word order is the result at the surface structure with a trigger is below. XP I Subject Verb V Object V' VP I' IP In this situation both the subj ect and the verb remain in th e verb phrase in their basegenerated positions. The trigge r is in SPEC, IP, but invers ion is not a result. Borer presents an additional scenario in which VSO word order occurs at surface structure. Here the subject does not raise, it remains w ithin the verb phrase; but the verb does raise to I. Borer (1995) does not discuss syntactically how or why the subject and verb of a given sentence do or do not move. Goldberg (2001) states that it could be hypothesized XP Verb i I Subject t Verb i Object V' VP I' IP
21 that within BorerÂ’s analysis strong or w eak inflectional featur es of the verb and agreement features of the subject determin e whether movement is overt or covert. However, within her analysis, as it is written, not determining the details of what syntactically motivates the subject and the verb to move allows sentences to be generated that show inversion, but that have no trigge r in SPEC, IP (Goldberg 2001). Sentences with inversion, but no trigger, are sometimes grammatical examples of Free Inversion. It is not clear though that this is BorerÂ’s (1995) intention when she left out an explanation for what motivates verb and subject movement. Borer does, however, discuss evidence that verb raising is opti onal. She does this by showing examples in which the words associ ated with the verb ar e in low positions as well as examples in which they are in inte rmediate positions (Goldberg 2001). As noted by Goldberg, Borer specifically uses Â“the part iciple-copula cluster of copula inversionÂ” to illustrate the optionally of verb raising (2001:6).1 What is most useful about Borer (1995) is its analysis of Modern Hebrew as a language in transition. As will be further di scussed later in this paper, I believe that Modern Hebrew is a language with coexis ting grammars. One of the two possible explanations I give for thes e coexisting grammars in Modern Hebrew is that they represent different levels of the transition that Borer believes Modern Hebrew to be undergoing. 3.2 Doron and Shlonsky (1990), Shlonsky (1997) Doron and Shlonsky (1990) and Shlonsky (1997) make a few assumptions about the clause structure of Modern Hebrew based on previous research un related to inversion 1 An illustration of the particle-copula cluster in Modern Hebrew and the possible word orders with them can be found above in senten ces (7-10) in section 1.2.
22 constructions. As does Borer (1995), both Doron and Shlonsky (1990) and Shlonsky (1997) state that in Modern Hebrew both the verb and th e subject begin in the verb phrase. Unlike Borer, both of these studies state that the verb and the subject raise obligatorily to a higher positi on outside of the verb phrase . Shlonsky (1997) notes that the benoni2 can also be involved in inversion as the result of a trigger since it acts as a verb in the present tense. The verb move s to I and the subject moves into SPEC-IP. Goldberg (2001) reviews Shlonsky (1997) and Borer (1995) and finds that based on data within the realm of Triggered Inversion, as well as other clause structures in Modern Hebrew, that the verb and th e subject obligatorily move out of the VP layer in Modern Hebrew. Specifically, in the case of Triggered Inversion, Doron and Shlonsky (1990) and Shlonsky (1997) have very similar, but not identical analyses. This section presents Doron and ShlonskyÂ’s (1990) analysis first. ShlonskyÂ’s (1997) analysis follows in terms of its modifications from Doron and Shlons ky (1990) along with poten tial pitfalls of this type of analysis. 3.2.1 Doron and Shlonsky (1990) Doron and Shlonsky (1990) state that what is occurring in cases of Triggered Inversion is a result of CP recursion. The di agram they use in their analysis is below. 2 The use of the benoni, specifically as a verb tense, was previously mentioned in footnote 1 of chapter 1 and in section 2.3.
23 CP CP XP CÂ’ C IP IÂ’ I VP VÂ’ VÂ… Doron and Shlonsky (1990) say that when the subject does not invert, and comes before the verb, that the subject is in SPEC, IP and there is no extra CP layer. In this case the verb is in I. When the surface structure word order is VSO, the verb moves to C, over the subject and the trigger is in the specifier pos ition of the lower CP. They state that the additional CP node forms a barrier for extraction. Other points in their analysis have been touched upon previously in this paper. Doron and Shlonsky (1990) state that the ex istence of the additional CP layer accounts for the fact that triggers are clause bound. The triggers cannot move above the nonlexically-marked node. Wh-triggers are an exception to this in that they can grammatically move out of the clause they affect in meaning. Doron and Shlonsky (1990) account for this by saying that wh-tri ggers do not have the additional CP layer. 3.2.2 Shlonsky (1997) One modification from Doron and Shl onsky (1990) is that Shlonsky (1997) includes their list of triggers as well as ne gative phrases and wh-i nterrogative phrases. He does mention slight differences between negative phrases and th e other triggers, but on the whole he gives all triggers the sa me analysis. The only difference Shlonsky (1997) lists between negative tri ggers and other triggers is th at he believes that negative triggers obligatorily force inversion. A nother modification of Shlonsky (1997) is not necessarily much of a difference from Do ron and Shlonsky (1990), but is worth noting
24 for its clarification purposes in Shlonsky ( 1997). He reminds readers that CP has many projections and that saying that the trigger is in the CP layer is enough for his analysis since discussing the COMP domain in great detail is beyond the purpose of his study. Since Shlonsky (1997) also believes that the inverted word order is a result of triggered I to C movement, he provides the following diagram for illustration. I to C movement: Shlonsky (1997) refers to feature-checking and criteria-based approaches for explanations. He does not adop t nor rule out one or the othe r. Instead, he describes how both can be used to with structure to ac count for both the SVO and VSO word order patterns. First, Shlonsky (1997) discusses Choms ky (1993) and his feature checking theory in accounting for the movement from I to C. Shlonsky (1997) says that this movement from I to C and the position of the trigger in SPEC, CP creates a connection between the verb and the trigger. According to Sh lonsky (1997:149) the trigger moving into SPEC, CP forces I to C movement in Â“formal wr itten HebrewÂ” only. He goes on to explain that strong features of the trigger need to be ch ecked in formal registers of Modern Hebrew with those of the inflected verb, whereas in colloquial registers the same features are weak, and do not need to be checked overtl y. When the features do not need to be Tr i gger C Ver b I I ' I P C' CP
25 checked before the surface structure I to C m ovement of the verb does not occur. Instead, they are checked at LF. This analysis us ing strong and weak features accounts for what appears to be an optional variation. Secondly, Shlonsky (1997) briefly states that the criteria approach of Rizzi (991) among others also work because under this theory the verb moves from I to C to satisfy a particular criterion. The SPEC position has the operator feature(s) that need to be coindexed with the head that has like features . Furthermore, the head has certain features that need to be coindexed wi th the trigger that is in th e SPEC position. He generalizes the varying criteria by stating: An XP marked [+F] must be in a SPEC-h ead configuration with an X marked [+F] and vice-versa. The remainder of ShlonskyÂ’s analysis of Triggered Inversion in Modern Hebrew focuses on variations based on the classes of triggers. In terms of wh-triggers, Shlonsky notes that it is the only trigger whose tr ace can cause Triggered Inversion. The other triggersÂ’ traces cannot license inversion as a result of non-wh-triggers being clause bound and unable to trigger inversion more than on ce in a given sentence. Shlonsky (1997) also refers to Borer (1984) when he mentions that relative clauses are unique in that observations have lead to the conclusion that inversion can only occur in the highest COMP of a given relative clause. He pos its a possible analysis in which a phonetically unrealized resumptive pronoun acts as a covert tr igger. He uses situations in which an overt resumptive pronoun may appear in the same position as evidence. Sentences illustrating this are below (Shlonsky 1997). 23. haadamSexaSavdaniSerinapagSaÂ… the-personthatthoughtDanithatRinamet Â‘The person that Dani thought that Rina metÂ…Â’
26 24. ??haadamSexaSavdaniSepagSarinaÂ… the-personthatthoughtDanithatmetRina Â‘The person that Dani thought that Rina meetÂ…Â’ 25. haadamSe otorinapagSaÂ… the-personthathimRinamet Â‘The person that Rina metÂ…Â’ 26. haadamSe otopagSarinaÂ… the-personthathimmetRina Â‘The person that Rina metÂ…Â’ Sentences (23) and (24) show the gram maticality judgments that result with inversion in relative clauses without overt resumptive pronouns. Sentences (25) and (26) provide evidence that resumptive pronouns can ac t as triggers in re lative clauses overtly. Again, Shlonsky (1997) hypothesizes that re sumptive pronouns may exist in a sentence like (23), but that they are covert. He states that these resumptive pronouns are basegenerated in COMP. It is not surprising that Shlonsky (1997) makes an exception with this particular type of tri gger in terms of it being base-generated because he makes a similar exception with negative triggers. Th is exception does accoun t for the given data with relative clauses and inversion; however , since my analysis does not accept Shlonsky (1997) in terms of his analysis of Nega tive Triggered Inversion as an exception of Triggered Inversion, I am hesitant to accept this exception. As stated in Levon (2001), separate research needs to be conducted on a ll of the triggers indi vidually to see if they each have unique properties. 3.2.3 Negative Inversion in Shlonsky (1997) Finally, Shlonsky (1997) discusses his account for the differences he observes about negative phrases acting as triggers. Using a criteria a pproach such as Rizzi (1991), Shlonsky (1997) states that his observation th at Negative Inversion is obligatory lies in
27 the fact that the feature [+Neg] cannot be base generated in the CP layer. It can only be generated in a lower posit ion on the negative head lo forcing the verb to raise over the subject in its effort to coinde x its features and for the Neg-cr iterion to be satisfied. The Neg-criterion is shown below as it is stated in Haegeman (2000). A. A Neg-operator must be in a Spec-Head configuration with an X-[NEG]. B. An X-[NEG] must be in a Spec-Head configuration with a NEG-operator. The explanation of Shlonsky (1997) poses a problem concerning obligatoriness. In the research process for this paper, native speakers of Modern Hebrew were consulted about varying sentences with and without inversion. I observed many native speakers not using inversion with negative triggers freque ntly and, according to them, grammatically. This conflicts with ShlonskyÂ’s (1997) claim that inversion is obligatory.
28 CHAPTER 4 LEVON (2001) Levon (2001) proposes an altern ative analysis that more accurately accounts for the use of Triggered Inversion in Modern Hebrew in that it recognizes that sentences with Triggered Inversion are semantically and pr agmatically different from their unmarked counterparts. This analysis does have so me weak areas and does not completely account for Triggered Inversion; howev er, Levon (2001) does come closer to the correct use and interpretation of Triggered Inve rsion than previous analyses. Like previous studies, Levon (2001) belie ves that there are differences between Triggered Inversion and verb second phenomen a. He does not try to form structural connections between the two. In his analysis, Levon (2001 ) states that Triggered Inversion is a type of focus construction, more specifically a specificational equative construction. Levon (2001) compares these eq uative constructions in Modern Hebrew to pseudocleft constructions in English. He goes on to say that both English pseudoclefts and Modern Hebrew Triggered Inversion are Â“language specific manifestations of the same constructionÂ” (Levon 2001:4). 4.1 Triggered Inversion as a Pseudocleft Construction 4.1.1 Glosses Before discussing the details of LevonÂ’s an alysis, it would be helpful to note some observations about the interpretation of sent ences with triggers. Often, Modern Hebrew sentences with triggers are glossed in the sa me manner as those in which the trigger is not preopsed. This assumes, however, that there is no difference in the interpretation of the
29 two types of sentences. Doron and Shlonsky ( 1990) highlight the fact that they have the same gloss. They state that although senten ces with preposed triggers have different discourse functions, they have the same truth conditions. I also encountered this difference in discourse function in speaking w ith native speakers of Modern Hebrew. For instance, out of context, several native sp eakers said that they would only use the unmarked word order in which the trigger is not preposed. In further conversations about the use of sentences with preposed triggers, they said that they would need a specific context in which the trigger was the answer to a previous statement or question to use the order with a fronted trigger. Levon (2001) also noticed a pragmatic differe nce; however, he does not go into the details of when each is used. He posits a need to capture this difference between the unmarked word order and sentences with pr eposed triggers. To do this, Levon (2001) uses different glosses for the sentences w ith and without preposed triggers. The following sentence from Levon (2001) and its gl osses provide an illustration of this. 27. harbepe ilim acraha-miStaraba-pSitaha-leilit manyactivistsdetainedthe-policein-the-raidthe-nightly (a) Â‘Many activists was who the police detained in the nightly raid.Â’ (b) Â‘The police detained many activ ists in the nightly raid.Â’ Gloss (a) is the additional gloss that Levon a dds in his analysis of Triggered Inversion. Gloss (b) is from Shlonsky (1997) and repres ents the way in which most sentences with or without Triggered Inversion are glossed. Following the general pattern of glossing Triggered Inversion in previ ous research, the following sent ences could also be glossed with gloss (b) above. 28. harbepe ilimha-miStara acraba-pSitaha-leilit manyactiviststhe-policedetainedin-the-raidthe-nightly
30 29. ha-miStara acraharbepe ilimba-pSitaha-leilit the-policedetainedmanyactivistsin-the-raidthe-nightly Levon (2001) notes that gloss (a) is a pseudocleft construction in Englis h and states that it should be analyzed as such. Some native speakers of Modern Hebrew que stion the use of a pseudocleft as the gloss for Triggered Inversion. They say that if they were to translate the English pseudocleft into English, it would not be a case of Triggered Inversion. However, they do admit that in translating the Modern Hebrew into the English, it is difficult to find an exact match and that the pseudocleft cons tructions used by Levon (2001) might be the closest match and the most ade quate gloss, although not perfect. 4.1.2 Evidence His first piece of evidence is based on Kiss (1998) in his discussion of both Triggered Inversion and pseudoclefts in terms of focus constructions that do not Â“manifest an exhaustive operatorÂ” (2001:4). Af ter checking his data with native speakers and reviewing Kiss (1998), it appears as if Levon (2001) is correct in saying that Triggered Inversion and pseudoclefts are simila r in terms of exaus tive identification, but incorrect in that they both do manifest an exhaustive operator. Although it appears that Levon (2001) incorrectly states that they do not express exhaustive identification, the use of exhaustive operators sti ll supports his analysis of Triggered Inversion being a pseudocleft construction since their exhaus tive identification facts are the same. The following set of sentences is adapted from Levon (2001).1 1 Levon (2001) marks the direct objects in these sent ences as accusative, but does not use the definite marker. This is ungramma tical according to the nativ e speakers in this study.
31 30. praximvexocoladnatandanile-Rina flowersand chocolategaveDanito-Rina Â‘Flowers and chocolate was what Dani gave to Rina.Â’ 31. praximnatandanile-Rina flowersgaveDanito-Rina Â‘Flowers was what Dani gave to Rina.Â’ This test is based on Szabolcsi (1981) and stat es that since sentence (31) is not a logical consequence of sentence (30), exhaustive id entification is expressed. Both the Modern Hebrew sentences and their English gl osses express exhaustive identification. Furthermore, neither pseudocleft construc tions in English nor Triggered Inversion in Modern Hebrew allows for wh-extracti on. Levon (2001) uses the sentences below to illustrate this. 32. harbesfarimnatandanile-rinaetmol manybooksgaveDanito-Rinayesterday Â‘Many books was what Dani gave to Rina yesteday.Â’ 33. *le-miharbesfarimnatandanitetmol to-whomanybooksgaveDanityesterday Â‘*To who was many books what Dani gave yesterday?Â’ 34. *miharbesfarimnatantle-rinaetmol whomanybooksgavetto-Rinayesterday Â‘*Who was many books what t ga ve to Rina yesterday?Â’ Examples (33) and (34) are ungrammatical in both Modern Hebrew and English. The final piece of evidence that Levon (2001 ) provides in support of his analysis of Triggered Inversion as a pseudocleft construc tion is based on their similarities in terms of scope. He provides the following examples. 35. Some boy is the problem in every school. 36. The problem in every school is some boy.
32 37. rutlozaxraet-kol-ha-xomer RutNEGrememberedACC-all-the-materials Â‘Ruth did not remember all of the materials.Â’ 38. et-kol-ha-xomerlozaxrarut ACC-all-the-materialsNEGremembered Rut Â‘All of the materials was what Ruth did not remember.Â’ Sentences (35) and (37) show that in both Modern Hebr ew and English, predicative sentences are often ambiguous in terms of scope. Sentence (35) has the following two interpretations available. a) In every school there is a boy, a different boy, that is a problem. b) The same boy is a problem for every school. Sentence (37) has the following two interpretations available. a) Ruth forgot every material. She did not remember anything. b) Ruth forgot some, but not all of the mate rials. She remembered some of them. Sentences (36) and (38) show that the corresponding specif icational equative sentences are not ambiguous and only have one possible interpretation which is the interpretation listed in (a) for both the Modern Hebrew and the English examples above. These sentences show that fronting a trigger in Modern Hebrew forces a wide scope interpretation of the trigger. 4.2 LevonÂ’s (2001) Structural Analysis Levon (2001) states that the tr igger is base-generated in the specifier position of the functional phrase in the operat or layer. The trigger is th en co-indexed with a long-chain to the position it is in when a given senten ce is interpreted. The subject and the verb begin within the verb phrase. Transformations then occur th at place the subject, and verb in different position than did previous studies. Levon (2001) states that the subject is raised to a neutral phrase projection within the inflectional layer. The subject is still below the verb in this analysis because this analysis does support V to I movement and
33 the neutral projection is belo w the landing site for the verb. Levon (2001) mentions that with the subject in this neut ral phrase, it receives a neutra l interpretation. The verb does not move to C or into the CP layer at all in this analysis. Finall y, the analysis of Levon (2001) also argues that a null pronominal c opula comes between th e trigger and the rest of the clause to which it belongs. Below is a sentence with its corresponding tree diagram to illustrate the analysis of Levon (2001). 39. le-rinanatandanipraxim to-RinagaveDaniflowers Levon (2002) glosses this sentence as Â‘Flowers was what Dani gave to Rina.Â’ However, based on Levon (2001) and his written analys is in prose and tree diagrams in Levon le-Rinai Pron proi Opi proj natank Danij t j titk praxim VP NeutP AspP IP XP AgrSP CP DP G' GP
34 (2002), I am assuming that he meant for the gloss to be something more like Â‘(To) Rina was whom Dani gave flowers.Â’ The tree above shows that the trigger is in the specifier position of the functional phrase in the operator layer (GP) and is co-indexed through a long-chain with the position in which it is interpreted. The traces of th e verb and the subject are in the verb phrase along with the remainder of the sentence. The subject is in the speci fier of the neutral phrase, which is below the verb, which is in AgrS. Pron is the null pronominal copula that is needed to syntactica lly show that the two arguments are part of an equation. Section 4.3 below provides LevonÂ’s (2001) r easoning for labeli ng the null copula as Pron. One problem with this tree is the fact that the subject is co-indexed with pro, which is in spec, AgrSP. This is a c-command vi olation. Levon (2001) is not clear as to why he believes spec, AgrSP needs to be filled with pro, although a plausibl e explanation is that it is for agreement purposes between the subj ect and the verb. One way to avoid this problem with c-command while maintaining proper agreement is to leave spec, AgrSP empty at the surface structure and to say that covert movement of the subject occurs at LF. Furthermore, it is important to note th at although Levon (2001) does not directly discuss the structure that would correspond with sentences in which inversion is not triggered, he does recognize that inversion does not always occu r when the trigger is present. He does not explain how, why, or when he believes that inversion occurs. He does seem to analyze all sentences with pre posed triggers as pse udocleft constructions, not just those with inversion.
35 4.3 Equatives and the Copula in Modern Hebrew The structure described above is base d on equatives in Modern Hebrew as described in Rothstein (1995). Modern He brew has a copula in the past and future tenses; however, in the present tense, the copul a is said to be defective. Many present tense sentences that would use a copula in other languages do not in Modern Hebrew. The following set of sentences adapte d from Levon (2001) illustrates this. 40. anihayitiSamen Ibe-PASTfat Â‘I was fat.Â’ 41. aniihyeSamen Ibe-FUTfat Â‘I will be fat.Â’ 42. aniSamen Ifat Â‘I am fat.Â’ In certain cases, however, Modern Hebrew does use nominative pronouns to represent the copula in the present tense. An exam ple of this from Levon (2001) is below. 43. danihumaryosef Dani3MSMr.Josef Â‘Dani is Mr. Joseph.Â’ Doron (1983) called this form of the copul a Pron. Based on Roth stein (1995) it can be concluded that sentence (42) does not have Pron because the adjective Samen Â‘fatÂ’ as a theta role that is filled by the pronoun ani Â‘IÂ’. On the other hand since mar yosef Â‘Mr. JosefÂ’ does not have a theta role that can be filled by Dani , sentence (43) needs the copula. Levon (2001) goes on to say that since Pr on allows for the syntactic licensing of equatives like those above, it can also allow for the syntactic licensing of equatives that are the result of Triggered Invers ion. This is illustrated below.
36 44. Â‘Dani is Mr. Josef.Â’ [dani[hu[mar yosef]DP]IÂ’]IP 45. Â‘Rina is who Dani gave flowers.Â’ [le-rinai[Pronarb[DPproi[CPOpi[AgrSPprojnatank[IPtk[AspP[NeutPdanij[AgrOP[VPtjpraxim tkti]]]]]]]]]]] The only difference noted by Levon (2001) be tween the equatives like in sentence (44) and the equatives that ar e the result of Triggered Inve rsion, as in sentence (45), is that the arguments are different. The argumen ts of the equatives above are DPs whereas the arguments of equatives that are a result of Triggere d Inversion are not.
37 CHAPTER 5 NEGATIVE INVERSION The previous two chapters have gone ove r the current analyses of Triggered Inversion. In this chapter I will discuss my hypothesis that certain negatives phrases require a separate analysis from Triggered Inversion. I base my analysis of Negative Inversion in Modern Hebrew on Haegeman and GuÃ©ronÂ’s (1999) analysis of Negative Preposing in English. 5.1 The Realization of Negation in Modern Hebrew Negation is formed in Modern Hebrew by adding the word lo Â‘notÂ’ in front of the verb, adjective, or noun being negated. Addi tional negation can occur in Modern Hebrew by using a few different sets of words. With each set, the word for Â‘notÂ’ still appears directly before what is being negated; however, an additional free morpheme appears before the word lo . For example, the representation of the concept of Â‘neverÂ’ is with the word lo plus an additional free morpheme, either le-olam or af-paÂ’am. This additional morpheme is glossed, in certain cases, as Â‘f orever,Â’ Â‘once everÂ’ or Â‘never.Â’ Shlonsky (1997) uses le-olam lo ; however, since native speakers told me that this is a phrase from a higher register, I also used af-paÂ’am lo to see if using a more colloquial expression would affect inversion. Grammatical examples of Negative Inversi on, like the one below, show that this additional morpheme does not need to be directly before the word lo , but that it can occur in a higher position.
38 46. af-paÂ’amDanlolomed once-everDannotstudy Â‘Never does Dan study.Â’ In this sentence af-paÂ’am and lo are not in sequential order because they are separated by the subject, Dan . In cases with sentential negati on, such as the sentence above, the additional morpheme is the negative trigger. 5.2 Sentential Negation as a Separate Category of Trigger in Modern Hebrew As previously mentioned, Shlonsky (1997) includes negative phrases as one of several triggers of inversi on. Levon (2001:2) claims that there are differences between negative and non-negative triggers, but do es not go into detail beyond stating that Negative Inversion is a cros s-linguistic phenomena and that Negative Inversion is not Â“stylistically marginal.Â” The analysis of this paper provides evid ence that constituent negation behaves as a part of other triggers of Triggered Inversi on and that sentential negation should be recognized as a separate category of tr igger, a trigger of Negative Inversion. 5.2.1 C-command and Scope In addition to the cross-linguistic te ndency for negation to cause inversion, negation is unique among the other triggers that Shlonsky (1997) lists in that certain types of negative triggers take different scope as a result of c-command. Many of the other triggers are not scopal expre ssions whereas some negative triggers are. The following sentences illustrate this point. 47. af-paÂ’amloSotadinakafe once-evernotdrinkDinacoffee af-paÂ’amdinaloSotakafe once-everDinanotdrinkcoffee Â‘Never does Dina drink coffee.Â’
39 48. le-dinanatanmoSesukariyot. to-DinagaveMoshecandies le-dinamoSenatansukariyot to-DinaMoshegavecandies Â‘(To) Dina is who Moshe gave candies.Â’ In the sentences in (47) the trigger is af-paÂ’am and it has scope over the entire proposition, Â‘Dina drinks coffeeÂ’ because it ccommands it. As mentioned in section 4.1.2 in relation to evidence for LevonÂ’s (2001 ) analysis of Triggered Inversion as a pseudocleft construction, fronted triggers take wide scope. In the sentences in (48) the trigger is le-Dina and it is not a scopal expressi on. In section 5.4, where I present my structural analysis of Negative Inversion, th is difference between triggers of Triggered Inversion and Negative Inversion becomes signi ficant. I argue that triggers that have scope over entire propositions interact w ith IPs and are directly followed by the proposition that they negate. They are not followed by any additional words or phrases that are not part of the IP being negated. Not all negatives c-command the entire proposition that follows them. The above example in which af-paÂ’am takes scope over its proposition has sentential negation because it negates the entire sentence as a re sult of its scope and c-command capabilities. Other negatives, constituent negative, only have scope over the argument that follows them. This is illustrated below. 49. af-paÂ’amloSotadinakafe once-evernotdrinkDinacoffee af-paÂ’amdinaloSotakafe once-everDinanotdrinkcoffee Â‘Never does Dina drink coffee.Â’
40 50. losukariyotnatanMoSele-dinaelaSokolad notcandiesgaveMosheto-Dinabutchocolate losukariyotMoSenatanle-dinaelaSokolad notcandiesMoshegaveDinabutchocolate Â‘Not candies is what Moshe gave to Dina, but (he gave her) chocolate.Â’ In the sentences in (49) the negative trigger is af-paÂ’am and it has scope of the entire proposition. In the sentences in (50) the negative trigger is lo . It is negative in meaning like af-paÂ’am lo ; however, it does not have scope over the entire proposition. It only has scope over the argument that is the rest of the trigger, sukariyot Â‘candies.Â’ I believe that in the above sentence lo is not acting as a trigger by itself. It is part of lo sukariyot Â‘not candiesÂ’ which acts as a larger constituent to trigger inversion. Furthermore, sentential negation has the same scope no matter where it is placed in a sentence while the scope of constituent nega tion may change if its place in the sentence is changed. The sentences below illustrate this point. 51. af-paÂ’amloSotadinakafe once-evernotdrinkDinacoffee Â‘Never does Dina drink coffee.Â’ 52. dinaaf-paÂ’amloSotakafe Dinaonce-evernotdrinkcoffee Â‘Dina never drinks coffee.Â’ Sentences (51) and (52) are different in terms of their di scourse functions in that (52) is the unmarked word order whereas (51) needs to be in a context in which it is a prompted response warranting focus on negati on. However; the fact that af-paÂ’am has scope over the entire proposition does not change between them. 5.2.2 Negation as a Trigger Sentential negation as a trigger does not fit into LevonÂ’s (2001) analysis with pseudocleft constructions whereas constituent negation does. This is expected because
41 Levon (2001) accounts for triggers of Tri ggered Inversion, not Negative Inversion. Lo+___ is a trigger of Triggered Inversion and, like other triggers of Triggered Inversion, does not take scope. Because af-paÂ’am does take scope, it is not a trigger of Triggered Inversion, but instead of Ne gative Inversion. The exampl e below shows that glossing sentential negative triggers as pseudoc lefts is not a favorable analysis. 53. af-paÂ’amloaxladinadag once-evernotateDinafish A. ??Â‘Never was when Dina ate fishÂ’ B. Â‘Never did Dina eat fish.Â’ Gloss (a) is the pseudocleft in terpretation of this sentence. This seems to be a problem in translation though because the English appear s to have the meaning that the action did occur, but this is not possibl e because it happened Â‘never.Â’ The Modern Hebrew sentence does not have this same cont radictory interpretation. The English gloss (b), like the Hebrew, does not have the same problem a nd while it is a marked construction, its grammaticality is not questionable. 5.3 The obligatoriness of Negative Inversion and Coexisting Grammars Before moving on to the details of my analysis of Negative Inversion, it is important to clarify my findings about th e obligatoriness of Negative Inversion. Shlonsky (1997) states that Nega tive Inversion is obligatory in all cases. While talking to native speakers of Modern Hebrew, I discov ered that SVO is actually a common word order pattern with sentential negative tri ggers and that many native speakers actually prefer this order. Levon ( 2002) also observes that inve rsion is not obligatory with Negative Inversion. He bases this judgm ent on his own grammaticality judgments as well as those of three consultants.
42 While the difference between preposing a nd not preposing a trigger is based on a given sentenceÂ’s discourse f unction; the difference between inversion and no inversion when the trigger is preposed is based on coexisting grammars as defined by Lightfoot (1998). Lightfoot (1998) states that it is po ssible for a given language to have more than one coexisting grammar and that individual speakers of the given language may have one or more of them available to them. Having a system of coexisting grammars accounts for what appears to be optionality in ma ny languages. In collecting grammaticality judgements for this research project, I met native speakers of Modern Hebrew who prefer SVO order, those that prefer VSO order, and those who had similar opinions about both orders. The existence of at least two coexisting gr ammars in Modern Hebrew could be the result of register variation and/or the position of Modern Hebrew as a language in transition. Levon (2002) stat es that he believes that it is a result of registers, while Borer (1995) discusses Modern Hebrew as being a language in transition, as discussed above in section 3.1. These two hypotheses are not necessar ily at odds with one another; that is it is quite possible that both extr alinguistic factors play a role in the formation and existence of coexisting grammars in Modern Hebr ew. The data I co llected supports both hypotheses in that of the native speakers in this study, those who strongly prefer SVO order seem to use Modern Hebrew only in colloquial settings and are of a younger generation. 5.4 A Structural Analysis for Negative Inversion As mentioned, I do not believe that any of the previous analyses of Triggered Inversion adequately account for what I have observed about Negative Inversion. Furthermore, I have mentioned that while I am not sure that any of the previous analyses
43 are perfect in their analysis of Triggered Inversion, I do be lieve that Levon (2001) is the most accurate analysis to date. Levon (2001) will be used as the accepted analysis for Triggered Inversion in this section. With this said, one aspect of analyses of Triggered Inversion that needs to be modified concerns constituent negation. Cons tituent negation is a trigger of Triggered Inversion. The structure below is taken from Levon (2002) and it is used to illustrate the use of lo+__, the manifestation of constituent nega tion in Modern Hebrew, as a potential trigger. 54. losukariyotnatanmoSele-dinaelaSokolad notcandiesgaveMosheto-Dinabutchocolate Â‘Not candies is what Moshe gave to Dina, but (he gave her) chocolate.Â’ Here lo+__ is the trigger and is the specifier of the fuctional projection (GP) in the CP layer. Pron represents the null pronominal copula that Levon (2001) posits as a key lo sukariyoti Pron proi Opi proj natank moSej t j titk le-dina ela Sokolad VP NeutP AspP IP XP AgrSP CP DP G' GP
44 element of his pseudocleft analysis. While this structure accounts for inversion with constituent negation, I do not think that a ps eudocleft construction accurately captures the observations I have made about sentential negative triggers. 5.4.1 Sentential Negative Tr iggers in the CP layer Since sentential negative tri ggers are phrases that have scope over the entire IP, I believe that they need to be in a CP layer that dominates the entire IP. This CP does not only contain the trigger at the surface structure, but also the verb if VSO word order is the result. In sentences with sentential negativ es acting as triggers nothing comes between the trigger and the IP over which it has scope. In all sentences in Modern Hebrew, the subject, the verb and the poten tial trigger begin in a low pos ition, within the verb phrase. Furthermore, in all sentences the verb move s to I and the subject moves from the verb phrase to SPEC, IP. When, for reasons of di scourse the trigger is fronted, the verb can either remain in I or move into the CP layer. The syntactic motivation for inversion is based on the Neg-Criterion, which was mentioned in section 188.8.131.52. The coexisting grammar that prefers inversion has the [Neg] feature base-generated on lo , forcing the verb to move into the CP layer to form a head-specifier relationship. In the coexisti ng grammar that does not show inversion, the [Neg] feature is base-generated in the CP layer and the verb remains in I. The explanation above is not detailed e nough; however, to account for all observed characteristics of Negative Inversion. What follows is based on Negative Preposing in English as described in Haegeman and GuÃ©ron (1999).
45 5.4.2 Sentential Negative Trigg ers in the Focus Projection Haegeman and GuÃ©ron (1999) state that th e negative trigger cannot be in the specifier of CP at surface structure. I believ e that this is also true of Modern Hebrew. The following sentences illustrate this. 55. * et-male-mimoSenatan ACC-whatto-whoMoshegave Â‘Who did Moshe give what?Â’ 56. ha-moraamraSe-af-paÂ’amlolomedetrut the-teachersaidthat-nevernotstudiesRuth Â‘The teacher said that never does Ruth study.Â’ Example (55) shows that CP adjunction stru ctures are not possible in Modern Hebrew because questions with more than one preposed wh-phrase are ungrammatical. Furthermore, sentence (56) shows that Ne gative Inversion can occur with an overt complementizer. If there is an overt comp lementizer in C and it appears before the negative trigger, it is not possible for the negative trigger to be in spec, CP. This is illustrated below. Here there is only one CP as stated above. Af-paÂ’am cannot be in spec, CP at the surface structure because at spell-out it is after Se Â‘thatÂ’ which is in C. This tree diagram shows that there must exist an additional landing site for the negative trigger that is between the complementizer and the rest of the sentence. spec Se C af-pa'am spec XP C' CP
46 Haegeman and GuÃ©ron (1999) label the landing site for the trigge r as the specifier of a focus phrase. Rizzi (1997:285) states th at the focus phrase is the landing site for an element Â“bearing focal stressÂ” and providi ng new information. Since preposing the trigger in Negative Inversion in Modern Hebe w is done in discourse situations in which the speaker wants to focus on the negative por tion of the sentence, it follows to adopt focus phrase as the landing site for sentential negative triggers in Modern Hebrew. While focus phrase is an adequate name for the XP in the tree above, what is more significant in this analysis is that it is not CP. The structure that I propose for Modern He brew in cases with preposed negative triggers is shown below w ith and without inversion. Without InversionWith Inversion As mentioned in section 2.1.1, it is sometim es difficult to determine if a preposed XP is a topic or a trigger when it is defi nite and no inversion o ccurs. When the XP is interpreted as a topic it appears in the specifie r of TopP. However, in the spefic case of negation, the XP is in spec, FocP, not TopP , as shown above. This is because the interpretation of preposed negation is one of focus and not topic. Topics are used to signal that the given sentence is going to be about what is spec, TopP (Haegeman 1999). t r i gger ver b Foc I P Foc' FocP CP t r i gger Foc I P Foc' FocP CP
47 While preposing negation does place a cert ain focus on it, the negation cannot become the topic of a sentence. Finally, Haegeman and GuÃ©ron (1999) state th at there is a topic phrase that exists above the focus phrase, but before the comple mentizer phrase in Negative Preposing in English. According to Doron and Shlonsky (1990) this does not carry over to Modern Hebrew because in Modern Hebrew a topic a nd a trigger cannot grammatically be in the same sentence. However, the grammatical ity of sentence (57) below, previously numbered as sentence (22) above, provides evidence that both can occur in Modern Hebrew. 57. etha-pa ilha-zematai acraha-mishtara ACCthe-activistthe-thiswhendetainedthe-police Â‘When did the police detain this activist?Â’ As in English, the Topic Phrase appears to be in a higher posi tion, above the Focus Phrase, within the CP layer. This is illustrated below. 5.4.3 Clarification of the Mo tivation for Inversion With the observation that the landing site for sentential negative triggers is the specifier of the focus phrase in the CP layer, when inversion occurs the feature [Neg] is spec C spec Top spec Foc IP Foc' FocP Top' TopP C' CP
48 base-generated on lo which forces the verb to move to Foc to form a relationship between the specifier and the head. When no inversi on occurs, the feature [Neg] is base-generated on Foc and the verb remains in I.
49 CHAPTER 6 CONCLUSION 6.1 Recapitulation of Negative Inversion In this paper I have discussed the main analyses of Triggered Inversion and I have come to the conclusion that, while not perf ect, with slight modification, Levon (2001) is the most accurate analysis. Although he does not go into great detail about the pragmatic differences between sentences with and wit hout preposed triggers, his analysis does accurately account for both. In my research I found that the trigger is not preposed in unmarked contexts and that the trigger is preposed when it is the focus of a utterance that is generated in response to a previous statem ent. The only significant alteration that I made to his analysis was the addition of lo+__ as a possible trigger. Although this trigger is negative, I believe that it acts more like non-negative triggers than like sentential negative triggers. I believe that sentential negatives are di fferent due to issues of scope and ccommand and that they are actually triggers of a separate process, Negative Inversion. Even though I believe Levon (2001) to be a mo re pragmatically perceptive approach to Triggered Inversion, I do not adopt hi s analysis for Negative Inversion. My analysis provides evidence that the la nding site for the tri gger is not in spec, CP, but instead in spec, FocP. Furthermore, wh en the trigger is in spec, FocP the feature [Neg] is either base-generated on lo which forces inversion, or in Foc which prevents inversion.
50 In addition, Modern Hebrew has two co existing grammars, one in which the inversion occurs and another in which SVO word order remains even with a preposed trigger. Native speakers may have one or both accessible to them. Based on conversations I had with consulta nts, as well as previous ob servations in the literature, I conclude that these coexisting grammars are a result of register variation. 6.2 Modern Hebrew as a language in Transition Finally, I would like to no te that although I do not adop t Borer (1995) in terms of Negative Inversion or Triggere d Inversion, the motivation of her paper, to show that Modern Hebrew is a language in transiti on is intriguing. While I do not have enough evidence to state the existence of coexisting grammars is based on Modern Hebrew being a language in transition instead of register variation, what data I did collect does work within this hypothesis. Lightfoot (1998) states that competing grammars, or coexisting grammars, often arise as the result of ongoing grammatical change. Furt her study into the role of coexisting grammars in Modern Hebrew would help to explain the variation in grammaticality judgements of native speaker s and potentially provide insight into syntactic changes that occur over time.
51 LIST OF REFERENCES Berman, Ruth Aronson. 1978. Modern Hebrew structure . Tel Aviv: University Publishing Projects. Borer, Hagit. 1995. The ups and downs of Hebrew verb movement. Natural Language and Linguistic Theory 13: 527-606. Chomsky, Noam. 1993. A minimalist progr am for linguistic theory. In The view from building 20 , ed. K. Hale and S-J Keyser, 1-52. Cambridge, MA: The MIT Press. Doron, E. and U. Shlonsky. 1990. Verb second in Hebrew. In Proceedings of the tenth west coast conference on formal linguistics , ed. Dawn Bates, 431-445. Stanford, CA: University CSLI. Goldberg, Lotus. 2001. On the obligatory nature of verb and subject movement in Modern Hebrew: a comparison of Borer (1995) and Shlonsky (1997). McGill Working Papers in Linguistics 15 (2): 1-35. Haegeman, Liliane & Jacqueline GuÃ©ron. 1999 . English grammar: a generative perspective . Malden, MA: Blackwell Publishers Ltd. Kayne, Richard and Jean-Yves Pollock. 2001. New thoughts on stylistic inversion. In Subject inversion in Romance and th e theory of universal grammar , ed. Aafke C. J.Hulk and Jean-Yves Pollock, 107-162. New York: Oxford University Press. Kiss, Katalin E. 1998. Identificational focus versus information focus. Language 74: 245273. Levon, Erez. 2001. On triggered inversion in Hebrew. Paper presented at ConSOLE X, Leiden University, The Netherlands, December 2001. Levon, Erez. 2002. Subject-verb inversion in Hebrew wh-interrogatives. Unpublished Manuscript, New York University. Lightfoot, David. 1998. The development of language: acquisition, change, and evolution . Malden, MA: Blackwell Publisherd Ltd. Rizzi, Luigi. 1991. The wh-criterion. DÃ©partment de linguistique gÃ©nÃ©rale et franÃ§aise. Technical report. UniversitÃ© de GenÃ¨ve.
52 Rizzi, Luigi. 1997. The fine struct ure of the left periphery. In Elements of grammar: handbook of generative syntax , ed. Liliane Haegeman, 281-337. Boston: Kluwer Academic Publishers. Rothstein, Susan. 1995. Small clauses and copular constructions. In Syntax and semantics 28, ed. A. Cardinaletti and T. Cardin aletti, 27-48. New York: Academic Press. Schwarzwald, Ora Rodrigue. 2001. Modern Hebrew . Muenchen: Lincom Europa. Shlonsky, Ur. 1990. Pro in Hebrew subject inversion. Linguistic Inquiry 21: 263-275. Shlonsky, Ur. 1997. Clause structure and word order in Hebrew and Arabic: an essay in comparative Semitic syntax . New York: Oxford University Press. Szabolcsi, Anna. 1981. The semantics of topic-focus articulation. In Formal methods in the study of language , ed. Jan Groenendijk, Theo Janssen and Martin Stokhof, 513541. Amsterdam: Matematisch Centrum. Taraldsen, Knut Tarald. 2001. Subject extrac tion, the distribution of expletives, and stylistic inversion. In Subject inversion in Romance and the theory of universal grammar , ed. Aafke C. J. Hulk and Jean-Yves Pollock, 163-182. New York: Oxford University Press.
53 BIOGRAPHICAL SKETCH Jodi Resnick received her B.A. from the University of Delaware in French Education with a minor in li nguistics in May 2001. She starte d her graduate studies at in the Program in Linguistics at the University of Florida in Augus t 2001 and will receive her M.A. in August 2003. While at the Univers ity of Florida she has taken courses in the core areas of linguistics as well as Modern He brew. In the fall, Jodi will begin working towards the completion of a Ph.D. in linguistics at Yale University.