University of Southern California Dissertations and Theses

On the status of possessives

(USC Thesis Other)

On the status of possessives

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content INFORMATION TO USERS This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer. The quality of this reproduction is dependent upon the quality o f the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back o f the book. Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6” x 9” black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order. UMI A Bell & Howell Information Company 300 North Zed) Road, Ann Arbor MI 48106-1346 USA 313/761-4700 800/521-0600 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ON THE STATUS OF POSSESSIVES by Nancy Mae Antrim A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fullfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (Linguistics) December 1996 Copyright 1996 Nancy Mae Antrim Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UMI Number: 9720176 UMI Microform 9720176 Copyright 1997, by UMI Company. A H rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. UMI 300 North Zeeb Road Ann Arbor, MI 48103 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UNIVERSITY OF SOUTHERN CALIFORNIA THE GRADUATE SCHOOL UNIVERSITY PARK LOS ANGELES, CALIFORNIA 90089 This dissertation, w ritten by Nancy Mae Antrim under the direction of h&x. Dissertation Committee, and approved by all its members, has been presented to and accepted by The Graduate School, in partial fu lfillm en t of re quirements fo r the degree of D O C T O R O F P H ILO S O P H Y Dean 11-26-97 Date DISSERTATION/ COMMITTEE n rpersort Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. On the status of possessives. Nancy Mae Antrim Co-chairs: Joseph Aoun and Mario Saltarelli The aim of this dissertation is three-fold: to capture both the adjectival and pronominal behavior of prenominal possessives; to provide a syntactic instantiation of the semantics of possession; and to account for the variation with respect to the prenominal possessives in Romance languages. Prenominal possessives have been traditionally analysed as adjectives; however, their behavior is not consistent with either attributive or predicative adjectives. Furthermore possessives demonstrate agreement with the possessor with respect to person and may show agreement with the possessed with respect to number and possibly gender. By taking into account the semantics of possession, the syntactic behavior of possessive constructions can be accounted for. To account for both the syntactic variation in prenominal possessives in Romance and the semantics of the possessive, I propose that the prenominal possessive is projected in an XP as a predicate structure with two semantic roles: possessor and possessed. Generating the possessive as a two-place predicate structure allows us to capture both their pronominal-like behavior with respect Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. to binding, and their adjectival-like behavior with respect to agreement. The adjectival-like agreement of the possessive would reflect the predication relation. This Possessive Phrase has as its specifier the possessor, as its head the possessive and as its complement the possessed noun. Via Spec-Head agreement, the possessive agrees in person with the possessor. The possessive in French, Spanish and Italian raises via head- to-head movement to D to check referenciality and where in French and Spanish the possessive and the definite article merge prior to Spell-out forming a clitic-like element with reduced agreement. In Italian, the possessive also raises to D where it incorporates with D, but does not merge retaining the form article + possessive. The merger of the article and possessive in French and Spanish accounts for the definiteness effects found with the prenominal possessive in these languages. The lack of merger in the case of Italian permits the use of the possessive without the article accounting for the lack of definiteness effects with the possessive. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ACKNOWLBGMBNTS ii Having arrived at the stage where one has the opportunity to reflect on the many people who have been instrumental in bringing one to this stage, it is difficult to know where to start. First I would like to thank all my committee members past and present: Joseph Aoun, Audrey Li, Mario Saltarelli, Jean-Roger Vergnaud, Roger Woodard and Maria Luisa Zubizarreta. Joseph Aoun, one of my co-chairs, was the one who first drew me to USC, having read his articles and books while I was at UTEP, and he has quided me while at USC. He taught me to focus on details and kept my feet firmly planted on the ground. Audrey Li was an inspiration as a teacher with her clear and logical presentations, as well as her patience. Maria Luisa Zubizarreta, although not always agreeing with my proposals, listened and challenged me to convince her; her challenges helped me to clarify my ideas and hopefully to argue them convincingly. Jean-Roger Vergnaud introduced me to morphology and sharpened my vision allowing me to see unimagined possibilities and subtleties. Observing him was like listening to variations on a theme by Philip Glass. Roger Woodard taught me what I know of historical linguistics, adding an important perspective to my research. My other co-chair was Mario Saltarelli. He has Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. iii been a constant and steadying force in my intellectual development. He introduced me to the work done in philology as well as linguistics with respect to the Romance languages. I spent many afternoons in his office discussing not only linguistics, but a myriad of subjects. For discussions and encouragement, dinners and conversation, and e-mail messages, thank you Mario. Aside from members of my committee there were other faculty members at USC who have influenced ny intellectual development. Alicia Gorecka expanded my knowledge of phonology, but more than that she expressed concern about and interest in my life outside the department, which was greatly appreciated. Haijime Hoj i honed my skills at making grammaticality judgements with his intricate English constructions. He also provided a safe haven for several of us during the Los Angeles riots for which I will always remain grateful. All I understand of semantics I owe to Barry Schein. I left his class with an appreciation of logic and a fondness for reading Quine. Ed Finigan and the Sociolinguistics Circle gave me an opportunity to explore areas of linguistics outside of my research; areas that I find increasingly relevant to my teaching. Bill Rutherford not only encouraged me to actively audit his class on second language acquisition, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. iv but also provided suggestions and support when I found myself teaching SLA. As for those I came in contact with while at USC: to Claudia Parodi, Judy Bernstein, Paola Beinca, Mario Montalbetti, Ian Roberts and Giuseppe Longobardi for their comments and contributions to different parts of this dissertation, thank you. The faculty at USC were not ny only mentors. I never would have arrived at this stage had it not been for those at the University of Texas at El Paso who encouraged me. Grant Goodall has been teacher, thesis advisor, mentor and friend. He has always believed in me and to that end provided support and encouragement whether via e-mail when I was in LA or meals and conversations when I was back in El Paso. I hope I have justified his faith in me. Jon Amastae taught me not only phonology, but also not to limit ny possibilities. Ed Blansitt, through his typology and universals class, developed my appreciation for language variation, leading to the focus of this dissertation. I returned to UTEP as a visiting professor while I completed my dissertation. I would like to thank both Richard Ford and Charles Elerick for their help and support in making my first year of teaching successful. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. V I would also like to thank Betty Bamburg and the Freshman Writing Program at USC for providing not only ny financial support, but also for giving me the opportunity to develop my teaching skills in working with non-native speakers of English. My program of study would not have been successful without the support and often encouragement of the department's staff. They made the department office a friendly, welcoming oasis. Thank you Kathy Stubaus for listening and always having an answer for my questions. Thank you Linda Carter for all your help with the copy machine and for taking the time to just chat. And thank you Laura Reiter for keeping me on top of the paperwork, listening to ny rantings and ravings, the laughter, the tears and the Thanksgiving dinner. To all my past and present fellow students at USC for their friendship: Pablo Albizu, Joe Allan, Lina Choueri, Daeho Chung, Abdesdam Elomari, Gorka Elordieta, Ibtisam Kortobi, Karin Megerdoomian, Keiko Miyagawa, Shu-Ing Shyu, and Shin Watanabe for their friendship. To my friends in the Spanish department, especially Debbie Gill and Liliana Paredes for long talks and late night suppers. To Yuki Matsuda for the Japanese lessons, Sunday breakfasts and study sessions, that made our first year enjoyable. To Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. vi Miao Ling Hsieh, who taught me how to make dumplings and shared my acting career, for her friendship, shopping trips, and letting me be a surrgate aunt to her daughter. To Jos6 Camacho and Liliana for their love and support, dinners and Trader Joe's. To Heather Goad and Nigel Duffield who served as inspirations for me and made me feel welcomed in Montreal. To Dong-In for his constant support, for Korean bar-ba-que, trips to the airport and his continued friendship. And to Teeanna Rizkallah, who put me up on my trips back to USC, for Star Trek, videos, and dinners, for allowing me to adopt her cats and for long conversations about everything except linguistics, but especially for her friendship. Finally ny thanks and gratitude to my mother and my children for their love and encouragement without which nothing else would matter. It is to ny mother, Mae Wilkes, and my children Heather Marie, Stephanie Mae, Megan Elizabeth, and Kenneth Edward that this dissertation is dedicated. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. vii TABLE OF CONTENTS Acknowledgements......................................ii Chapter 1 Possessive Constructions................... 1 1.1 Expressing Possession...................... 2 1.1.1 Genitive Constructions................. 2 1.1.1.1 Grammatical Relations Expressed by Possession.................... 3 1.1.1.2 Additional Relations Expressed by Genitive...................... 4 1.1.2 Possessive Forms...................... 7 1.2 Possessive Forms Classified as Adjective or pronoun............................ 11 1.2.1 Spanish................................ 12 1.2.2 French................................. 14 1.2.3 Italian................................ 17 1.2.4 Prenominal Possessives in Romance........19 1.2.5 The Proposal........................... 21 1.3 Organization of Thesis...................... 23 Chapter 2 The Status of Possessives...................29 2.1 The adjectival status of possessives.........29 2.1.1 Attributive Adjectives..................35 2.1.1.1 Predicates......................... 35 2.1.1.2 Elliptical constructions............37 2.1.1.3 Ne/en stranding.....................37 2.1.1.4 Modification....................... 38 2.1.1.5 Status as Attributive Adjectives.... 41 2.1.2 Predicative Adjectives..................42 2.1.2.1 Individual vs. stage level...........42 2.1.2.2 Theta-role assignment...............44 2.1.2.3 Status as Predicative Adjectives.... 45 2.2 Pronominal status of possessives.............46 2.2.1 Binding................................ 46 2.2.2 Weak Crossover Effects..................47 2.2.3 Possessives as Pronouns.................49 2.2.4 Pronouns and Articles...................50 2.2.5 Possessives Adjective or Pronoun?........................ 52 2.2.5.1 Modifiers.......................... 52 2.2.5.2 Coordination....................... 54 2.3 Possessives as Predicates................. 59 2.3.1 Predication............................. 59 2.3.1.1 Williams 1980...................... 60 2.3.1.2 Napoli 1989........................ 62 2.3.2 The Semantics of Predication.............63 2.3.3 The Semantic Representation of Possessives................................ 65 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. viii Chapter 3 The Composition of DP...................... 67 3.1 Previous Proposals for Possessive Structures......................................71 3.1.1 Earlier Treatments..................... 71 3.1.1.1 Jackendoff (1977)................. 72 3.1.1.2 Abney (1987)...................... 78 3.1.1.3 Chomsky (1988).................... 80 3.1.1.4 Lyons (1986)...................... 82 3.1.1.5 Olson (1989)...................... 84 3.1.2 Giorgi and Longobardi (1991)............89 3.1.3 Zubizarreta (1987).................... 98 3.1.4 Tremblay (1991)....................... 102 3.1.5 Valois (1991).........................106 3.2 Decomposing DPs............................109 3.2.1 Projections in DP..................... 109 3.2.1.1 Valois............................ 109 3.2.1.2 Picallo........................... 112 3.2.2 Adjectival Status and Structure........114 3.2.3 Pronominal Status and Structure........117 3.2.3.1 Koopman........................... 117 3.2.3.2 Cardinaletti...................... 119 3.3 Arguments and Predicates ................. 120 3.3.1 Possessives and Argument Structure 121 3.3.1.1 Giorgi and Longobardi............. 121 3.3.1.2 Grimshaw..........................122 3.3.2 Possessives as Predicates............. 123 3.3.2.1 Predication Structure............. 125 3.4 DP Structure...............................129 Chapter 4 Accounting for Variation.................. 131 4.1 Possessive structure....................... 132 4.1.1 Old French and Old Spanish.............133 4.1.2 Inalienable Possession................ 137 4.2 Merger.....................................139 4.2.1 Pronominal Forms...................... 141 4.2.2 Conclus ion............................ 14 6 4.3 Definiteness and the Possessive............ 147 4.3.1 Existential Constructions............. 148 4.3.2 Partitive Constructions............... 152 4.3.3 Further Evidence...................... 156 4.3.3.1 Indefinites....................... 156 4.3.3.2 Emphasis.......................... 157 4.3.3.3 Romanian.......................... 158 4.3.3.4 Marsian........................... 160 4.3.4 Alternative explanations.............. 162 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ix 4.4 Determiners................................ 163 4.4.1 Articles, Possessives and Inalienable Possession..................................167 4.4.1.1 Kinship........................... 167 4.4.1.2 Body parts........................ 171 4.4.1.3 Extensions to English............. 174 4.4.2 Articles, Possessives and Predicates... 178 4.4.2.1 Small Clause Analysis............. 179 4.4.2.2 English Possessives and Predicates................................181 4.4.3 Vocatives............................. 182 4.4.3.1 Expletive Article................. 184 4.4.3.2 English Vocatives................. 185 4.5 Possessive Placement....................... 187 Chapter 5 Cliticization and Merger.................. 193 5.1 Merger.....................................193 5.1.1 Prosodic influences................... 198 5.1.2 Adjective Placement................... 198 5.1.2.1 Dialectical Variation in Word Order................................201 5.1.2.2 Dialectical Variation and Demonstratives............................ 202 5.2 Possessives as Clitics..................... 203 5.2.1 Previous Works........................ 203 5.2.2 The Notion of Clitic.................. 204 5.2.3 Clitic Properties................... 207 5.2.3.1 Emphasis.......................... 211 5.2.3.2 Alternation with strong forms......212 5.2.3.3 Syntactic Independence............ 213 5.2.3.4 Placement......................... 215 5.2.3.5 Coordination...................... 217 5.2.3.6 Phonological Adjustment........... 223 5.2.4 French Prenominal Possessives......... 224 5.2.5 Clitic Parameters..................... 229 5.2.6 Italian Possessives................... 233 5.2.7 Possessive Clitics and Structure.......238 5.3 Summary....................................239 Bibliography........................................242 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter l Possessive Constructions Possessives have long been a topic of discussion in the study of language. Recent articles within the framework of generative grammar have focused on the syntax of possessive constructions (Abney 1987, Tellier 1988, Valois 1991, Giorgi and Longobardi 1991, Tremblay 1991, Picallo 1994, among others). Assuming the traditional grammar classification of possessives, they have treated possessive forms as either pronouns or adjectives. However, in so doing previous accounts have been unable to fully account for the distribution or behavior of these possessive elements crosslinguistically. I propose that in order to fully account for possessive forms their status as either adjectives or pronouns (or possibly both) must be taken into account. But before considering their classification we need to look at the various ways possession can be expressed in order to reach a clearer understanding of these constructions. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 1.1 Expressing Possession Languages differ internally as well as crosslinguistically in the means employed to express possession. 1.1.1 Genitive Constructions Perhaps the first distinction that needs to be made is that between possessive and genitive. Genitive constructions involve case marking of a grammatical relation either by preposition or morphological case, as illustrated in (l) for French and German, respectively. l. a. le livre de Pierre the book of Peter Peter's book b. die Schussel des Hundes the dish the-gen dog-gen the dog's dish c. Johanns Fahrrad John-gen bicycle John's bicycle Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 1.1.1.1 Grammatical Relations Expressed by Possession The grammatical relationships that can be denoted by possession (Agard, 1984) include kinship, property/attribute, product/activity/thought, action by the modifier and action upon the modifier. These are illustrated for French in (2). 2.a. le p&re de 11 enfant (kinship) the father of the child b. le restaurant de la gare (property/attribute) the restaurant of the station c. l'opinione du doyen (product/activity/thought) the opinion of the dean d. les pri&res des fiddles (action by the modifier) the prayers of the faithful e. 1 *6stablissement de la loi (action upon the modifier)1 the establishment of the law The wide variety of relations that can be expressed by possession, as well as the vagueness of possessives, led Williams (1982) to propose the following Det Rule. 1 The examples are from Agard, 1984. The glosses are my own. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 4 3. The Det Rule The relation between the possessive NP and the following N' can be any relation at all. Although this is clearly not the case, as can be seen with respect to the part-whole relationship expressed with the possessive in English. 4.a. the ship’s rudder b. *the rudder's ship In (4a) the relationship of the part to the whole entails a possessive interpretation (i.e. the rudder that belongs to the ship) ; however, this possessive interpretation is not possible with the inverse relation in (4b) .2 l.l. 1.2 Additional Relations Expressed by Genitive However, the genitive construction can express relationships other than possession. Curme (1952) in discussing the grammar of German notes seven other 2 For a detailed discussion of the semantics of possessives see Barker, 1995. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5 relations expressed by the genitive. These are illustrated for German in (5) .3 5. a. die Nachkommen Abrahams (genitive of origin) the descendants Abraham-gen Abraham's descendants b. die Ruge des Lehrers (subjective genitive) the reproof the-gen teacher-gen the reproof of the teacher c. die Erziehung der Kinder (objective genitive) the education the-gen-pl children the education of the children d. ein Dach schattender Buche (genitive of material or composition) a roof shady beech-gen-pl a roof of shady beeches e. der G6tt der Liebe (descriptive genitive) the god the-gen love the god of love 3 To illustrate this point, I used German examples because German nouns are clearly marked for genitive case; whereas, the majority of languages that will be discussed in this dissertation do not employ overt genitive case. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 f. das Laster der Trunksucht (appositive genitive) the vice the-gen intemperance the vice of intemperance g. der erste Vers des Liedes (partitive genitive) the first verse the-gen song-gen the first verse of the song4 In (5a) the relationship is one of source, cause or authorship. The subjective genitive shown in (5b) "represents a living being as associated with an action in the relation of author" (Curme:478). The objective genitive in (5c) , as the name suggests, indicates the object of some activity. While not common to prose, the genitive of material and composition, shown in (5d) functions poetically. Curme reports that this genitive is instantiated with von (of) in prose. Example (5e) illustrates only one of the descriptive genitives, that of quality or characteristic. Under descriptive genitive there is, also, the genitive of measure. The appositive genitive in (5f) acts, as its name implies, by expanding upon the word it modifies. Finally, the partitive genitive in (5g) indicates a part of a collective whole. 4 The examples are from Curme, 1952. The glosses are my own. The names for these various genitive functions are those of Crume. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 Since genitive constructions permit more interpretations than what have been called either possessive adjectives or pronouns, I will not address the syntax of these constructions, leaving that, for the most part, to future research. Nor will I consider prepositional constructions as possessives. Following Barker (1995) , I consider the prepositional phrase as in (2a) a syntactic argument of the noun rather than a possessive construction, although the relationship expressed is one of possession. 1.1.3 Possessive Forms In addition to prepositional phrases and genitive case marking, languages may employ prenominal and/or postnominal possessive forms, as shown in (6) for Italian. 6. a. il mio libro the my book b. il libro mio the book try These are generally considered, in traditional grammars, / to be adjectives (Tekavfic, 1972; Lausberg, 1965) . Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 The interpretation of these possessives is more limited than that of the genitive constructions. The relationships expressed in (2) can be expressed with a possessive, as shown in (7). 7.a. son p4re (kinship) his (the child) father b. son restaurant (property/attribute) its (the station) restaurant c. son opinion (product/activity/thought) his (the dean) opinion d. leurs pri&res (action by the modifier) their (the faithful) prayers e. son etablissement (action upon the modifier) its (the law) establishment However not all the genitive relations given in (5) can be represented by the possessive form. Only four of these relations (genitive of origin, subjective genitive, objective genitive, and appositive genitive) can be expressed with the possessive forms. Moreover, a possessive used in an appositive construction is ambiguous between possession and apposition. A possessive cannot be Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9 used to express the genitive of composition or a descriptive genitive, as shown in (8) . 8. a. ein Dach schattender Buche: *sein Dach a roof shady beech - gen-pl: *its roof a roof of shady beeches b. der Gdtt der Liebe: *sein Gdtt the god the-gen love: *its god the god of love In (8a), the possessive refers to the relationship between a part of the roof) and a whole (what the roof is attached to), not the composition of the roof. In (8b), the possessive refers not to the nature of a god, but to the relationship of a god to a religion or a people. As for the partitive genitive, it has been suggested (Zubizarreta, p.c.) that the postnominal possessive, at least in Spanish, may have a partitive interpretation.5 While there is a considerable amount of overlapping between genitive constructions and possessive forms, the possessive forms are narrower in their interpretation. 5 This will be discussed more extensively in chapter 4. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 10 Despite the unavailability of some genitive interpretations for the prenominal possessive, Olson (1989) argues that the prenominal possessive, in German, is a genitive case-marked personal pronoun. This characterization, however, fails to take into account the adjectival nature of the possessive; although, she does derive the agreement features from the structural position of the possessive. Cornilescu (19 93) looking at the assignment of genitive case in Romanian observes that the genitive case marking captures a variety of theta-roles in addition to its possessor role. These are illustrated in (9) , her examples (34-35). 9. a. tradarea cauzei (Theme) the betrayal of the cause b. tradarea lui luda (Agent) Juda's betrayal c. surpriza lui Ion (la vederea musafirilor) (Experiencer) John's surprise (at the sight of the guests) d. cartea lui Ion (Possessor, alienable possession) John's book Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 11 e. surisul Giocondei (Possessor,inalienable possesssion) Gioconda's smile f. floarea cimpului (Locative) the flower of the field To account for genitive case assignment she proposes, based on control facts, an agreement phrase where genitive case is assigned outside of NP. The possessor phrase, she assumes, albeit tentatively, is generated in the Spec NP position. This separates genitive case from possession syntactically suggesting that the genitive is a morphological case relation which may not imply a semantically coherent reading. This is consistent with the use of genitive case in Latin where it is diacritically required with some verbs (i.e. memini "I remember'', obliviscor "I forget", misereor "I pity") . (Allen, 1970) 1.2 Possessive Forms Classified as Adjective or Pronoun Traditionally possessives have been classified as either possessive adjectives or possessive pronouns (Lausberg, 1965; Grevisse, 1969; Tekav££, 197 2; among others). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 12 1.2.1 Spanish Spanish possessives have two forms: an adjectival form that is unstressed and appears only before the noun; and a pronominal form that is considered a "full" form, as it shows agreement in both number and gender, while the shorter form shows agreement only in number, except for first and second person plural, as shown in the following chart (Butt and Benjamin 1988) . This agreement is with the item possessed, not the possessor. Although considered adjectival, these forms do not occur with determiners SPANISH Short Form Masculine Feminine Singular 1st 2nd 3rd Plural 1st nuestro/s 2nd vuestro/s 3rd mi/mis tu/tus su/sus su/sus nuestra/s vuestra/s Eull .Eom Masculine Feminine Singular 1st 2nd 3rd mio/mios tuyo/tuyos suyo/suyos mia/mias tuya/tuyas suya/suyas Plural 1st 2nd 3rd nuestro/s vuestro/s suyo/suyos nuestra/s vuestra/s suya/suyas Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 13 While these "full" forms occur separately, functioning as nouns, they can also occur postnominally functioning as adjectives. The full form, functioning as a pronoun, requires the addition of the article after prepositions. Compare (10a) and (10b). 10. a. dDe quiin es el coche? Mio. of who is the car my-masc-sg Whose car is it? Mine, b. dEn qu£ coche vamos? En el mio. in what car go-3rd ps pi in the my-masc-sg Which car are we going in? In mine. The article is also required when the possessive is either the subject or object of a verb, as shown in (11) .6» 7 6 The examples in both (8) and (9) are from Butt and Benjamin, 1988. 7 The pronominal form may be an instance of ellipsis; then, the form would be el [e] nuestro analogous to (i). i. Mis videos son buenos, pero los [e] tuyos no. This form will be discussed in chapter 4 section 4.2.1. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 14 11. a. Los dos videos son buenos, pero el nuestro es mejor. the two videos are good, but the our is better the two videos are good, but ours is better b. Coge el mio. take the my-masc-sg take mine Finally, the article, also, occurs with respect to emphasis. The emphatic use of the possessive will be discussed in more detail in chapter 4, section 4.3.3.2. 1.2.2 French The possessives considered adjectives in French "qualify" a noun being structurally dependent on the noun. These forms always precede the noun they modify. They agree in number and gender with the thing possessed, not the possessor in the singular, while they agree in number only in the plural, as seen in the following chart (Judge and Healey, 1985). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 15 FRENCH Masculine Feminine sg/cl ag/pl Singular 1st mon/mes 2nd ton/tes 3rd son/ses Plural 1st 2nd 3rd notre/nos votre/vos leur/s ma/mes ta/tes sa/ses The situation in French is not the same as in Spanish with respect to agreement. In Spanish gender agreement is formed with an inflectional morpheme: -a for feminine; -o for masculine. This inflectional morpheme extends to other adjectives. Gender agreement in the French prenominal possessives is not consistent with the general rule of agreement found with adjectives in which the feminine is formed by adding -e to the masculine form. Additionally the masculine form is used in front of singular feminine nouns beginning with a vowel or silent /h/, as in (12b). Reminiscent of the use of the masculine form of the article with stressed initial vowels words in Spanish, shown in (12a) . 12. a. el alma Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. the-ms soul-fm 16 b. mon amie my-ms friend-fm There is both a stressed and an unstressed form. Only the unstressed forms are used in prenominal position. Moreover, these unstressed adjectival forms, like their Spanish counterparts, do not occur with a determiner. The stressed form is an earlier form and now considered archaic. These forms are illustrated in the following chart (Judge and Healey, 1985). Stressed Forms Masculine Feminine __________sg/pl_________________________________sgy.pl___ Singular 1st mien(s) mienne(s) 2nd tien(s) tienne(s) 3rd sien(s) sienne(s) Plural 1st n6tre(s) 2nd vdtre(s) 3rd leur(s) Possessives functioning as pronouns replace the possessive adjective + noun8. They appear with the definite article, as shown in the following chart: 8 This is the traditional interpretation. An alternate possibility is that the "so-called" pronominal form is an elliptical construction in which the noun has been deleted leaving the article and an adjectival possessive. Note that this form follows the general adjective gender agreement rule. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 17 Masculine sct/d 1 FRENCH Feminine sct/d I Singular 1st 2nd 3rd le mien/les miens le tien/les tiens le sien/les siens la mienne/les miennes la tienne/les tiennes la sienne/les siennes Plural masculine feminine Dlural 1st 2nd 3rd le notre le votre le leur la notres la votre la leur les notres les votres les leurs Morphologically the form of the possessive pronoun is the same as the stressed adjective form with the addition of the article. 1.2.3 Italian Unlike Spanish and French, the forms for both the possessive pronoun and the possessive adjective are the same in Italian. These forms agree in number and gender with the possessed and require a determiner with common nouns, as shown in the following chart: Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 18 ITALIAN ------------Masculine_____________________Feminine Singular 1st il mio/i miei la mia/le mie 2nd il tuo/i tuoi la tua/le tue 3rd il suo/i suoi la sua/le sue Plural 1st il nostro/i nostri la nostra/le nostre 2nd il vostro/i vostri la vostra/le vostre 3rd il loro/i loro la loro/le loro These possessive forms have been traditionally- described, as noted, as either adjectives or pronouns.9 9 While this dissertation is primarily concerned with Italian, French and Spanish as representative of Romance languages, the possessive patterns presented are also found in other Romance languages. In Romanian the possessive always appears in conjunction with a definite article, as shown in (i) . i. E cartea mea. be-3s book-det my-fs It is my book. (Mallinson, 1986) The definite article in Romanian is a suffixed clitic. Catalan requires a determiner when the possessive precedes the noun and may have one when it follows, as seen in (ii). ii. a. un teu projecte or un projecte teu a your plan b. el teu projecte or el projecte teu the your plan (Wheeler, 1988) In Sardinian the possessive always requires a determiner except with kinship relations or terms of endearment, as in (iii) . iii. a. su libru meu the book my b. unu libru meu a book my (Jones, 1988) Occitan and Rhaeto-Romance appear to follow the French/Spanish pattern; whereas Portuguese follows the Italian pattern (Harris and Vincent, 1988). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 19 1.2.4 Prenominal Possessives in Romance Prenominal possessives in the languages under discussion (e.g. French, Spanish and Italian) occur in one of two ways. In Spanish and French, the prenominal possessive is a reduced form preceding and appearing with a noun: Possessive + Noun. In Italian, the prenominal possessive is preceded by a determiner, which is required in the case of common nouns: Article + Possessive + Noun. This variation remains to be fully accounted for despite previous attempts to do so (Giorgi and Longobardi 1991, Zubizarreta 1987, among others). There are other notable differences between these two instanciations of the possessive. First the degree to which agreement is reflected in the possessive. The Italian possessive shows full agreement as to gender and number with the item possessed for all persons. The French prenominal possessive shows agreement with respect to number in all forms, but as to gender only in the singular. Gender agreement is lacking in the plural of first - third person singular, as well as, in first - third person plural. The Spanish prenominal possessive has number agreement for all forms, but shows no gender agreement for first - third person singular or third person plural. I will argue that the variation with Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 20 respect to agreement is a reflection of the degree to which the prenominal possessive behaves like a clitic in each of these languages. A further point of contrast between the prenominal possessive in these languages involves definiteness effects. The French and Spanish prenominal possessives behave as + definite with respect to existential environments; whereas, the Italian possessive does not behave as + definite. This definiteness, I will argue, is a result of a merger of the possessive with the determiner resulting in the possessive acquiring a definiteness feature from the determiner in French and Spanish. As Barker observes, prenominal possessives appear to behave semantically like both definites and indefinites. Like indefinites, possessives can introduce novel/new participants in a discourse (13b) and, on the other hand, like definites they can refer to know/old participants (13c). 13.a. A student walked into the room. b. She had her daughter with her c. She was nervous and her daughter was crying. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 21 With respect to discourse, it would appear that possessives cannot be classified as definite or indefinite. However, by definiteness, I mean, following in the Russellian tradition, the presupposition of uniqueness; but not necessarily a presupposition of identif iability, allowing then for a nonspecific interpretation as well as a specific one.10 1.2.5 The Proposal To account for both the syntactic variation in prenominal possessives in Romance and the semantics of the possessive, I propose that the prenominal possessive is projected in an XP as a predicate structure. I propose that the possessive is a two-place predicate with (along the lines of Napoli 1989) the semantic roles: possessor and possessed. Generating the possessive as a two-place predicate structure allows us to capture the semantics of possession. It further accounts for the adjectival-like agreement but pronominal-like behavior of the possessive. In generating the possessive as an XP, I am assuming standard X-bar theory in which a phrase has a head, a specifier and a complement, as in (14) . 10 cf. Jackendoff 1972. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 22 14 XP / \ Specifier X' / \ X YP Head Complement I will also assume the DP-Hypothesis proposed by Abney (1987), in which he proposes that Noun Phrases are selected by a Determiner Phrase. In looking at English, Abney proposes that both a determiner and a genitive case marker can occur as heads of DP. I will depart from this assumption, as it weakens X'-theory permitting two different elements as heads of the same category. I propose that the Possessive Phrase is generated as an XP, as shown in (15) 15 D1 D PossP = XP P Poss' P o o s NP Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 23 The Possessive Phrase has as its specifier the possessor, as its head the possessive and as its complement the possessed noun. The possessive behaves differently and is generated in an XP because of the number of its semantic roles, via Spec-Head agreement, the possessive agrees in person with the possessor. The possessive in French, Spanish and Italian raises via head- to-head movement to D to check referenciality, following Chomsky (1992), where in French and Spanish the possessive and the definite article merge prior to Spell-out forming a clitic-like element with reduced agreement. In Italian, the possessive also raises to D where it incorporates with D, but does not merge retaining the form article + possessive. The merger of the article and possessive in French and Spanish accounts for the definiteness effects found with the prenominal possessive in these languages. The lack of merger in the case of Italian which permits the use of the possessive without the article accounts for the lack of definiteness effects with the possessive. 1.3 Organization of Thesis In order to fully explain the variation in possessive constructions crosslinguistically it is necesary to take into account the dual nature, requiring adjectival-like Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 24 agreement and having pronominal reference, of these forms. Combining this with their varying degrees of cliticness, we will be able to explain the variation in these forms. By projecting the possessive as a predication structure, we are able then to capture the semantics involved. In chapter two the adjectival and pronominal nature of the possessives will be discussed. Despite their failure to satisfy the general properties of adjectives (i.e. function as an attribute, function as a predicate, modification by an intensifier and occur in comparative and superlative forms), possessives have been treated as adjectives within the generative framework (Giorgi and Longobardi 1991, Bernstein 1993, among others). Looking at possessives from this perspective it will be shown that possessives behave as adjectives with respect to agreement only. Considering them as pronouns by looking at their binding properties and weak crossover effects, they appear to be more pronominal in behavior; however to capture both the agreement facts as well as their pronominal reference, I will argue that possessives should be viewed as two- place predicates. Following Napoli's (1989) distinction between semantic and thematic roles, the possessive has two semantic roles: possessor and possessed. The Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 25 adjectival-like agreement reflects this predication relation. In chapter three I will discuss the composition of DP as proposed by earlier analyses. These analyses have argued for the syntax of the possessive based on its status as an adjective (Tremblay 1991, Bernstein 1993, Giorgi and Longobardi 1991, among others), as a determiner (Abney 1987, Giorgi and Longobardi 1991), as a pronoun (Olson 1989), as an argument (Giorgi and Longobardi 1991), as an a-adjunct (Grimshaw 199 0), and as an adjunct (Zubizarreta 1987) . All of these previous analyses, while contributing to our understanding of the complexity of the possessive, have failed to provide a full explanation of the variation found with respect to the possessive. These analyses are inadequate because of this very complexity. The possessive appears adjectival because of its agreement, but adjectival analyses are unable to fully account for the possessive. Taking the pronominal aspects into account provides a fuller picture, but the adjectival-like agreement must still be accounted for. By considering the possessive as a two-place predicate, incorporating the semantics of predication into the syntactic structure, we are better able to account for the behavior of the possessive. I will argue for the base Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 26 generation of the possessive in a PossP, where the possessive is the head of its own projection. The possessive is projected in the Head of POSSP as a predicate. In chapter four, drawing on data from Old French and Old Spanish, I propose that the basic prenominal possessive construction in Romance is article + possessive + noun. The close relationship between the definite article and the possessive in French and Spanish is examined with respect to inalienable possession. I propose that the variation between French-like possessive constructions and Italian-like possessive constructions can be accounted for within the structure proposed in chapter three by the merger of the possessive with the article. This will be demonstrated with respect to definiteness. Looking at existential and partitive constructions, as well as, indefinites and emphasis, it will be shown that French and Spanish prenominal possessives demonstrate definiteness effects; whereas, the Italian prenominal possessive does not. Finally in this chapter we will consider the environments in Italian where the possessive occurs without the article: inalienable possession, predicates and vocatives. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 27 In Chapter five, I will argue that the same configuration occurs in French and Spanish. However, through a process of cliticization the article and possessive have mergered morphologically and cliticized to the noun in these languages. Following Marantz (1988), morphological merger will be discussed. This morphological merger results in the reduced prenominal form occuring in D position in French and Spanish at Spell-out. The relationship between the article and the possessive in Italian results from a syntactic incorporation of the article and possessive yielding the Spell-out form: article + possessive. The reduced prenominal possessive by virtue of its merger with the article acquires clitic-like properties. Others, notably Valois (1991) and Authier (199 2) for French and Rivero (1986) for Spanish, have suggested clitic status for the prenominal possessive. Their status as clitics will be considered in light of the properties attributed to verbal clitics as shown by Jaeggli (1982) and Kayne (197 5). Further distinguishing between the phonological, syntactic, morphological and semantic properties of clitics, along the lines of Wanner (1987), the degree to which these prenominal possessives are clitic-like will be discussed, with respect to these Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 28 properties, Spanish and French prenominal possessives are clitic-like; whereas, the Italian prenominal possessive is not clitic-like. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 29 Chapter 2 The Status of Possessives As we saw in Chapter 1 possessives have been traditionally classified as adjectives or pronouns. In this chapter we will look at the properties of possessives with respect to this classification. 2.1 The adjectival status of possessives To be considered as an adjective, in the traditional sense of the word, four properties are considered. First, an adjective can occur within a noun phrase functioning as an attribute, as in "a tall tree". Secondly, adjectives can occur as predicates, as in "the tree is tall". Moreover, adjectives can be modified by an intensifier, as in "the very tall tree". Note, however, that the number of intensifiers able to modify a possessive adjective are limited. Finally, adjectives can occur in comparative and superlative forms, as in "a taller tree, the tallest tree". Not all adjectives have all four properties which Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 30 makes the classification of adjectival-like forms problematic. The possessives forms in both Romance and Germanic languages exhibit these properties to varying degrees suggesting that either they are not true adjectives or that the category is gradient and fuzzy. This is contra the assumption that generative grammar has operated on.1 Adjectives appear to vary in their interpretation with respect to their position. Prenominal adjectives are generally considered attributive, while postnominal ones are predicative. Bolinger (1967) looking at English argues against deriving attributive adjectives from relative clauses. He observes that prenominal adjectives have a sense of customariness; whereas, postnominal adjectives exhibit a sense of temporariness, as shown in (1) . 1. a. a responsible man b. a man responsible for . . . (Bernstein 1993) c. un viejo amigo a friend of longstanding 1 For a discussion of fuzzy categories see Comrie (1981), Corrigan, Eckman and Noonan (1989), Matthews (1981), Taylor (1989) among others. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 31 d. un amigo viejo an aged friend In (la) responsible carries the meaning of trustworthy; while, in (lb) responsible means accountable. In Romance languages certain adjectives can occur both pre- and postnominally with a difference in meaning. The prenominal position giving a nonrestrictive reading as opposed to a restrictive reading for the postnominal position, as shown in (2) . 2. a. las olorosas flores (nonrestrictive) b. las flores olorosas (restrictive) the fragrant flowers (Bernstein 1993) In prenominal position olorosas entails an inherent property of flowers; however, in postnominal position olorosas is limited to a particular set of flowers. This can be clearly seen within the context of a sentence, as in (3). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 32 3. Las flores olorosas valen mds. The flowers fragrant cost more. The fragrant flowers cost more. Here the meaning is the same as a restrictive relative clause, which would restrict the flowers under consideration to only those which are fragrant. This implies that there are flowers which are not fragrant. The prenominal position refers to a property of flowers not to the existence of fragrant and non-fragrant varieties. While olorosas has the same semantic and lexical interpretation in these contexts, this is not the case with all adjectives. Consider, for example, the adjective maravillosas "marvelous" which does not show the same entailment. And as seen in (lc-d), the pre/post position can entail differ semantic interpretations. Interpretations according to position for adjectives do not appear to apply for possessives, except in the cases of Spanish and Italian. Apparently, there is no difference in reading between pre- and postnominally positioned possessives. The postnominal position of the possessive can be used for emphasis, but this does not entail a difference in the reading, rather it entails an intensification of the reading. In Spanish, however, there Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. does appear to be a difference in meaning, as given in (4). 4. a. Llamaron a mis hermanas. They called to my-pl sisters. b. Llamaron a hermanas mias. they called to sisters my-fm-pl. c. des soeurs & moi some sisters to me some of my sisters In (4a), for some speakers, the intended reading is they called all my sisters; whereas in (4b) the reading is they called some people who happened to be my sisters. A reading, that in French, would be accounted for by using the dative construction shown in (4c) . For other speakers, the postnominal position has an indefinite rather than a logical quantification reading. (4b) may also have a partitive interpretation as in the French example (4c) . While noting a difference between possessives and other adjectives in respect to placement in Italian, Longobardi claims postnominal possessives have only a Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 34 contrastive specification.2 However, there is an apparent partitive reading with respect to the postnominal position and it may be this reading which supplies the contrastive focus. A contrastive reading singles out a given subset of a group as opposed to another subset. A partitive reading also selects a subset of a larger set. Consider the Italian sentences in (5) . 5. a. Hanno chiamato le mie sorelle. they called (the-all?) my sisters b. Hanno chiamato delle sorelle mie per partecipare alia sfilata they called of the sisters my to participate in the parade They called some of my sisters to participate... Saltarelli (p.c.) notes the possibility of (5a) being interpreted as all the sisters; whereas, (5b) has a partitive interpretation. 2 Longobardi USC class lecture notes: 15 Sept. 1994. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 35 2.1.1 Attributive Adjectives While position may not provide any insight into the adjectival nature of possessives, it may be possible to maintain an adjectival account for possessives by employing the characteristics of a particular class of adjectives. Bernstein (1993) discusses four characteristics of "regular attributive" adjectives.3 2.1.1.1 Predicates First, they may always appear predicatively. This is the case with possessives in Italian, Portuguese, and Spanish (for the full form) , as well as the independent English possessive, as seen in (6) , respectively. 3 The terms attributive and predicative are used here with the most general sense of use. An attributive adjective is one that modifies a noun as in (i). i. a big; house A predicative adjective is a property predicated of the noun and is found in a predicate structure, as in (ii). ii. The house is biff. Note that with some adjectives they can be used both attributively and predicatively. Traditionally, attributive adjectives precede the noun. Semantically they entail what the speaker attributes to the noun; whereas, postnominal predicative adjectives describe the noun. See Quirk et al. 1985; Judge and Healey, 1985; Butt and Benjamin, 1988. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 36 6. a. Questo denaro i mio. (It) This money is mine b. 0 livro 6 meu. (Po) The book is mine c. Es el mio (*mi). (Sp) (It) is the mine d. This book is mine. While the possessive may occur as a predicate, in French after itre, "be", it functions to emphasize a distinction, as in (7) . 7. Ce livre-ci est le mien et celui-l& est le tien. This book is mine and that one is yours. In non-contrastive predicative environments, the possessive is represented by itre + k + a disjunctive pronoun, as in (8) . 8. a. Ce livre est k lui. This book is to him. b. *Ce livre est le sien. This book is the his. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 37 2.1.1.2 Elliptical constructions Secondly, "regular" attributive adjectives can appear in elliptical nominal constructions. Possessives in the Romance languages under discussion do not occur in these constructions, as shown in (9). 9. a. *Ella es mi propria. (Sp) She is my own b. *Elle est ma propre. (Fr) c. *Ella e mia propria. (It) d. *Ela £ miriha propria. (Po) Both forms of the English possessive can occur in elliptical constructions, as shown in (10). 10. a. If you need a car, you can use my own.4 b. If you need a car, you can use mine. 2.1.1.3 Ne/en stranding Thirdly, these adjectives are stranded by ne/en. This is the case in Italian, as shown in (11) . 4 There seems to be a varying degree of acceptability for ($l0a) among English speakers. However, (i) is acceptable. i. If you need a car, you can use your own. Here "own" may be anaphoric rather than an intensifier. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 38 11. Ne ho visto uno mio. of them I have seen a my However, it is not the case in French, as seen in (12) . 12. a. *J'en ai trouv6 une mienne. I of them found a mine Of them, I found one of mine. b. *J’en ai trouv6 ma. I of them have found my c. J'en ai trouv£ ion I of them found one I found one of them 2.l.l.4 Modification Finally, these adjectives can be modified by an intensifier. This appears to be the case in the Romance languages, under discussion, where the possessive appears prenominally, as shown in (13) . 13. a. mi propria hermana (Sp) my own sister b. la mia propria sorella (It) c. a minha prbpria irma (Po) d. ma propre soeur (Fr) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 39 This is not the case for postnominal possessives in Spanish, and Italian if the intersifier is adjacent to the possessive, as seen in (14). Recall that French does not permit postnominal possessives. 14. a. *la hermana propria mia (Sp) b. *la sorella propria mia (It) However, both Spanish, to some extent, and Italian allow for prenominal modification with a postnominal possessive, as shown in (15). 15. a. la stessa sorella mia (It.) the same sister my my own sister b. ? la misma hermana mia (Sp.) the same sister my Portuguese, on the other hand, does allow postnominal modification with a postnominal possessive, as in (16a), but not further modification, as in (16b) . Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 40 16. a. a propria irma minha b. *a minha muito propria irma English possessives appear to be intensified by- modification. The possessive can be modified by own,5 which provides two possible meanings, as shown in (17) . 17. a. This is my own copy of LGB. b. Maki prepares her (own) recipes. In (17a) own implies mine and not anyone else's. In (17b) the addition of own emphasizes coreference with the subject. This raises the question of whether own is an intensifier or, in fact, an anaphor. Ommitting own makes coreference possible but not necessary. Own cannot modify the independent possessive; however, the possessive plus own can occur independently, as seen in (18) . 18. a. *mine own b. I brought my own. 5 Own can also be used with the nominal genitive, i. That’s Maki's own recipe. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 41 Note, also, that own follows the possessive while adjectival modifiers such as very precede the adjective. If, following Giorgi (1983) , propria, own, is an anaphor then possessives may not be modified and would fall under Bernstein's mero class of adjectives. 2.1.1.5 Summary: Status as Attributive Adjectives To summarize, possessives in Romance are not consistent with the characteristics of "regular" attribute adjectives. This can be seen clearly in the following chart: Regular .Attributive JV3iec.tiv.es Predicate Elliptical ne/en intensifier ----------------------------------------ere_______post Fr no no no ? no Sp yes no N/A ? no It yes no yes ? no Po yes no N/A ? yes En pre- no yes N/A ? iud yes_______yes_______n/a________________ ae Note that the possessive traditionally classified as a pronoun behaves like an attibutive adjective in respect to predicate and elliptical constructions. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 42 2.1.2 Predicative Adjectives If possessives are not consistently attributive, do they function predicatively, as in Italian and Portuguese, shown here in (19)? 19. a. Questo denaro & mio. (It) This money is mine, b. 0 livro e meu. (Po) The book is mine 2.1.2.1 Individual vs. stage level If they function predicatively, the question arises as to whether they are stage level or individual level predicates. Diesing (1992) in her discussion of indefinites suggests several properties that aid in delineating these two types. Traditionally, stage level predicates have been considered temporary states; whereas, individual level predicates have been considered permanent states. Possession would appear to vary in this respect depending on the item possessed. Inalienable possession would by definition be a permanent state. Although possible extensions of inalienable possession, such as suggested by Vergnaud and Zubizarreta (1992) could have a temporary status. This destinction proves problematic, in Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 43 the same way that the temporary-permanent distinction proved problematic for Diesing. She provides additional properties involving both semantic and syntactic distinctions. Following Diesing, stage level predicates permit there-insertion, while individual level predicates do not. This is related to the distinction between having an event argument with stage level and the lack of an event argument with individual level. The lack of an event argument provides for a generic reading but not an existential reading. Individual level predicates allow only the generic reading; whereas, stage level predicates allow either a generic or an existential reading. Spanish, French and English do not allow there- insertion, as seen in (20a-c). Italian and Portuguese, on the other hand, do permit there-insertion, as seen in (20d-e) .6 6 An. explanation for the non-occurrence of the French and Spanish prenominal possessive in existential constructions will be discussed in chapter 4. It will be argued that in Spanish and French the possessive and the article are conflated giving the possessive a definiteness feature, in Italian and Portuguese, on the other hand, the prenominal possessive does not conflate definiteness as evidenced by the separate use of the article. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 44 20. a. *Hay mis hermanas en el parque. There are my sisters in the park. b. *11 y a mes soeurs qui arrivent. There are my sisters that arrive. c. *There is my sister in the garden. d. Ci sono sorelle mie che arrivano domani. There are sisters my that arrive tomorrow There are sisters of mine that are arriving tomorrow. e. Tern amigos meus que gostam de fumar. have friends my that like to smoke There are friends of mine that like to smoke. In (20c) while not accceptable with an existential reading is acceptable with a list reading. 2.l.2.2 Theta-role assignment A further distinction relates levels of predication to the distinction between raising verbs and control verbs. Stage level predicates, like raising verbs, do not assign a theta-role. Individual level predicates, like control verbs, do assign a theta-role, meaning "has the property X". Applying this to possessives draws a Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 45 distinction between possession and inalienable possession. Kratzer (1989) may provide some insight into this. Kratzer analyzes a group of predicates as lacking an event argument and not assigning a theta-role. Included in this group, which Kratzer calls individual-level unaccusatives, is the predicate: belongs to. One might assume following Kratzer that possessives, since they denote the relation belongs to, are included in this group. 2.1.2.3 Status as Predicative Adjectives Moreover, following Bernstein (1993), adjectives that occur predicatively may be modified and may occur in elliptical constructions. This is not the case as has been shown for these possessives. These properties of predicative adjectives, as they apply to possessives, are summarized in the following chart. Predicative Adjectives Modified Elliptical There- Event _________________________insertion______ Theta- _Role Fr Sp It Po En no yes yes yes no no no no no no yes yes no no no no no no no no pre- yes ind no yes no no. no no. no no Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 46 I suggest that the reason the possessives do not pattern as adjectives is because of their pronominal status. Semantically possessives both denote a state and a reference. They denote a state of possession and as denotators of a state, they are semantically adjectives. They denote a reference in that they refer to a noun that they substitute for and as such are pronominal. 2.2 Pronominal status of possessives Valois (1991) notes the pronoun like properties of the French possessive "adjective".7 These properties are also evident in Spanish, Italian, Portuguese, and English. 2.2.1 Binding First, the possessive "adjective" can be bound by a c-commanding QP, as shown in (21). 21. a. La photo de chaquei photographe de sai ville pr6fer6e. (Fr) The picture of each photographer of his favorite town. 7 The French, examples are from Valois (1991). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 47 b. La foto de cadai fotografo de sui ciudad favorita. (Sp) c. La fotograf ia di ognii fotografo della suai citta favorita. (It) d. A foto de cadai fotografo da suai cidade favorita. (Po) e. The picture of eachi photographer of hisi favorite town. The second property possessives have in common with pronouns is the ability to bind a reflexive, as shown in (22) . The binding of reflexives is not possible with referential adjectives. 22. a. Soni portrait de lui-memei his picture of himself b. Sui foto de el mismoi c. La suai fotografia di s6 stessoi d. ? a suai foto de si mesmoi e. hisi picture of himselfi 2.2.2 Weak Crossover Effects A further test for pronominal status involves weak crossover effects. The ungrammaticality of wh-sentences (Fr) (Sp) (It) (Po) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 48 with the wh-trace co-indexed with a preceding pronoun has been accounted for under the leftness condition (Chomsky, 1976; Koopman and Sportiche, 1982). The ungrammaticality decreases in the case of a non c-commanding pronoun as in (23) . 23. a. Whoi loves hisi mother? b. *Whoi does hisi mother love ti? Possessives in the languages under discussion vary in their behavior with respect to WCO, as seen in (24). 24. a. Quemi ama a suai mae? (Po) *Quemi a suai mae ama ti?8 b. Quii sai m6re aime-t-elle? (Fr) Quii est-ce que sai m6re aime ti? c. Quieni ama a sui madre? (Sp) A quieni ama sui madre ti? d. Chii ama suai madre? (It) *Chii suai madre ama ti? 8 Note that in Brazilian Portuguese (BP) sua is 3rd person only in written style. In spoken BP sua is understood as 2nd person. (Mendez, p.c.) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 49 In Italian, Portuguese, and English WCO effects can be observed; however, these effects are not present in French and Spanish. In Spanish and French because it is possible to get a pair list reading in response to the question, there are no WCO effects with possessives, bringing into question their status as pronouns. If, as claimed by Giorgi and Longobardi (1991:155), possessives are adjectives in Italian, then the WCO effects observed are difficult to explain. A possible explanation arises, not from their status as adjectives, against which we have argued, but from the availability of stylistic inversion for both Italian, as seen in (25a) and Portuguese, as seen in (25b) . 25. a. Chii amaj suai madre tj ti? b. Quemi amaj a suai roae tj ti? Taking into account stylistic inversion, Italian and Portuguese would pattern like French and Spanish with respect to WCO effects. 2.2.3 Possessives as Pronouns The pronominal status of English and Romance possessives is similar, as shown in the following chart. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 50 Pronouns Bound by Reflexive WCO effects c-commanding QP binding_____________with, inv Fr yes yes no Sp yes yes no It yes yes yes no Po yes yes (?) yes no En________ yes____________yes____________yes______ n/a Possessives in both English and Romance pattern the same in respect to reflexive binding and being bound by a c- commanding QP. However, these languages divide as to WCO effects. English shows WCO effects; whereas, French, Spanish, Italian and Portuguese do not. 2.2.4 Pronouns and Articles When considering the possessive forms traditionally classified as pronouns, it is of interest to note that these forms require the article, unlike other pronouns which reject it, as seen in (26) . 26. a. le mien versus je/me (Fr) (the) mine vs. I/me Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 51 b. el mio versus yo/me (Sp) (the) mine vs. I/me c. il mio versus io/mi (It) (the) mine vs. I/me The possessive element here appears to substitute for a noun not a noun phrase, thus requiring the article to give it argumental status by making it a DP, following Longobardi1s (1995) and Stowell's (1987) proposals that arguments are DPs. And, in fact, in Italian, it is exactly the nonargumental positions where the article is dropped, as will be discussed in more detail in chapter 4. Thus the so-called possessive pronouns are distinctly different from other pronouns. Another possibility to account for the presence of the article with the "so-called" possessive pronoun would be to treat these constructions as analagous to elliptical expressions in which there is an empty N, as in (27a) . 27. a. el ec bianco b. el ec mio Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 52 The article, here, required to bind the variable, following Higginbotham (1987), supporting the predicate analysis X propose. While the traditional treatment captured the adjectival nature and the pronominal nature of possessives, the marked division between adjective and pronoun obscures the interaction between these aspects of the possessive. It appears, then, that the distinction between possessives as adjectives or as pronouns cannot be maintained. Possessives as both adjectival and pronominal must be considered independently. 2.2.5 Possessives: Adjective or Pronoun? If we consider possessives as adjectives in the broadest sense, i.e. as modifiers of a noun, we are unable to fine tune the nature of that modification, since possessives do not conform to the categorical properties of the various subcategories of adjectives. 2.2.5.1 Modifiers In respect to their adjectival nature, possessives appear to function more as restricted modifiers. So that "my house" corresponds to "the house that belongs to me." Gross (1986) suggests that the possessive, in fact, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 53 corresponds to the form "the definite article . . . relative". In his analysis the possessive "determiner" has as its source a single phrase, as seen in (28). 28. a. La peur qu'il avait 6tait grande. The fear that he had wass great, b. Sa peur £tait grande His fear was great.9 This correspondence is reflected in dative possessive constructions, where the notion belongs to is expressed by a prepositional phrase, e.g. in French la maison i . moi (the house to me = my house) . The availability of expressing possession either with a possessive element or by a dative construction led Langacker (1968) to suggest that the possessive form was derived from Le + N + k + Pronoun, providing one source for both constructions. However, as discussed in Chapter 1, these two constructions (i.e. the prenominal possessive and the prepositional construction) are different. In deriving one from the other we are unable to explain these differences. 9 The French examples are from Gross (19??) . The glosses are my own. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 54 2.2.5.2 Coordination One further property common to both adjectives and pronouns that one would expect to find in possessives if they are, as argued, both adjectival and pronominal, is that of coordination. Both adjectives and pronouns can be coordinated, as in (29) , respectively. 29. a. un chien petit et marron Fr. a dog small and brown b. Tu y yo tendremos que juntamos la semana que entra. Sp. You and I will have to meet the week that follows. You and I will have to meet next week. However, these possessive forms can not be coordinated, as shown in (30) for both French and Spanish, respectively.10 10 Note that they can be coordinated in the fully inflected form, as in (i) , leading to the two traditional classes of "weak" and "strong" (lausberg, 1965) . i. Perdimos el mio y el suyo. We lost mine and his. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 55 30. a. *mon et ton livre my and your book b. *mi y tu libro Coordination constructions with possessives, in Spanish, depend on the number of individuals involved; the second occurrence of the possessive can be deleted if the nouns refer to the same individual or aspects of the same thing, as shown in (31) . 31. a. mi madre y mi padre my mother and my father (different people) b. my amigo y colega my friend and colleague (same person) c. su paciencia y valor his patience and courage (aspects of same virtue) (Butt and Benjamin, 1988) In (31b) , there is one referent and thus one possessive. This correlates with the use of the determiner, shown in (32) . Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 56 32. el misterio o enigma del origen the mystery or enigma of the origin11 Postnominal possessives behave differently than their prenominal counterparts in Spanish. The postnominal can be coordinated, as in (33) , with the presence/absence of the second article reflecting a difference in interpretation. 33. a. la casa tuya y mia the house your and my b. la casa tuya y la ec mia In (33a) there is only one house involved; whereas, in (33b) there are two houses. In respect to French, in coordinated structures the use of the possessive follows that of Spanish, where the possessive is repeated when reference is to different individuals or aspects, as shown in (34) . 11 When the nouns form a single complex idea or mean essentially the same thing, only the first article is necessary, unless the nouns vary with respect to gender, in which case both articles would be used, as in (i) (Butt and Benjamin, 1988). (i) las aulas y los equipos the classrooms (fm) and the equipment (ms) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 57 34. a. ma m&re et mon p6re 1st sg fm mother and 1st sg ms father my mother and my father b. *ma m£re et p4re c. *ma m&re et soeur my mother and sister d. mon ami et coll&gue my friend and colleague In (34d), one referent/individual is understood; while in (34a-c) there are two individuals involved. Likewise, in Italian coordinated structures the possessive must be used with each element of the coordination if those elements refer to different "individuals", as shown in (35). 35. a. il mio soprabito e il mio cappello 1st sg ms coat and 1st sg ms hat my coat and my hat b. *il mio soprabito e cappello If the coordinated structure refers to one "individual", then only the first instance of the possessive is Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 58 necessary, as in (36) where the item functions as both a coat and a blanket. 36. Questo 6 il mio soprabito e coperta. This is (the) my coat and blanket. It seems strange that a property held in common by both adjectives and pronouns should not be available to possessives, since they behave at the same time like adjectives and like pronouns. This can be explained if we consider their status as clitics, as well as, their relationship to the determiner. If, as will be argued in Chapter 5, possessives are clitic-like then they could not be coordinated, as clitics can not be coordinated. With respect to the possessive's relationship to the determiner, assuming Higginbotham (1985, 1987b), the determiner acts as a binder, theta-binding a variable. As long as the possessed element refers to an "individual," it functions as a variable; however, when two "individuals" are involved, there are two variables that must be theta-bound requiring the presence of two binders. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 59 2.3 Possessives as Predicates Setting aside for a moment the pronominal properties of possessives, let us reconsider their status as adjectives. In section 2.2.1, we found that the possessive forms were adjectival in only one respect: agreement. According to Napoli (1989), gender-number agreement represents evidence of a predication relationship. 2.3.1 Predication A number of proposals have been made in the literature with repect to the analysis of predicate noun and adjective constructions of the type shown in (37), where red is predicated of the barn.12 37. we painted the bam red. These constructions have been either analysed as complex predicates (Chomsky, 1955; Larson, 1988; among others) or as small clauses (Williams, 1975; Chomsky, 1981; Stowell, 1981; among others). 12 There is smother interpretation possible where red would be the msmner in which the barn was painted. Under this interpretation, red would act as an adverbial-like modifier of the verb paint. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 60 2.3.1.1 Williams 1980 Williams (1980) argues that "NPs do not have predicates in their complement structure." Observing that English possessive 's structures with gerunds are permissible; whereas, these constructions in determiners are not, as shown in (38) , he appeals to the distinction between obligatory control and nonobligatory control. 38. a. *John's arrival dead b. John's arriving dead (Williams 1980: 218) He claims that obligatory control is a special case of predication and therefore should not be permitted in NPs. And, in fact, the control properties exhibited in NPs are those of nonobligatory control: (a) the antecedent need not c-command; (b) an antecedent is not necessary; (c) the controller can be outside the NP. These are illustrated in Williams' examples (73-75), repeated in part here as (39). 39. a. the attempt by John to leave b. any attempt to leave c. John left orders not to be disturbed. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 61 Williams offers no explanation of why predicates are absent from the complement structure of NPs, concluding that it "remains a mystery." It seems strange, in light of recent attempts to demonstrate parallels between IP structure and DP/NP structure that such a "mystery" should arise. If we look more closely at DP structure in relation to the environments where predication is possible, perhaps we can come to, at least, a partial answer. Following Williams there are two environments for predication. Grammatically governed cases meet certain structural descriptions, shown in (40) (Williams' example 16) ; while, the thematically governed cases involve predicates within a VP, where "the predication is of the theme of the verb of the VP," shown in (41) (Williams' example 17) . By positing a PossP where the structural conditions for grammatically governed predication can be met, we are able to account for predication in NPs. That there is no predication in the complement structure of NPs is due to a lack of thematically governed cases in NPs. 40. a. NP VP John died b. NP VP X John left nude./John left singing. c. NP be X John is sick./John is near Larry. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 62 41. John gave Bill the dog dead. 2.3.1.2 Napoli 1989 Napoli (1989) follows Williams in assigning a separate predication structure. She distinguishes this structure from lexical structure, noting that in predicate structures the role players receive semantic roles; whereas, in lexical structure the arguments receive theta- roles. The distinction between semantic roles and theta- roles relates to the number of roles available. Theta- roles are "a few gross semantic features like agent, patient, theme ..." while semantic roles are "unlimited, detailed semantic properties" (Napoli, 1989:30). A further distinction necessary for Napoli and, in deed, for predication theory, is that between predication and modification, illustrated by (42) (Napoli's examples (1-29) and (1-30)). 42. a. We painted the bam red. Predicate b. We painted the red bam. Modifier In (42a) red is predicated of the bam; while, in (42b) red assigns a property to the bam, but says nothing about Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 63 the barn being painted red.13 As Napoli explains, red in (42b) is within the N* of the N that it is modifying and as such it is part of the reference of the entire NP.14 Emphasizing the semantic nature of predication, Napoli concludes that there is no "one-to-one correlation between semantic entities and syntactic entities" (Napoli, 1989: 78). 2.3.2 The Semantics of Predication While Napoli draws a sharp distinction betwen semantics and syntax to expain the nature of predication, others (Rothstein, 1983; Zubizaretta, 1985; Gunnarson, 1986; Culicover and Williams, 1986: Hornstein and Lightfoot, 1987) have proposed syntactic analyses. These 13 in (42a) there are actually two predicates affecting the -barn: the Verb, paint and the Adjective, red. For a full discussion of primary and secondary predication and the relationship between them see Rothstein (1983) and Napoli (1989) . 14 An alternate interpretation, suggested by Saltarelli (p.c.), would have red predicated of paint in (42a) . Employing a Larsonian shell, bam and red would be complements at different levels with red in the place of the direct object. This seems to be the case in Italian, where di rosso (of red) is like an indirect object, as in (i) . i. Ha dipinto di rosso la casa. (he) has painted of red the house He painted the house red. ii. Ha dipinto la casa rosa. a.He painted the house red. b.He painted the red house. Note that in (ii) the position of the adjective makes the sentence ambigous between the two readings given. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 64 range from Rothstein's distinction between primary and secondary predicates to the postulation of the governing of PRO by Hornstein and Lightfoot to the elimination of PRO by Culicover and Wilkins. All of these implying that the semantics can be, if only in part, captured in the syntax. Chierchia (1985) arguing from the tradition of logical semantics claims that "there must exist some systematic relation between syntax and semantics" (Chierchia, 1985: 417). Assuming a Fregean view of properties which holds that properties have two modes of being as predicates and as singular terms, he argues that these modes are represented syntaxically in the English nominal system by the contrast between NPs and CNPs (common noun phrases) . NP is associated with semantic type e (individual expressions) and are therefore nonfunctional; whereas, CNP is associated with semantic type <e,p> (where p represents well-formed formulas) and are functional. Within the verbal system this distinction in modes of being is reflected in the contrast between finite and nonfinite. I will assume Chierchia1 s contention that there does exist a syntactic instanciation of semantics within grammar. Further I will argue that by taking into account Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 65 the semantics of possession, the syntactic behavior of possessive constructions can be accounted for. 2.3.3 The Semantic Representation of Possessives One aspect of possessives that has proved problematic is that the possessive demonstrates agreement with the possessor with respect to person and agreement with the possessed with respect to number and possibly gender. If we analyze possessives as two-place predicates15 then both these relationships can be realized: the pronominal nature of the possessive from a variable in Spec position being coindexed for person with the possessive and the adjectival nature of the possessive from the complement variable being coindexed with an NP providing number and gender agreement. If we take possession as an event of belonging, then as an event it requires participants. Possession then assigns two, what Napoli (1989) refers to as semantic roles as opposed to theta-roles: possessor and 15 The possessive as a two place predicate paralleling vp structure may explain the placement of the English possessive marker 's. This marker could reflect the contraction of the verb have, which denotes possession. It would be contracted with the subject which can be phrasal. So that (i) would have been derived from (ii) . i. the Queen of England's hat ii. The Queen of England has a hat. I will leave this for future research. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 66 possessed, The possessive relationship would then he represented semantically as in (43b). 43. a. my book b. POSS (I, book)16 This semantic representation entails two semantic roles for the possessive: the possessor (I) and the possessed (book). In keeping with X'-theory, syntactically (in a first approximation) the possessive would be generated as a head with the possessor as a specifier (external argument - subject) and the possessed as a complement (internal argument) . In chapter 3, I will argue for such a representation. 16 This follows a suggestion by Montalbetti (p.c.). He, however, distinguishes between weak and strong possession, where weak possession is relativized to the object possessed. The non- relativized case he considers economic ownership. This distinct may also account for the variation wrt inalienable possession. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 67 Chapter 3 The Composition of DP Following Pollock's (1989) decomposition of INFL into AGR and Tense, attempts have been made to establish a parallel between NP and VP structures. Bernstein (1993) observes a parellelism between adjective placement in NPs and Belletti's (1990) treatment of adverbs in VP. Abney (1987) proposes that the head of an NP is D, reinterpreting NPs as DPs. Giorgi and Longobardi (1991) suggest a layered NP with N' identified as the domain of internal arguments, N" the domain of the external thematic argument and N'" the domain of the possessor. While considerable attention (references) has been given to the functional categories present in CP, less attention has been given to their NP counterparts. A number of functional categories have been proposed for the CP including Tense, AGR - both Subject and Object, as well as Indirect Object, Focus, Neg, etc. If the DP mirrors CP, then, can we find evidence of similar Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 68 functional projections in DP? And if so what are these categories? Following Chomsky's (1992) minimalist program, in which he proposes that morphology plays a crucial role in accounting for parametric differences, I suggest that a morhological decomposition of nominal elements will reveal the functional projections within the DP. A number of languages show overt agreement as to number and gender with the head noun, as in (1) for French. 1. les petits princes the+pl little+masc+pl prince masc+pl the little princes Consequently, both Number (Valois 1991) and Gender (Picallo to appear) projections have been proposed. Looking at the possessive with a noun phrase in Romance languages, agreement in number and gender are also evident in certain positions, as in (2) for Italian. 2. i miei cappotti the+masc+pl my+masc+pl coats+masc+pl (the) my coats Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Possessives, showing adjectival-like agreement with pronominal-like behavior as we argue in Chapter l, show more than number and gender agreement. The pronominal element agrees in person with its referent while the adjectival element agrees in number and gender with the item possessed. This suggests that a possessive must be checked for possibly three agreements.1 Possible agreements since complete agreement is not overtly present prenominally for all persons in French and Spanish. Expanding on the structure suggested by Picallo, shown in (3) , the possessive could be generated in the spec of NP moving then to GeP for checking gender and then to NuP to check number. Since D following Stowell (1987) marks referentiality, the reference (person) would be checked there. 1 The idea of a separate projection for person was first suggested to me by J.-R. Vergnaud. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 70 3. DP /\ D1 NUP I / \ D Nu' GeP I / \ Nu Ge' NP I / \ Ge Poss N' / X N Compl The possessive appears to collapse two-three functional categories. The number depending on whether AGR is further decomposed into Number and Gender. For French the prenominal possessive shows agreement in number and gender for 1st, 2nd and 3rd person singular, but only shows number agreement for 1st, 2nd and 3rd person plural. The Spanish prenominal possessive shows agreement in number and gender only for 1st and 2nd person plural; all other forms show agreement only in number.2 This appears to account for the French and Spanish prenominal forms. Note the similarity to the agreement pattern of the definite 2 See the charts given in Chapter 1 for the actual forms. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. article. The singular definite article in French shows overt gender agreement, but non-overt number agreement; whereas, the plural definite article shows overt number agreement, but non-overt gender agreement. However in Spanish the definite article shows overt gender agreement in both the singular and the plural, but only overt number agreement in the plural. However, this structure does not reflect the semantics of the prenominal possessive as a two-place predicate. In this chapter we will begin by examining recent proposals concerning the structure of DP/NP. Then we will propose an expanded structure to account for the complexity of the possessive forms. This expanded structure, while abandoning any approach relying on the expansion of agreement projections, captures both the adjectival-like agreement features and the pronominal features of the possessive by treating possession as a two-place predicate heading its own projection. 3.1 Previous Proposals for Possessive Structures 3.1.1 Earlier Treatments The question of where the possessive is in English has been widely discussed in the literature (Jackendoff 1977, Abney 1987, Chomsky 1988, Lyons 1986, among others). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 72 3.l.l.l Jackendoff (1977) Jackendoff proposes a triple bar system for N where the genitive occurs in complementary distribution with demonstratives. While observing three semantic roles for the specifier of NP: demonstrative, quantifer and numeral, Jackendoff notes that certain of these elements can co occur, as shown in (4) (Jackendoff1 s 5.3) a. Fred's some the each ★ those all dwarf(s) which no any b. Fred's many the few dwarfs those several which To account for these co-occurrences, Jackendoff proposes two specifier positions: one for the article and one for the quantifier. He reaches the generalization that genitive NPs "arise" only in the determiner position, as shown in (5). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. POSS/D N" QP N1 NP | N This suggests that a determiner could occur wherever a possessive occurs; but this is not always the case, as shown in (6) . 6. a. We watched his every move, b. *We watched the every move. And this is certainly not the case if we try to extend Jackendoff's proposals to Italian. In Italian not only does the possessive occur with the article but with the demonstrative as well, as shown in (7). 7. a. la mia casa (the) my house b. questa mia bella casa ♦this my beautiful house Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 74 Furthermore, in Italian there appears to be some variation in the placement of the possessive with respect to other adjectives and numeral quantifiers3, as seen in (8) , as well as, with demonstratives, as in (9) .4 8. a. la mia bella casa (the) my beautiful house b. ??la bella mia casa c. le mie due belle case (the-pl) my-pl two beautiful houses my two beautiful houses d. le due mie belle case 9. a. ?questa bella mia casa this beautiful my house b. quelle due belle mie case these two beautiful my houses c. quelle mie due belle case d. quelle due mie belle case 3 cf. Comilescu (1982) and Dumitrescu and Saltarelli (1996) for the order between cardinals, adjectives and possessives in Romanian. 4 This variation in word order willl be addressed more fully in Chapter 5. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 75 Clearly the possessive in (8b and d) and in (9a, b and d) are not in determiner position. Nor is the postnominal possessive in Spanish in a determiner position, as seen in (10), since the determiner position is already filled in (10a); however the postnominal possessive can occur without an article preceding the noun as in (10b) . 10. a. la casa mia the house my my house b. Hay hermanas mias en el jardin. there are sisters my-pi in the garden there are some of my sisters in the garden This calls into question whether the possessive and the article occupy the same position in English; although, it appears that in English, Spanish and French the possessive has conflated with the definiteness within the determiner position, thus supporting Jackendoff's Poss/D. However, that question is further raised again when we consider all. English appears to pattern like Spanish and French in respect to the prenominal possessive in that these possessives cannot co-occur with the determiners that Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 76 Jackendoff (1977) classifies as demonstratives, i.e. definite article, interrogatives and possibly indefinite a and singular some. These possessives, however, can co occur with quantifiers, as in (11) . 11. a. all my books b. todos mis libros Sp c. tous mes livres Fr Note that the possessives in French, Spanish and English can co-occur with the quantifier all, but their occurrence with other quantifiers varies in acceptability, as shown in (12) . 12. a. iry many books b. ?mis muchos libros Sp c. *mes plusieurs livres Fr This does not parallel the occurrence of all with pronouns in English, as shown in (13) . 13. a. all of these people/all of them b. all these people/*all they came Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 77 c. I saw them all/*these people all d. I spoke to them all/*to these people all To account for the co-occurrence of a pronoun with all Maling (1976) proposes Quantifier Pronoun flip (Q-flip). However, in the case of a possessive there is no Q-flip, as shown in (14) . 14. a. all my books b. *my all books This could be attributed to the partially adjectival status of the possessives which would be reflected in their position relative to the quantifier. Jackendoff in accounting for the complementary distribution within English of the possessive and determiner-like elements captures a semantic connection between the possessive and the definiteness feature of the determiner position; however, his proposal cannot be extended to the Romance languages nor does it expain the interaction between the possessive and the quantifier all. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 78 3.1.1.2 Abney (1987) Abney (1987) suggests that the possessor role is assigned by D. He assumes that the NP provides a predicate over individuals and that the determiner is a functor which relates that predicate to the rest of the sentence. The possessive would then be generated in the Spec of dp, since lexical determiners appear to be in contrastive distribution with possessors in English. He gives the following tree for the structure of DP in English, assuming that the determiner is the site of person, number and gender features. 15. rnp DegP qp N However, by allowing two different types of elements (genitive case marker ’s and determiners) to appear as heads of a DP, X1-theory must be weakened. If we are to be Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 79 consist with X'-theory, then we would have to posit both a Determiner Phrase and a Genitive Phrase, a differenciation that Martin (1995) makes. Hudson (1989) noting differences between determiners argues for strong Determiners appearing in Spec of DP, while weak Determiners appear in Spec of NP. This distinction between strong and weak determiners will be discussed further in reference to the possessive in Chapter 4. Extending Abney's proposal to Romance encounters the same problems that extending Jackendoff1 s does with the added problem of accounting for all agreement features under one head, that being D. There is evidence from the analysis of Romance languages as well as other languges that projections for both number and gender are motivated. Ritter (to appear) looking at Hebrew and Valois (1991) working on French and English propose a separate projection for Number based on the distribution of adjectives. Carstens (1991) (as discussed in Valois 1991) proposes the necessity of a separate number phrase to account for the position of the numberal words in Yapese (an Austronesian language). Picallo (1991) argues for the separate projection of a Gender Phrase in Romance to Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 80 account for the gender agreement between nouns and determiners as well as adjectives. 3.1.1.3 Chomsky (1988) Chomsky (1988) suggests that for English genitive case is assigned in the context [NP _ X] , as in (16). 16. a. John's book b. his reading the book For Chomsky the realization of case occurs under government. In English the 's has been considered the realization of genitive case, but it is not assigned under government. Chomsky (1988) proposes a genitive rule that assigns case to the configuration. Assuming that only [-N] categories assign case then of-insertion is required for case in the N of NP constructions. If this were, in fact, genitive case marking then the English 's should occur for all genitive functions (see Chapter l section 1.1.1 for discussion). This is not the case. First, there are clear lexical restrictions on the use of the genitive ’s marking: the first noun should be a proper name, personal noun or a noun with personal Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 81 characteristics, but not an inanimate noun, as shown in (17) . 17. a. Ann's car b. *this house's roof If 's is a case of genitive case-marking, then, these constructions should be (17a) or not be (17b) amenable to of-insertion for genitive case marking in N of NP constructions; however, this does not prove to be the case. In fact, the reverse is true, as shown in (18) . 18. a. *the car of Ann b. the roof of this house These lexical restrictions appear to block some genitive functions. Secondly, the example in (17a) shows that so-called genitive case marking is unavailable for the genitive of material or composition. Nor can *s be used for the genitive of description, as shown in (19) . 19. *love's god the god of love Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 82 The apposition genitive and the partitive genitive are also unavailable for this so-called genitive case-marking, as shown in (20). 20. a. *intemperance's vice the vice of intemperance b. *song's verse the verse of a song Nor do subcategories of the partitive constructions, such as the partitive constructions of quantitative and qualitative have a genitive counterpart, as seen in (21) , respectitively. 21. a. a glass of water ~ *a water's glass b. this kind of research ~ *this research’s kind 3.1.1.4 Lyons (1986) Lyons (1986) also suggests that genitive case is assigned to the NP by possessive of as well as in the Spec position.5 He correlates the distribution of the two 5 Note the contrast between (ia) and (b) where (b) is possessive, but (a) is not and (ii) where both are possessive, i. a. a picture of John b. a picture of John's ii. a. a friend of Mhry (apposition genitive) (partitive genitive) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 83 constructions (i.e. NP's NP and N of NP) to these possibilities of genitive case assignment. The Spec positions that occur with pronouns which can bear case and the PP possessives occuring with full NPs which do not bear case. However the genitive 's morpheme can occur with full NPs. He concludes that nouns can assign inherent genitive case to both Spec and its complement. Then genitive case is realized as ’s prenominally and as of postnominally. The directionality of the case assignment would account for the variation in case realization: the noun directly assigns case to its specifier and indirectly via a preposition to its complement. But, then, how are we to account for the doubly marked possessives, as in (22)? 22. a book of his father's Following Lyons, the noun, book, would assign genitive case to its complement, father, via the preposition of; but, the complement is also genitively case-marked by the 's. This is counter the assumptions of standard case b. a friend of Mary's The post-genitive is restricted in its usage. The post modifier must be human and definite; whereas, the head noun must be indefinite. Given these restrictions, it appears that the postgenitive construction has a partitive component. The of in this case then may not be assigning genitive case. This possibility was not discussed by Lyons. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 84 theory, since the noun would be doubly marked for genitive case. A case assigner assigns a given case only once and each case recepient receives only one case. Given these assumptions, it becomes questionable as to whether we can say the noun, as a case assigner, assigns genitive case in two different ways by means of 's and the preposition of. 3.1.1.5 Olson (1989) Olson (1989) re-examines the status of the possessive in German with respect to Abney's DP hypothesis. She suggests that the possessive is a genitive marked personal pronoun. This case-marked pronoun, then, functions as a specifier of the DP. The structure for this is given in (23) . Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 85 23. DP / \ DP D' I / \ D' DET NP DET POSS N' 2ps 3ps N sg sg | gen dat 3ps fem sg dat fem I dein -er Katze your cat Comparing the pronominal possessive and the determiner, she suggests that their structure varies with respect to agreement relationships. The determiners show agreement for the features person, number, case, and gender with the noun; whereas, the possessives have inherent person and number features that do not necessarily agree with the noun, in addition to, the agreement features of the noun, as shown in (24). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 86 24. a. der Mist [3ps Sg Nom Masc] the dirt b. dein -er Katze 2ps sg GEN 3ps sg DAT Fem 3ps sg DAT FEM These agreement facts parallel those discussed for French and Spanish, where possessives reflect the person of the possessor and the agreement features of the noun possessed. To resolve this apparent clash in features, Olson proposes that the agreement features of the noun are copied onto DET establishing a Head-to-Head relationship with the noun. The agreement shown by the possessive, then, is a "reflex" not only of its structural position, but also of its morphological structure. She cites coordination facts as evidence for the possessive being a maximal projection functioning as a specifier. The possessive can coordinate with other possessive phrases, as in (25).6 6 The German examples are from Olson; however, the glosses are mine. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 87 25. a. Sie treffen sich in deiner und Karls Lieblingsnape. They meet themselves in your and Karl's favorite bar. They meet one another .... b. Ohne seines Vaters und meinen Rat sollte er nichts untemehnem. Without his father' s and my advice he should not untertake anything. However, the possessive cannot be coordinated with determiners, as shown in (26). 26. a. *lch wasche deinen und den Wagon. I wash your and the car. b. *Karl reparient weder das noch mein Telefon. Karl repaired neither the nor my telephone While her account attempts to reconcile the categorical duality of possessives, it cannot be extended to the Romance languages because of the degradation of agreement between the possessive and the noun, as in Spanish, and the occurence of the article with the possessive, as in Italian. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 88 According to Olson the agreement features of the noun are copied onto DET, if this were the case in Spanish then the possessive should reflect gender agreement with the noun which is not the case, as seen in (27a) . 27. a. *mia casa my-sg-fm house-sg-fm my house b. mi casa The correct form for the Spanish possessive in this example is given in (27b) , in which there is agreement as to number but not as to gender.7 Olson's account is unable to explain this reduction in agreement features. Nor will her account explain the presence of the determiner with the possessive in Italian, given that in her account both the specifier and the head positions in DP are both filled with the pronoun and the genitive marker, respectively. 7 I will argue in Chapter 5 that the prenominal possessive in Spanish has undergone a process of cliticization resulting in the loss of gender agreement. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 89 3.1.2 Giorgi and Longobardi (1991) Giorgi and Longobardi (1991) propose a possessive parameter to account for the distribution of the prenominal possessive. Treating prenaminal possesssives as genitive NPs, they argue that the possessive is realized as a determiner in English and French, while realized as an adjective in Italian. They note that prenominal possessives are in complementary distribution with elements traditionally treated as determiners, as shown in (28) for English and (29) for French. 28. a. *a my book b. *each my book c. *this/that my book d. *some my books e. *three my books f. *many my books 29. a. *le mon livre b. *un mon livre c. *chaque mon livre Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 90 d. *quelques mes livres e. *trois mes livres f. *plusieurs mes livres (Giorgi and Longobardi, 1991: 153-4) All of these constructions are possible in Italian, as illustrated in (30). 30. a. un mio libro b. ciascum mio libro c. questo/quel mio libro d. alcuni miei libri/qualche mio libro e. tre miei libri f. molti miei libri (Giorgi and Longobardi, 1991: 154) They relate this to three other properties that distinguish English, on the one hand, from Italian, on the other. First, when focused or used contrastively possessives occur postnominally in Italian, as shown in (31) . 31. a. El il mio cane. It's (the) my dog. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 91 b. E il cane mio. (it) is the dog Poss-lP It's ny dog! This cannot be taken as a matter of fact for Italian. Postnominal possessives can indeed be used with semantically contrastive readings, but prenominal possessives can also be used contrastively, as in (32) . (Saltarelli, p.c.) 32. Questa & la mia casa e quella & la tua. This is (the) my house and that is (the) Poss-2P This is my house and that is yours. Secondly, the possessive can occur as sentential predicates in Italian, as shown in (33) . 33. a. Questo denaro & mio. This money is mine. b. E tua questa borsa? Is this purse yours? c. Quei biglietti sono suoi. Those tickets are his/hers. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Finally, in Italian possessives may cooccur with a gapped head noun, as illustrated in (34) . 34. a. Metti i tuoi libri vicini ai miei. Put (the) your books next to+the my b. *Put your books next to my. (Giorgi and Longobardi, 1991: 155) The question here remains as to whether ai miei is used pronominally or as an elliptical adjective (i.e. ai miei pro) .8 To account for the distributional and property differrences between Italian-type languages and English- type languages, Giorgi and Longobardi propose the following parameter for UG in (35): 35. Possessive elements are syntactically specified to be realized on the surface as Adjectives or as Determiners. 8 If this is an elliptical construction, then the question remains as to whether the ellipsis form is pre- or postnominal: la [e] tua or la tua [e]. Saltarelli (p.c.) observes that with respect to ellipsis, Romance would appear to need a licenser (la) and an identifier (tua). The position of the empty category in these constructions will be discussed with reference to Spanish in Chapter 4 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 93 In Italian, the possessive would be realized as an adjective, while in English and French, the possessive would be realized as a determiner. Following from this hypothesis, they note that the possessive NPs without articles in English and French behave semantically like NPs containing a definite article in that they may occur in theta-maked positions as referential arguments. However, Italian possessives without the article are indeterminate and when singular occur in non-thematic positions such as in vocatives, predicative and exclamatory, as shown in (36) .9 36. a. Amico mio, vieni qui! My friend, come here! b. Gianni 4 mio amico. Gianni is my friend. c. Dio mio! My God! (Giorgi and Longobardi, 1991: 156) 9 It is the absence of the article that is important not the position of the possessive as the possessive can occur both prenominally and postnominally in these contexts, as in (i) . i.a. Gianni 6 amico mio. b. Mio Dio! (Saltarelli, p.c.) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 94 They conclude that French and English possessives are able to saturate an NP in the same way that determiners do, while Italian possessives are not able to just as adjectives are not able to.10 Their Possessive Parameter is, as they state, "orthogonal to the one we have proposed" (Giorgi and Longobardi, 1991: 153), that being the Directionality parameter which relates to the direction in which a language assigns case and theta-roles. This parameter accounts for the occurrence of both prenominal and postnominal adjectives in Romance, but only prenominal ones in Germanic. Taking the possessive as an adjective in Italian, Giorgi and Longobardi propose that the possessive is generated postnominally and then moved and adjoined to NP prenominally, as illustrated in (37). 10 For another view see Higginbotham (1989) and Ritter (to appear) . They consider adjectives as predicates and therefore they cannot appear as arguments as nouns can. under their view only predicates need to be saturated contrary to Giorgi and Longobardi1s view. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 95 37 . Nmax AP N' NP poss N e (Giorgi and Longobardi 1991:162) However, as Giorgi and Longobardi, themselves, note this categorical distinction can not be maintained consistently. There are examples that would be classified as determiners but occur in positions available to mio and not to my or mon, as shown in (38) - (41) . 38. a. Questo libro k mio. (It) b. *This book is my c. *Ce livre est mon (Fr) d. *Este libro es mi. (Sp) 39. a. Questo libro sembra mio. b. *This book seems mine c. *Ce livre semble mon d. *Este libro parece mi. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 96 40. a. Lo considero mio b. *1 consider it my c. *Je le consid&re mon d. *Lo considero mi. 41. a. Metti i tuoi libri vicino ai miei b. *Put your books next to my c. *Mets tes livres pr&s de mes d. *Ponga sus libros junto a mi.11 To account for this, they propose to treat these elements as having "weak" or "strong" distribution. A weak distribution is "defined as a prenominal position in an NP with a lexically realized head." (158) A strong distribution would apply to all the other positions available to determiners and adjectives, such as postnominal positions, predicative positions and NP positions with an empty head. This distinction proves problematic as it has to be modified language internally as well as cross-linquistically with items being marked +/- strong, +/- weak + weak/strong, as well as neutralizations of this distinction. This limits the 11 The Italian, French and English examples are from Giorgi and Longobardi (1991:154-55). The Spanish examples are mine. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 97 explanatory force of the generalization Giorgi and Longobardi propose in their Possessive Parameter. They classify English mine as a +strong determiner meaning that it can only occur in predicative positions or in empty headed NPs, but why would a determiner be so restricted? Their + categorical distinction obscures the relationship between the pronominal feature of the possessive and definiteness of the determiner position. Furthermore, it blurs the distinction between determiner and adjective, if strong elements can occur in both. I will argue in Chapter 5 that the distribution can be better accounted for when the status of prenominal possessives as clitics is taken into account. Their analysis proves problematic in its extension to other Romance languages. Within their framework, the manifestation of possessive elements as either determiner or adjective cannot be mantained with respect to Spanish. In Spanish they would have to say that the possessive is manifested as both determiner and adjective: as determiner when prenominal; as adjective when postnominal. A further question to consider is the place of the Romanian enclitic possessive within their determiner or adjective classification.12 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 98 3.1.3 Zubizarreta (19 87) Responding to Rappaport's (1983) claim that the status of POSS is a semantically restricted grammatical function, Zubizarreta (1987) observes that the semantic arguments that POSS is resticted to do not form a natural class, since it includes agents, "affected objects" and time adverbials, as shown in (42) respectively. 42. a. the Roman's destruction of the city (agent) b. the city's destruction by the Romans (affected object) c. yesterday's destruction of the city (time adverbial) (Zubizarreta 1987: 46) However, Zubizarreta supports the inituition behind Rappaport's analysis with respect to the noun phrase. "The relation between predicate-argument structure and syntactic structure is systematically more transparent in the noun phrase then in the sentence." (Zubizarreta 1987: 46) . To this end she proposes that there are two levels of 12 For a thorough discussion of the Romanian enclitic possessive see Dumitrescu and Saltarelli 1996. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 99 lexical representation. There is a lexical-semantic level (S-R) which encodes the selectional properties of a lexical item. These are not unordered lists of lexical specifications; they are a set of structured predicate- argument relations. There is, also, the lexico-syntactic level (L-R) which is composed of "frames" that mediate the mapping between syntax and the arguments from S-R. This mapping is achieved by one of two distinct mechanisms; predication and projection. Predication assigns a value to a variable; whereas, projection relates an external variable to each level of syntax.13 She formulates a rule of predication as (43). 43. If an open predicate P is present at a syntactic level L, then P has as sister a category C with index i at level L and the variable borne by P is interpreted as having the value i at level L. (Zubizarreta 1987; 13) In order to interpret modifiers and distinguish them as predicates or arguments, she proposes the Rule of Modification, given in (44). 13 This is essentially the Projection Principle proposed by Chomsky 1981. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 100 44. A modifies B in the context [C . . .A . . .B . . .] iff C immediately dominates A and B, C is a projection of B, and B is not a head. If A is an adjunct predicate which contains a variable x, then B or the head of B contains an arg-variable with the index i and x is assigned the value of i. If A is an adjunct argument with index i, then B or the head of B contains a variable x and the value i is assigned to x. Observing that the genitive in an NP is optional, Zubizarreta (1987) suggests that the genitive is an adjunct in the Spec of the NP, which assigns a value to a argument variable in the lexico-semantic structure in the head of N1 by means of modification. She assumes that both a determiner and a genitive phrase are dominated by a Spec node at the same time, as illustrated in (45) . Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 101 45. NP l.j R Spec m« Det NP* S N NP j John discussion of the book j 3 S-R: (P Y) , X 1=R by predication j=x (external arg-variable in the S-R of discussion) by modification (Zubizarreta, 1987) In (45), John's modifies the N' discussion of the book and assigns a value to the external arg-variable in the S-R of discussion. (Zubizarreta, 1987:53) Via modification the external arg-variable becomes saturated. This saturated variable is no longer able to form an open predicate capable of being saturated via predication.14 However, the open variable is not available to saturation by both modification and predication as these are triggered by different structures. Zubizarreta captures the predicational aspect of the possessive and argues, in part, for the interdependency of 14 cf. Higginbotham 1983, 1985 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 102 semantics and syntax. But by assuming both a determiner and a genitive phrase dominated by a single complex Spec node, she is unable to explain the exclusion of an overt Det with English, Spanish and French prenominal possessives. She states that their co-occurrence is not possible in English, but offers no explanation as to why, In Chapter 4 I will argue that this can be explained as the conflation of determiner and possessive. This structure does allow for an account of the Italian and Portuguese possessives with the article. 3.1.4 Tremblay (1991) Tremblay (1991) combines Giorgi and Longobardi's adjective interpretation with Zubizarreta's focus on a predication type relation. Claiming that the possessive pronouns function as adjectives occurring both pre- and postnominally, she proposes that the distinction is between attributive and predicative. Attributive adjectives, as shown in (46a), specify the nominal. 46. a. Un livre int^ressant a book interesting b. Ce livre est interessant. This book is interesting. (Tremblay, 1991) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 103 Predicative adjectives, as in (46b) assign a property to a specific nominal. In respect to the possessive, attributive possessives would be specifiers of a nominal, while predicative possessives would assign a property to a nominal.15 Following Higgins (1973) only the subject NP is referential in predicative sentences; whereas, in equative sentences the subject and the predicative are referential.(Tremblay, 1991) She argues that the possessive pronominal with the copular is equative, involving two referential NPs, as shown in (47a) for French. 47. a. [Cette maison]± est [la sienne]i this house is (the) yours b. [e] est [SC NPi NPj.] (Tremblay, 1991) The two NPs are base-generated inside a small clause, as illustrated in (47b). She further proposes that pro is present with possessive pronominal constructions and that this empty head is licensed by the presence of a 15 A similar claim is made by Dumitrescu and Saltarelli (1996); however, they derive this distinction from two different modes of saturation. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 104 possessive pronominal. Reference, then, comes from the presence of the empty head. This structure is illustrated in (48). (The index comes from the nominal head.) 48. a. Cette maison est la mienne. This house is DEF mine b. [Cette maisonili est [la mienne proj.] i (Tremblay, 1991) This possessive pronominal is generated in prenominal position, as illustrated in (49b).16 16 She argues for a prenominal position for the possessive pronominal based on historical evidence. She cites Bruneau and Brunot (1969) who state that in Old French the possessive form corresponging to mien was used with an overt head and when it was so used it preceded the head, as in (i) . i. Sire, por coi m'avez traie? Ge non ai, voir la moie amie. Gr6visse (1980) also recounts the use of the pronominal possessive before the noun, as in (ii) . ii. J'ai retrouv4 1'autre jour un mien article. (Montherlant, Solstice de juin, p. 222) Although Gr6visse also records the use of this form postnominally, as in (iii). iii. Cette oeuvre mienne (Gide, Journal, 5 mai 1942) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 105 49. a. le mien b. Spec /\ A D NP o l NP N' I I mien N I pro (Tremblay, 1991) While possesssive pronouns in Tremblay's terminology, such as le mien, are not clitics, possessive adjectives (again in her terminology) , such as mon, are clitics and are generated in the Spec of DP. From this position, the possessive adjective can identify an empty category in postnominal position, unlike the possessive pronoun which licenses the empty head, but does not identify it. The structure for possessive adjectives proposed by Tremblay (1991) is illustrated in (50b) . Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. o 106 50. a. Ton livre b. Spec I Ton 1 / \ D NP [+def^^ N* NP I N pro [+dat] I livre Recall that Tremblay supports the status of the possessive adjective as a clitic by their cooccurence with adjectives, but that at first blush a similar treatment for Spanish was not possible. Nor is this analysis available for Italian and Portuguese, since in both these languages the determiner position is lexically filled. Furthermore, we have seen in Chapter 2, sections 2.1.1 and 2.1.2 that classifying possessives as predicative or attributive is problematic. 3.1.5 Valois (1991) Valois (1991) also pursues a clitic approach to the possessive. While the tests he employed would seem to rule out possessives as clitics, we will see in Chapter 5, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 107 section 5.2 that the possessive is clitic-like in its characteristics. He further observes that the possessive is marked for genitive case, which is not the case with adjectives; thereby discounting the possessive as an adjective. He argues that possessives are both XPs and determiners and as such they go through the Spec of NumP and then cliticize onto D, as illustrated in (51). 51. DP / \ D NumP / \ Spec Num' DetP He further proposes that the possessive receives case in SpecNumP. SpecNumP as a case position, then, accounts, according to Valois, for the co-occurrence of the possessive pronoun with a determiner in Italian, but not in English and French. As Valois admits this depends crucially on the SpecNumP not being available for full DPs. This seems to be an ad hoc stipulation. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 108 While we could say that in English, French and Spanish the possessive is realized as an NP; whereas, in Italian and Portuguese it is realized as a DP, this fails to account for the absence of the determiner in Italian and Portuguese in certain environments; environments which would preclude a derterminer position as they are non argument positions. These environments include, as noted by Giorgi and Longobardi (1991) , vocative, predicative and exclamatory contexts, as given in (52). 52. a. Amico mio, vieni qui! my friend, come here b. Gianni & mio amico Gianni is my friend c. Dio mio! My God! These various previous analyses have treated the possessive as a determiner (Abney 1987), as a genitive case-marked pronoun (Olson 1989) , as a clitic (Valois, 1992; Tremblay, 1991), as an argument (Giorgi and Longobardi, 1991), or as an adjunct (Zubizarreta, 1987). While showing insight into the nature of possessives, none of these analyses has been able to fully account for the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. distribution of the possessive or the variation within Romance. 3.2.0 Decomposing DPs The nature of the noun phrase has been the object of recent research (Abney 1987, Szabolcsi 1983, Giorgi and Longobardi 1991, Valois 1991, Picallo 1990 among others). Abney's (1987) DP hypothesis, discussed in section 3.1.1.2, proposes that NPs are headed by a functional category in much the same way as VPs are headed by INFL. He suggests this functional category is Determiner concluding that NPs are DPs. Szabolcsi (1983) also notes the parallel between CP and DP. She proposes a functional category between D and N following from the Hungarian possessor agreeing in number and person with the head noun. This nominal agreement parallels verbal agreement. 3.2.1 Projections in DP Both Number (Valois 1991) and Gender (Picallo 1994) projections have been proposed. 3.2.1.1 Valois Valois assuming the parallelism between CP and DP suggests that these structures differ in respect to the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 110 type of functional projections they contain. He proposes that the differences between languages is accounted for by the same mechanism accounting for differences between VPs, namely the presence or absence of head movement. He argues for a functional category between D and N suggesting that this category is Number Phrase which contains the number features of the DP. He proposes the following structure for French. 53. NP Head Movement in French DP / \ D NumP /\ Num NP* I / \ +/- Spec N*' pi ^ /\ /j^ agent^N* NP N / \ N NP I ^ theme —noun (Valois 1991:53) In briefly addressing the position of the prenominal possessive in English and French, he assumes that the possessive pronoun is projected in SpecNP* and Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Ill obligatorily moves to Spec of NumP, which he proposes is the case position for the possessive pronoun. He concludes that the difference between French and English is that "a pre-nominal possessive pronoun in French is case-marked once (by the noun) , while it is case-marked twice in English (by both the noun and SpecNumP." (valois 1991: 103) Italian possessives present a problem with respect to case assignment since the Det, following Valois, occupies the SpecNumP position. To account for the agreement between the possessive and the noun, Valois proposes that a mechanism of feature transmission operates between the determiner and the possessive.17 The determiner, then, transmits its case to the possessive. Valois' proposes three different types of case assignment for the possessive in order to account for the agreement18: in SpecNumP, by the noun, and by feature transmission. In doing this he posits that in English the possessive is case-marked twice which is contra standard case theory. Further either his proposal for case transmission in Italian would have to be extended to 17 This mechanism is similar to that which operates in Italian impersonal passives allowing for the expletive pronoun to transmit its features including case to the post-verbal subject. 18 The central point of the framework he adopts equates case positions with agreement positions. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 112 account for the agreement between the noun and the postnominal possessive or a separate mechanism would be needed. Finally his account is unable to explain case assignment/agreement in the cases where the determiner is absent. 3.2.1.2 Picallo Picallo (1994) proposes that the possessive in Catalan is generated in the Spec of NP, as shown in (3) repeated here as (54). 54. DP /\ D* NuP I / \ D Nu* GeP I / \ Nu Ge' NP I / \ Ge Poss N’ / \ N Compl Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 113 Picallo assumes both a Nu(mber)P and a Ge(nder)P for Catalan with the noun moving up through GeP and then NuP. If the noun moves and the possessive stays in situ, the result will be a postnominal possessive. When the possessive moves up first, it will be in a position to merge with the determiner as I will propose it did in French and Spanish. Modifying the structure suggested by Picallo, shown in (54) , to account for person, then the possessive is generated in the Spec of NP moving then to the Spec of GP for checking gender and then to the Spec of NuP to check number. Finally moving to the Spec of DP, where, since D following Stowell (1987) marks referentiality, the reference (person) would be checked. Evidence for a separate projection for person comes from the possessive constructions in Isthmus Zapotec, a language spoken in Mexico. Here the possessive and the person morphemes are separate, as shown in (55) . 55. a. s-palu-be poss. stick 3rd person sg. his stick Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 114 b. s-palu-lu poss. stick 2nd person pi. your stick If we project the pronominal reference in D, the problem remains as to how to account for the presence of the article with the possessive in Italian and Portuguese, as well as, the presence of the possessive in the absence of the article in non-argumental positions. Picallo maintains essentially an adjectival analysis for possessives in Catalan. However, it is this adjectival nature of the possessives that is being questioned. 3.2.2 Adjectival Status and Structure If we consider the partly adjectival status of possessives, then we need to account for their structural position in relation to the structural position of adjectives. Bernstein (1993) accounts for the different classes of adjectives by proposing multi-adjunction sites for adjectives. She proposes three classes of adjectives that are distinguished by their adjunction sites. Adjectives are adjoined to NumP, as in (56a) , to NP, as in (56b), or to XP, as in (56c). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 56.a. une large valle& DP D NumP / \ AP NumP I / \ large Num NP une vailed N i t i b. une vallee large DP / \ D NumP I / \ xine Num NP I / \ valle6 AP NP i | | large N t i Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 116 c. un mero accidente DP / \ D AP I / \ un A NumP I / \ mero Num NP I I accident N 1 I t i (Bernstein 1993) The structures proposed in (56a and b) account for the difference in distribution of pre- and postnominal adjectives. While these structures allow for an account of the differences in readings between the pre- and postnominal adjectives by providing distinct syntactic positions for the three classes, they do not provide an account for the derivation of the semantic interpretation. Bernstein is unable to explain why adjunction to NumP is associated with a non-restrictive reading and adjunction to NP is associated with a restrictive reading. She suggests that adjunction to NumP "permits" a nonrestrictive reading for the adjective; while, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. adjunction to NP "gives" a restrictive reading. Bernstein notes that this parallels Belletti's (1990) adverb adjunction to either AgrP or VP. Belletti proposes that sentential adverbs are adjoined to AgrP, while VP adverbs are adjoined to VP. Since possessives do not yield restrictive/nonrestrictive readings based on position, neither of these sites would seem probable. The structure proposed in (56c), also, proves problematic. This structure does not permit a postnominal position as the trace of the moved noun would not be governed by its antecedent since the A0 would intervene leading to a violation of the HMC (Head Movement Constraint) . These possible adjunction sites for adjectives do not appear to be possible sites for the possessive. 3.2.3 Pronominal Status and Structure It would appear that if an adjectival position is not available then the possessive might be generated in a pronominal position. 3.2.3.1 Koopman Koopman (1993) suggests that pronouns always occur in either Spec or Agr positions. She proposes both an AgrP and a NumP as functional projections of DP. Pronouns take Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 118 the place of a noun. They are specific and they have both number and person features. Having these features, they must be checked. They can be checked either by head movement to a functional category or by movement of some projection containing the pronoun to a Spec position where these features would be checked under spec-head agreement. She claims that the basic structure is the same crosslinguistically with the surface variations in order accounted for under movement. The internal structure she proposes for DP is given in (57) . 57. DP /\ Spec / \ D NumP SpecT^X^ Num Spec N' N I pronoun In her discussion of pronouns she omits person features, but suggests that these should project in PerP between D and Num. In addition to the variation from movement, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 119 languages would also vary as to which functional projections are overt and which are what Koopman calls silent (covert) . She suggests that in English NumP is always overt and D and Agr are silent. Since possessive prenominals do not take the place of a noun, they are not generated in N; rather, they are generated in Spec of NP as the external argument of a noun. They then move via spec-to-spec movement to check their features. 3.2.3.2 Cardinaletti Cardinaletti (1993) looking at Romance, as well as, German pronominals proposes a variation in internal structure depending on whether the pronominals are strong pronouns or clitic pronouns. According to her analysis strong pronouns are projected as full DPs and have the lexical category NP embedded under DP; whereas, clitic pronouns are projected as functional projections. She further distinguishes between clitics and weak pronouns, which contain a further functional projection. These are shown in (58) , her (54) ,19 19 She proposes a support morpheme for weak pronouns observing the more morphologically complex nature of weak pronouns as opposed to clitics which consist of either a case morpheme or an agrement morpheme marking phi-features, as she shows for Italian in (i). i. WEAK CLITIC Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 120 58. a. strong DP I / \ D° ^SuppP Supp° NP LA N° b. weak DP / \ D° SuppP^ Supp® c. clitic DP I D® While her analysis is concerned with object pronouns, 3rd- person object pronouns in particular, she suggests that the analysis could be extended to subject pronouns. She does not consider the pronominal status of possessives. The prenominal possesives under consideration will be argued to be clitic-like in chapter 5, section 5.2; however, they show more than case or agreement features, they also reflect reference. While this would appear to be captured by projecting the possessive clitic under D®, it fails to adequately capture the semantics involved. 3.3.0 Arguments and Predicates If, as it appears, we are unable to capture the structure of prenominal possessives by treating them as lo-ro gli/le,* lo/la/etc. (Cardinaletti, 1993:? Loro contains the support morpheme r which is not found on clitic pronouns. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 121 adjectives or pronouns; then, we must consider what alternative(s) exist. 3.3.1 Possessives and Argument Structure Moving away from the categorical distinction between pronoun and adjective, the only possible alternative would be the classifying these possessives as either arguments or predicates. 3.3.1.1 Giorgi and Longobardi Considering the possessive as an argument of the noun, Giorgi and Longobardi (1991) propose a hierarchical organization of DP-internal arguments based on extraction and binding facts. This hierarchy is shown in (59) . 59. Giorgi and Longobardi's Thematic Hierarchy N"' N" possessor N' external argument N internal argument Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 122 This hierarchy led them to propose the Possessivization Principle given in (60). 60. Possessivization Principle: The unique phrase allowed to appear as a possessive is the hierarchically highest genitive argument of an NP. (Giorgi and Longobardi 1991:68) While this hierarchy accounts for the relation between arguments, it does not account for the variation between Romance and Germanic languages.20 3.3.1.2 Grimshaw This assumes, in part, Grimshaw's (1990) proposal that possessives are a(rgument)-adjuncts. Distinguishing between arguments, adjuncts and a-adjuncts, she claims that possessives resemble arguments in respect to licensing, but are not theta-marked. If possessives are arguments, then, under the Theta-Criterion they would be obligatory as part of the argument structure; however, they are optional. If adjuncts, possessives could occur in 20 To account for the variation they propose the Head-Subject Parameter. This parameter explains variations in Theta and case assignment between Romance and Germanic languages. It does not apply to the variation within Romance wrt possessives. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 123 adjunct positions; however, their behavior is not consistent with the behavior of adjuncts. Expletives are unable to occur in possessives, as seen in (61). 61. a. *It's claim that John left, b. It is claimed that John left. (Grimshaw, 1990) Finally, possessives can not raise into non-argument positions, as in (62). 62. *John's appearance to have left. (Grimshaw, 1990) Further, as Stowell (1987) observes, if a possessor theta-role can be assigned then all nouns would have a possessor theta-role as part of their lexical entry. This can be avoided by distinguishing between theta-roles and semantic roles, a distinction Napoli (1991) makes with respect to predication relationships. 3.3.2 Possessives as Predicates Given that the structure suggested by Picallo, shown in (54) , captures the external argument-like status of the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 124 possessive, it still maintains an adjectival analysis of the possessive and as such must be abandoned. If, as I argue in chapter 2, the possessive is a predicate, then, its syntactic structure needs to reflect this relationship. I propose that the possessive is generated in its own projection with the possessor in spec of PossP and the possessed as a complement reflecting the external and internal arguments of a two-place predicate, as shown in (63) . 63. / \ D POSSP = XP / \ P Poss' Poss NP The possessive behaves differently and is generated in an XP because of the number of its semantic roles, via Spec- Head agreement, the possessive agrees in person with the possessor. The possessive in French, Spanish and Italian raises via head-to-head movement to D to check Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 125 referenciality, following Chomsky (1992), where in French and Spanish the possessive and the definite article merge prior to Spell-out forming a clitic-like element with reduced agreement. In Italian, the possessive also raises to D where it incorporates with D, but does not merge retaining the form article + possessive. This structure is able to capture the predicative relationship expressed by the possessive, while allowing for both the pre- and postnominal structures and their subtle variation in interpretation.21 3.3.2.l Predication Structure The structure I am proposing for the possessive follows Bowers (1993).22 He proposes a Predication Phrase having the structure shown in (64), his (8). 64. PrP / \ (siobj ect) NP Pr ’ / \ Pr XP (predicate) X = {V, A, N, p} 21 For the purposes of this analysis, this structure is being proposed for Romance languages. It remains for future research whether this can be extended to other languages, without argument, I am assuming at this point that it can be. 22 Sdnchez (1995) also follows Bowers in proposing a predicate phrase within NP. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 126 Applying Napoli’s semantic roles to predication structure what would be the external argument is in the case of possessives the semantic role of possessor which is projected in the Spec position. The predication relation then holds between the semantic role in Spec and the complement of Pr which would be the possessed. Projecting a predicate phrase within DP has been proposed (Dumitrescu and Saltarelli, 1995) to account for the semantic distinction between the adjectives that can occur in both pre- and postnominal positions. Noting that the postnominal position of these adjectives gives a restricted, extensional reading and allows for modification, they propose that these adjectives are projected as predicates with a referencial empty category that is saturated via theta-identification (a la Higginbotham (1987)), as shown in (65)23 23 The subscripts 1 and 2 do not play a formal role. They are used only to distinguish between the two adjectival phrases. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. NP /\ DP / \ D AP e A. N1 / \ N AP i 2 / \ e A* i I A la hennosa casa blanca the beautiful house white The prenominal counterparts of these adjectives are projected as predicates that are saturated by the determiner D via theta-binding, allowing them to be open to metaphorical interpretation, as illustrated in (66) . 66. a. el pobre hombre the poor=wretched man b. el hombre pobre the man poor=economically improvished Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 128 The adjectival-like agreement of the possessive would reflect the predication relation as claimed by Napoli (1991) . The possessive would then be checked for referentiality in D. If we check the pronominal reference in D, the problem remains as to how to account for the presence of the article with the possessive in Italian and Portuguese. I will propose that in Italian and Portuguese the pronominal reference is moved to head position with the determiner in Spec of DP. Then the article and possessive would agree under Spec-head. I suggest that this is also the case for French and Spanish, where there is no overt article. If, as Kayne (1993) suggests, the specifier is an adjoined position, there I propose that the article and the possessive are conflated or merged in French and Spanish, as suggested by Vergnaud and Zubizarreta (1992) for French as shown in (67); with the merged form retaining the features both have in common.24 24 In a talk at UC Irvine (Jan. 29, 1994) , Tsai also suggests a merger of a reduced form of the English definite article with an indefinite morpheme to account for the formation of some English pronominals, as illustrated in (i). i. pronominals th -ey th-en th-em th-ere th-at Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 129 67. a. [DP pronoun[D' determiner]] - * ■ suppleted form b. eux + les - * • leurs c. eux + le -+ leur Proposing a merger of the definite article with the possessive in French and Spanish will account for the surface variation in possessives in Romance languages. 3.4 DP Structure Just as morphology has motivated the decomposition of IP into its components (i.e. Tense, AgrO, AgrS) so can the morphology of the DP. Looking from the perspective of the possessive, we have seen analyses for decomposing DP agreement into number and gender; and, now, we can extend that decomposition to person. For now, we will project person under D. As a functional element person is a closed lexical class and, as we have seen from the discussion of clitics, it is often phonologically weak. Furthermore it is inseparable from its complement. The variation found in languages will not be a result of different structures, but rather the extent to which the basic structure must be expanded to reflect the morphology of a given language, as well as, capture the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 130 semantics involved. To this end we are proposing a PossP reflecting the POSS as a two-place predicate. In Spanish and French, I will argue, the determiner and the possessive, through a process of cliticization and morphological merger are reduced to the present day prenominal possessive. In the case of the postnominal possessive in Spanish, the noun has moved as Cinque (1993) proposes, raising to D, following Longobardi (1994), and the possessive remains in situ. Since the possessive would not be adjoined to D, it would not acquire a definiteness feature thus permitting a partitive-like interpretation postnominally; although, it would still be within the feature checking domain for referentiality. In the next chapter, I will discuss this merger in respect to a unified underlying structure for Romance possessive constructions. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 131 Chapter 4 Accounting for Variation Prenominal possessive constructions in Romance languages surface in one of two distinct patterns: (a) article + possessive + noun as in Italian (la) and Portuguese (lb), (b) possessive + noun as in French (2) and Spanish (3).l»2 1 There is a third pattern that can be found in Romanian and Marsian (an Italian dialect). This pattern involves a clitic: N-clitic Poss, as in (i) for Romanian. i. soru-sa sister-his his sister This is not a general pattern as it is restricted in use, notably to kinship terms. The Marsian form is even more restricted. For a discussion of the Romanian data see Comilescu (1993, 1995) and Grosu (1988) . The information on Marsian is from Mario Saltarelli (p.c.) 2 Mario Saltarelli (p.c.) informs me that in the Venetian and Toscan dialects (i) is acceptable, (c.f. TekavCfi (197 2) who further cites Rohlfs (1968, 1969) for the Toscan data). i. la su casa There appears to be a process of historical change involved here. The change in the possessive construction appears to progress in the following way: a. Art + Poss (number & gender) + noun b. Art + Poss (number only) + noun c. Poss (number only) + noun and if we add the Romanian, then d. N-cliticPoss I suggest that the change was motivated by changes in the features on either the possessive, the article or both as a result of a gradual cliticization of the possessive. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 132 1. a. La sua casa & in Italia. (the) his/her house is in Italy, b. o seu l&pis (the) his/her pencil (masc) 2. a. sa maison his/her house b. *la sa maison 3. a. su casa his/her/your house b. *la su casa (the) his/her/your house 4.1 Possessive structure Cinque's proposes (1993) that adjectives are base generated to the left of N, as in (4a) with the noun subsequently raising to a functional head between N and D, as shown in (4b). 4. a. [ dp .. ( np AP [ n 1 N conqpl.] ] ] b. L'invasione italiana t dell'Albania the invasion Italian of Albania the Italian invasion of Albania (Cinque 1993) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Although we have seen in Chapter 2 that possessives are only nominally adjectival in that they agree with their noun, they must occur in a configuration which allows for this agreement. To that end we will begin by following Cinque and propose that the Possessive Phrase is also generated to the left of N: the resulting variation between pre- and postnominal forms as evidenced in Spanish and Italian resulting from movement. I propose that the underlying possessive construction is Art + Poss + Noun in these Romance languages. 4.1.1 Old French and Old Spanish Historically Old Spanish and Old French are suggestive of this proposal. In both Old French and Old Spanish the prenominal possessive appears with the article resembling the system of Italian, as seen in (5) and (6) , respectively. 5. a. la soe amie the (fm. sg.)his (fern, sg.) sweetheart (fern, sg.) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 134 b. un mien enemi mortel a (masc. sg.) my (masc. sg.) enemy (masc. sg.) mortal3 c. li miens cuers the (masc. pi.) my (masc. pi.) heart (masc. pi.)4 6. a. las suas fijas (El Cid, In. 275) the (fm. pi.) his/her (fm. pi.) daughters (fm. pi.) b. los mios dias (El Cid, In. 220) the (masc. pi.) my (masc. pi.) days (masc. pi.)5 The Old Spanish use of the article with the possessive is not restricted to penisular Spanish, but can also be found in New World Spanish. The evidence here (7) suggests a transitional state between the Old Spanish in (6) where the article + possessive show both number and gender 3 The example in (5a) is from Chretien de Troyes, Erec et Enide, line 296 and (5b) is from Chrestien de Troyes, Yvain, line 4912 found in Einhom, E. 1974. old French: a concise handbook. 4 Example (5c) is from La Chasteleine de Vergi, verse 77 3. D. Artega (p.c.) 5 The Old Spanish examples from El Cid were pointed out to me by Claudia Parodi (p.c.). Penny (1991) notes the frequent use of the definite article with the prenominal possessive. This use became rare by the early sixteenth century. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 135 agreement and Modern Spanish in (3) , where the possessive shows only number agreement. 7. a. en dote y casamiento con la nuestra hija (Tucuman 1611) in dowry and marriage with the our+fm+sg daughter b. este mi distrito (Bs. As. 1738) this my+sg district c. este mi recurso (Bs. As. 1803) this my+sg appeal6 Additionally evidence from Andean Spanish shows the determiner co-occurring with the possessive as shown in (8)7 with the indefinite. 6 The examples are from Fontanella de Weinberg, B. 1992. El Espanol de America. Madrid: Mapfre. The glosses are my own. Fontanella de Weinberg reports: existen asimismo, construcciones en que la funcion de demostrativo esta desempenada por el participio dicho, que puede estar precedido o no de articulo: cuatro mil y setenta pesos de la dicha su dote (Tucuman 1607) de quien el dicho su padre era devoto (Pto Rico 1647) 7 Javier Galvan (p.c.) informs me that there are Mexican dialects which also allow for the article with the possessive, as in (i) . i. Recoge todos los sus juguetes de su hermano. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 136 8. uno mi hermana mayor8 one my sister older one of my older sisters (Paredes, p.c.) This use of the possessive with an indefinite article is also attested for Guatemalan Spanish, as shown in (9) . 9. a. Me voy a tamar una mi tacita de caf6 antes de salir. me (I) go to drink (a) Poss-lp-sg cup of coffee before to leave I'll just drink a little cup of coffee before I go b. Voy a comprar unas mis diez vacas Holstein para poner una mi lecheria. (I) go to buy some Poss-lp-pl ten cows Holstein in order to start a Poss-lp-sg dairy. I'm going to buy about ten Holstein cows for myself in order to start me a dairy. (Martin, 1985) 8 Liliana Sanchez (p.c.) suggests that agreement in second languages is unstable. In Standard Spanish she proposes incorporation of AgrO head-to-head. In bilingual Spanish she suggests two possibilites: AgrO movement doesn't take place or it does, but the morphemes are not being assigned their masculine or feminine features yet. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 137 The use of the indefinite article in (9a) suggests a partitive interpretation for the construction; however, this is clearly not the case in (9b) .9 4.1.2 Inalienable Possession Furthermore, inalienable possessive constructions are suggestive of the relationship between the definite article and the possessive in French and Spanish. In both French and Spanish, cases involving body parts and clothing where the ownership is clear may be expressed by the definite article, as seen in (10) for French and (11) for Spanish. 10. a. Pierrei a lesi mains sales. P. has the hands dirty b. Mariei a lesi cheveux roux. M. has the hair red c. Pierrei tient le livre dans lai main. P. is holding the book in the hand 9 A possible explanation for both the Andean and Guatemalan use of the indefinite with the possessive may be found in language contact. The form originally introduced by the Spanish was reinforced by similar forms found in the native languages resulting in the maintance of a converged form that did not continue in Modem Spanish. Martin (1985) proposes a convergence hypothesis to account for the Guatemalan data. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 138 11. a. Mariai le extendi6 lai mano a Pedro. M. extended the hand to Peter b. Arr6glatei eli pelo you tidy the hair In (10a) les mains sales are Pierre's hands, in (10b) les cheveux is Marie's red hair and in (10c) la main is Pierre's. In (11a) la mano is Maria's hand and in (lib) it is your hair that you are being asked to tidy. This possessive reading is not available with the indefinite article, as shown in (12) for French (Judge and Healey, 1985). 12. a. *Pierrei a unei main sale. P. has a hand dirty b. *Mariei a desi cheveux roux. M. has some red hair c. *Pierrei tient le livre dans unei main. P. is holding the book in a hand Kinship relationships are also considered inalienable and the definite article without the possessive occurs in these constructions, as shown in (13) for Italian. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 139 13. Giannii ha perso ili padre. Gianni has lost the father Gianni lost his father. In Italian the appearance of the determiner implies inalienable possession. In (13) the father can only be interpreted as Gianni's father.10 The merger, which I am proposing, of the definite article and the possessive would sanction the relationship and alternation between definiteness and possession. 4.2 Merger Given an historical common underlying structure, how then can we account for the present variation? I propose that in French and Spanish the definite article has merged with the possessive. This merger yields a possessive agreeing in number only in Spanish and French for some forms, as was seen in the charts in Chapter 1. Recall that in Spanish, there is only agreement in number for first and second person singular and for third persons in both singular and plural. While in French, there is agreement 10 It is interesting to note that in Totonac, a language spoken in Mexico, the possessive morpheme is prefixed to the possessed and functions as a determiner; although, the language does not have determiners. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 140 in number only for all plural first, second and third person. Full agreement, that is agreement in both number and gender, for all forms appears only when the possessive occurs post nominally, in Italian and Portuguese there is no merger resulting in the form Art + Poss + Noun. This is reflected in agreement in both number and gender in the declension of the possessive, as was seen in the charts in Chapter l.H The merger of the definite article with another element is not without precedence. In French, the interrogative adjective quel (what) is merged with the definite article to form an interrogative pronoun, lequel (which one). There is further evidence suggestive of a merger between the article and the possessive with a resulting reduced clitic-like form from the history of Spanish. Penny (1991) reports that the loss of tonicity and reduction in form was accompanied by the loss of the definite article resulting in the rarity of the article + possessive + noun form by the beginning of the sixteenth century. 11 Recall that loro (3p-pl) does not reflect agreement. Agreement is shown by the article and I have suggested that this is why loro always occurs with the article. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 141 4.2.1 Pronominal Forms If the prenominal possessive form is the result of a merger, then, should the noun move before the merger in French and Spanish, the form Art + Noun + Poss could surface. This is the case, as seen in (14a) for Spanish, although it is not the case for French, as in (14b) . 14. a. el amigo mio the friend my b. *le ami mien The pronominal or elliptical form of the possessive has full agreement and arises from the movement and subsequent deletion of the noun. The resulting trace blocks the merger of the article and the possessive, as shown in (15) . 15. a. la casa mia (Sp) the house my b. la eci mia ti the my c. *la maison mienne (Fr) d. la eci mienne ti Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 142 The difference in the grammaticality of (15a and c) could be accounted for if the deletion after movement is obligatory in French, but not in Spanish or whether the movement itself is obligatory. The pronominal form of the article and possessive (e.g. el mio), then, would reflect the deletion of the noun head. That the noun is, in fact, the head in the Spanish case is suggested from semantic interpretation according to Montalbetti (p.c.). In (16b) the deletion of the noun reflects a semantic interpretation differing from the possessive shown in (16a) , which does not have an empty head. 16. a. Ese avi6n es mio. that airplane is mine b. Ese avi6n es el e mio. that airplane is (the) mine For Montalbetti (16b) reflects a strong sense of economic ownership, as in the English paraphrase, that plane is mine. Whereas, (16a) reflects a weaker sense of possession as the plane being the one I have a ticket for, although the ownership reading is still possible. Employing Vendler's (1968) notion of relative predication, Montalbetti explains the ambiguity of the Spanish example. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 143 The mio in (16b) is predicated of the empty head allowing for the relativization of the predication; whereas, the mio in (16a) without an empty head to predicate over permits the non-relativized reading.12 This contrast also reflects a difference with respect to theta-binding. In (16b) the presence of the determiner implies an independent variable so that the possessive is theta-bound by the determiner; whereas, in (16a) it is not. In French, there is also a pronominal form that consists of the article plus the possessive (e.g. le mien - the mine sg. masc.); however, this is not the result of what by necessity would be an obligatory delection of the noun after movement, since a postnominal possessive is not possible.13 Finally, a merger between the determiner and the possessive suggests an explanation of why the prenominal possessive under an adjectival analysis can not, unlike other adjectives, stand alone, as shown in (17) for French and (18) for Spanish. 12 Montalbetti distinguishes between STRONG possession, which is the paradigmatic non-relativized case (i.e. economic) and WEAK possession, which is the relativized case. (Montalbetti, p.c.) 13 I will leave to future research the explanation for the historical development from Old French to Modem French; however, for the time being I suggest that the difference between French and Spanish wrt the postnominal possessive may be attributed to the lack of pro-drop in Modem French and the availability of movement. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 144 17. a. le livre, il est bleu. the book, it is blue b. *le livre, il est mon. the book, it is my 18. a. el libro, es azul the book, (it) is blue b. *el libro, es mi the book, (it) is my Since the determiner is clitic-like in that it cannot stand alone, then the resultant merged form would also be clitic-like and, as such, be incapable of standing alone. Now, instead of two forms traditionally referred to as possessive adjective and possessive pronoun, there is one underlying form: article + possessive + noun. The traditional possessive adjective results from a morphological merger between the article and the possessive; whereas, the possessive pronoun results from the deletion of the noun, at least for Spanish. Furthermore, noun movement is optional in Spanish permitting a postnominal possessive. Since a syntactic merger assumes N-movement, the merger in Spanish is Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 145 syntactic and occurs late at a point prior to spell-out. For French, the merger is a lexical phenomenon resulting in the syntactic effect of no postnominal possessives. In the case of Italian the possessive and the determiner do not merge; instead the possessive which has been considered an adjective (Giorgi and Longobardi, 1991) has undergone a process of pronominalization by incorporating the adjective into the empty noun site. This accounts for the pronoun-like behavior of the Italian possessive that was discussed in section 2.2 of chapter 2. The structure for merger and incorporation is given in (19) . 19. a. Incorporation b. Merger la mia la mia mi In (19a) the possessive raises, but does not adjoin to the determiner; whereas, in (19b), the possessive adjoins to the determiner. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 146 4.2.2 Conclusion Up to this point I have proposed a unified account for the distribution of the adjectival, pronominal possessive in Romance by suggesting the same underlying structure for the possessive in each of these languages in keeping with the current proposals emphasizing the similarities between languages (Chomsky 1992, Cinque 1993, Kayne 1994). By teasing apart the functional categories expressed in the possessive (as discussed in chapter 2), we are able to capture both the possessive's limited adjectival and primarily pronominal features, and to take a first step in accounting for parametric variation in possessive constructions. This proposal has provided an analysis that accounts for the distribution of the possessive without positing two types - adjective vs. pronoun. It provides for one form demonstrating both adjectival and pronominal (or perhaps more appropriately predicational, as was discussed in Chapter 2) features whose distribution and variation results, in part, from movement. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 147 4.3 Definiteness and the Possessive If, as I have proposed, the surface variation of possessives in Romance is based on a morphological merger14 of the definite article and the possessive, then the question arises as to what evidence there is for this merger. I propose that the differences with respect to definiteness between pre- and postnominal possessives reflects this merger.15 In Spanish and French, these prenominal possessives are marked +DEF as a result of their merger with the definite article, and as such do not require licensing for definiteness. Possessives in Italian and Portuguese, on the other hand, are not marked for definiteness and as such require licensing. Proposing the necessity of a definite D° for licensing a possessive is not without precedence, in discussing English possessives, Kayne (1993) claims a definite D° must be present, since a possessive AGRO is not sufficient to license a DP in its SPEC. According to Kayne this definite element in English 14 The concept of merger will be discussed in more detail in Chapter 5. 15 The notion that the possessive could vary in respect to definiteness is not without crosslinguistic support. Jim Huang (p.c.) reports that possessives in Chinese can have either a definite or indefinite interpretation depending on the environment. This suggests that the Chinese possessives are also not marked for definiteness. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 148 must be phonetically unrealized. However, if my in English is lexically marked as definite, then it is in the determiner position. If it gains its definiteness by merging with the determiner, as I claim for Romance, then the Possessive has moved to D°- 4.3.1 Existential Constructions Turning to empirical evidence for saying these possessives are not marked for definiteness in Italian and Portuguese, I will show that Italian and Portuguese possessives can occur in an environment where definites are excluded. One such environment involves existential "there". "There" insertion is not possible with a definite, as in (20) for English. 20. a. There is a man in the garden, b. * There is the man in the garden. Note, however, that existential constructions have been questioned as a diagnostic of indefiniteness. More than one reading is possible with there constructions in English. Woisetschlaeger (1983) observes three possible readings: existential, generic and list. In the latter two - generic and list - a definite is possible, as shown in (21) . Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 149 21. a. There is the wine we were going to use in the desert, (list) b. There was the air of the successful businessman about him. (generic) (woisetschlaeger, 1983) If we discount these two possible readings and concentrate only on an existential reading; then, following Milsark (1974) there predicates existence excluding definiteness which presupposes existence. It would follow, then, that if the possessive constructions, under discussion here, are used in an existential construction, their definiteness should be revealed.16 In Spanish these existential constructions with possessives are not possible, as in (22). 22. *Hay mis hermanas en el parque. There are my sisters in the park. 16 Note, however, that this assumes that the generalization concerning there-insertion applies to Italian and Portuguese. The structures involved with French/Spanish, on the one hand, and Italian/Portuguese, on the other, differ and bear further research. One possible explanation would be the nature of the definite article in these languages. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 150 This is also the case for French, as in (23) .17 23. *11 y a mes soeurs qui arrivent. There are my sisters that arrive. This follows from the definiteness of the possessive in both French and Spanish. Italian and Portuguese, on the other hand, allow for the possessive to occur in existentials. In both languages the possessive occurs in these constructions without the article, suggesting that the possessives are not marked for definiteness, as seen in (24) for Italian and (25) for Portuguese. 24. Ci sono sorelle mie che arrivano domani.18 There are sisters my that arriving tomorrow There are sisters of mine that are arriving tomorrow. 17 Note that (23) becomes grammatical with a list reading as in (i) , but not under an existential reading as in the contrast between (ii) and (iii). i. II y a Jean qui rentre dans la pi§ce. ii. il y a ion homme qui rentre dans la pi£ce. iii. *11 y a 1'homme qui rentre dans la piSce. 18 In Italian a generic reading is possible in an existential construction with the article, as in (i), which will be discussed more fully in section 4.4. i. Ci sono i tuoi amici in giardino. However it is not possible to get a definite or specific reading. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 151 25. Tem amigos meus que gostam de fumar. have friends my that like to smoke There are friends of mine that like to smoke. Grimshaw (1990), among others, suggests that the definiteness of a phrase is determined by that of its possessive, as in (26) (her 20). 26. a. There's a man's shirt on the chair. b. *There's the man's shirt on the chair. The NP in these cases has the definiteness of the possessive despite the fact that the determiner is associated with the possessor (man) and not the head (shirt) . If this is the case then the possessives permitted in existential expressions would also be indefinite. Note that it is only the postnominal form of the possessive without the article that can occur in existential expressions in Spanish, as shown in (27). 27. a. *Hay tus amigos en el jardin. There are your+pl friends+pl+masc in the garden b. Hay amigos tuyos en el jardin. There are friends your+pl+masc in the garden There are your friends in the garden. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 152 4.3.2 Partitive Constructions A second environment where definiteness effects (DE) can be observed involve partitives. Partitive constructions in French and Italian allow for en/ne- cliticization only when the NP is indefinite, as shown in (28) for French. 28. a. *11 s'est construit les maisons. There is built the houses. b. II s'est construit trois maisons. c. II s'en est construit trois. The definiteness of the possessive predicts that en/ne- cliticization with a possessive should be possible in Italian, but not in French.19 This is the case, as seen in (29) . 29. a. Ne ho visto uno mio. of them I have seen a my (It) 19 Note that with the dative construction the possessive is grammatical with en-cliticization. i. J'en ai trouv6 une a moi. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 153 b. *Ne ho visto il mio. c. *j'en ai trov6 la mienne. (Fr) d. *J’en ai trouv6 ma.20 Based on their occurrence in existential contexts, as well as the evidence from ne-cliticization, possessives in Italian and Portuguese can not be marked for definiteness. This merger, I will claim is the result of a process of cliticization. It is possible only with the definite article since it adds no additional semantic interpretation to the NP aside from definiteness. Chomsky's (1965) notion of recoverability would preclude the reiteration of definiteness with other determiners that contribute additional semantic information since the additional semantic information would not be recoverable, as in (30). 30. a. *chaque mien livre each my book b. *quelques miens livres some my books 20 Recall that Spanish and Portuguese do not have a comparable cliticization. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 154 c. *plusieurs miens livres several my book d. *ces miens livres these my books In (30a) chaque (each) carries the meaning of every one of two or more, in addition to the definiteness. In (30b) quelques (some) adds unspecificity. And in (30c) plusieurs (several) has the additional meaning of more than two but not many. Finally, in (30d) ces (these) has a deictic meaning. Nor are the examples in (30) possible with mon/mes, if as I claim these forms have already merged with the definite determiner then the determiner position would already be filled. All of these constructions are possible in Italian, as illustrated in (31). 31. a. ciascun mio libro21 each my book b. alcuni miei libri/qualche mio libro some my books 21 There seems to a degree of acceptability wrt to the position of the possessive here. For some speakers there is a clear preference for the possessive to be postnominal. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 155 c. questo/quel mio libro this/that my book (Giorgi and Longobardi, 1991: 154) with respect to the occurrence of cardinals with possessives, as in (32) and (33) for Spanish and Italian respectively, there appears to be a preference for the postnominal possessive. 32. a. dos amigos mxos two friends poss-lp-ms-pl two friends of mine b. *dos mis amigos 33. a. due amici miei b. ?due miei amici Cardinals, being weak binders, can co-occur with prenominal possessives in Italian, but not with prenominal possessives in Spanish. This provides further evidence for the definiteness of the prenominal possessive in Spanish. Given the referential nature of possessives, the possessive form would need to raise to D to check the reference feature in French and Spanish similar to the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 156 verb raising in French as a result of strong AGR. There the article and the possessive in French and Spanish, since they agree in phi-features would through a process of cliticization merge. The PF conditions would reflect the morphological properties. 4.3.3 Further Evidence Additional evidence supporting this analysis comes from the formation of indefinites and the use of emphasis. 4.3.3.1 Indefinites In considering the formation of the indefinite, it is not possible to say in French un mon livre, since the NP would be specified as both indefinite and definite; whereas in Italian un mio libro is grammatical, the possessive being licensed by the indefinite determiner, un. The indefiniteness of un in French is clear despite the ambiguity between "one" and "a". Even interpreted as a number, ua cannot occur with the possessive, as seen in (34a) ; although, a numeral can occur with the possessive, as seen in (34b). 34. a.*mon un livre my a/one book Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 157 b. mes trois livres my three books 4.3.3.2 Errqphasis Further evidence for this merger comes from the emphatic use of the determiner. In both Spanish and French, the use of the article gives the interpretation, "it’s the one which belongs to me". Emphasis22 is reflected in the use of the article with the possessive in postnominal position, as shown in (35a) or by the article plus the fully inflected unmerged form of the possessive as in (35b and d) . 35. a. la casa mia (Sp) the house my Art + N + Poss b. la mia c. *la maison mienne (Fr) d. la mienne e. *la mi 22 Here when we speak of emphasis we refer to focusing the hearer/reader's attention on a particular element. This is achieved in different ways, among them phonetic stress and word order. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 158 Recall that French does not allow for the postnominal form. The merged form, as in (35e) could not occur with the article as it already contains the features of the article (i.e. definiteness), is situated in the same syntactic position that the article occurs in, and as a clitic reguires a host, which in the case of the possessive is the noun. In Italian emphasis can, also, be reflected by the postnominal position of the possessive, as seen in (36) . 36. a. E il mio cane. It's (the) my dog. b. E il cane mio. It's (the) dog mine! 4.3.3.3 Romanian23 There are four types of possessive constructions in Romanian. The/ may occur postnominally, as in (37a) . 37. a. fratele sau brother-the his b. *frate sau 23 The Romanian data is from Dumitrescu and Saltarelli (1996) . For further discussion of Romanian possessive constructions see Comilescu (1992) and Grosu (1988) . Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 159 This possessive construction reguires the presence of the article cliticized to the preceding noun. Without the enclitic article the construction is ungrammatical. The possessive can also occur as a clitic attached to the noun, as in (38a). 38. a. frate-su brother-his b. *fratele-su Note that in this construction the possessive is a reduced form. Also the article cannot occur within this construction, as in (38b) . with an indefinite noun the possessive occurs with what has been called the possessive article preceding it, as in (39) . 39. un frate al meu a brother POSS-the my a brother of mine Dumitrescu and Saltarelli (1996) suggest that the so called possessive article is necessary to saturate the referencial variable within the possessive. Finally there is a prenominal possessive construction, as in (40) . Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 160 40. rapausatul meu frate late-the my brother my late brother This prenominal possessive only appears following a prenominal articulated adjective. In three of the four constructions the possessive occurs with a definite article. This article is necessary for saturating the referencial variable of the possessive predicate. This article is excluded from the reduced enclitic construction arguably because the determiner and the possessive have been conflated. 4.3.3.4 Marsian24 Additional support comes from the southern Italian dialect Marsian. In Marsian possessives only occur postnominally, either in unstressed (clitic) or stressed forms, as shown in (41) . 41. a. frate-me brother - Poss lp sg my brother 24 The Marsian facts were pointed out to me by Mario Saltarelli. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 161 b. ne/se frate maye a/the brother Poss lp sg ms a brother of mine In (41a) the possessive me is enclitic to the noun. This form cannot appear with either an indefinite or a definite article nor can it occur with a stressed or doubled possessive, as seen in (42). 42. a. *ne/se frate-me b. *ne/se frate-me maye In (4lb) the possessive maye is not enclitic and can only follow the noun. Contrary to the clitic form, the postnominal fully inflected form must occur with an article, as in (43). 43. *(ne/se) frate maye This parallels the reduced prenominal possessives we have considered in French and Spanish: mon and mi. The fact that these unstressed forms in Romanian and Marsian cannot occur with an article suggests, as I claim, that their Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 162 French and Spanish equivalents have "merged" with the article. At this point, I have proposed that Romance languages have the same under lying structure (or diasystem) for what have been traditionally called possessive adjectives and possessive pronouns. I have shown that Italian and Portuguese possessives are not marked for definiteness, because of their availability for existential constructions and their occurence with indefinite articles. Further I have provided evidence indicating that French and Spanish possessives are marked for definiteness as evidence of a merger between the article and the possessive as pari: of the cliticization process to account for the surface variation in Romance of these constructions. 4.3.4 Alternative explanations I have suggested that the motivation for the merger of the definite article and the possessive in Spanish and French was motivated by the article and the possessive both being marked for definiteness. However, the question arises as to whether the Spanish and French possessives are in themselves definite or that the definiteness comes from the determiner with which they have merged. In Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 163 looking at Old French (like Modern Italian) , it would appear that the possessive, in fact, is not marked for definiteness, as the stressed adjectival form can co-occur with an indefinite, shown in (44). 44. un mien ami a my friend (Judge and Healey: 41) Given this, then, the motivation for the merger may be other than definiteness. In the next section we will consider the role of the determiner and in Chapter 5 we will look at the cliticization of the prenominal possessive as possible motivations. 4.4 Determiners The evidence supporting a definite [+DEF] feature on the possessive is unfortunately not conclusive, since the full inflected form can occur in existential contexts, as shown in (27) in addition to the Old French data given in (44) . An alternative possibility that must be considered is that of a feature specification on the definite article or an expletive-nonexpletive contrast between the determiners. In Italian a generic reading is possible in an existential construction with the article, as in (45c) . Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 164 45. a. There is still (*the) soup in the pot. Eng. b. Todavia hay (*la) sopa en la olla. Sp. c. C'd ancora (la) zuppa nella pentola. It. However it is not possible to get a definite or specific reading.25 This suggests that it may, in fact, be the article that carries a degree of definiteness. This use is also possible with the possessive, as in (46) . 46. Ci sono i tuoi amici in giardino. There be+3rd+pl the+pl your+pl friends in garden There are your friends in the garden. Further support, suggesting that it is the feature specification of the article rather than the possessive involved, comes from the occurrence of the postnominal possessive in existential contexts in Spanish, as shown in (27) repeated here as (47) . 47. a. *Hay tus amigos en el jardin. There are your+pl friends+pl+masc in the garden 25 In this example, if the article is ommitted, one obtains a partitive reading. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. b. Hay amigos tuyos en el jardin. There are friends your+pl+masc in the garden There are your friends in the garden. However, if we can generalize Dobrovie-Sorin1 s (1993) DR (Determiner Raising) then it should be possible to get an existential reading with Italian and Portuguese, but not French and Spanish for possessives. Looking at the denotation and scope of indefinites, Dobrovie-Sorin suggests that only a DP that stays in its base position as a sister to NP saturates the predicate denoted by NP. This DP functions as a constant, that is it denotes an individual; it is referential. On the other hand, a DP whose determiner has been raised does not denote an individual. DR leaves an open variable which allows for three possible interpretations at LF: (a) existential quantification; (b) complex predicate formation; (c) generalized quantification. The reading depends on the syntactic position position of the open DP. She supports the distinction between open and closed DPs with arguments from clitic-doubling, WCO effects and donkey-ef fect s. Extending her analysis, in Italian and Portuguese the definite determiner would DR resulting in a non- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 166 referential open DP. In French and Spanish the determiner could not raise since if, as I suggest, it has merged with the possessive. This would suggest that the postnominal position could occur in French or Spanish in an existential environment. This is in fact the case for Spanish, as seen in (47b) above. The possibility of DR would then provide an explanation for why the article can occur in existential environments in Italian, as was seen in (46) above. Further consideration needs to be given to the nature of the determiner. Stowell (1987) proposes that the determiner serves two logical functions in what he terms a common noun phrase (CNP). The determiner either closes the predicate by binding its open argument or it converts the phrase into a referential expression. Since a possessive is inherently referential the determiner would be necessary to serve as the head of a referential DP. Functioning as the head of a referential DP, the determiner could, then, be an expletive filling a structural position, but not contributing definiteness to the phrase. Then the distinction would not be between a weak/strong feature specification on the determiner, but rather one of expletive/non-expletive. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 167 Whether there is a definiteness feature on the possessive or a variation in respect to the determiner is not crucial to the stucture; what is important is that there is a difference in features which will account for the variation. What remains to be accounted for are the incidences in Italian and Portuguese where the article does not occur. 4.4.1 Articles, Possessives and Inalienable Possession There are several environments in which either the possessive or the determiner may be ommitted. These environments involve inalienable possession. 4.4.1.1 Kinship The use of the article in Italian varies with respect to inalienable possession as shown with kinship terms in (48) . 48. a.mia sorella / la mia sorella my sister b.le mie sorelle / *mie sorelle (the) my sisters Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 168 c.la nostra nonna italiana / *nostra nonna italiana (the) our grandmother Italian d.il loro figlio / *loro figlio (the) their son Singular unmodified kinship terms, such as (48a) do not require the article, with the exception of loro (their) , as in (48d) . The article is required when the noun is plural, as in (48b) and when it is modified, as in (48c) . I suggest that the article can be dropped if there is one and only one possible referent involved. In (48b) the plural indicates more than one; whereas, in (48a) the singular does not imply the existence of other sisters, therefore there is only one referent. Likewise in (48c) the use of the modifier (italiana - Italian) specifies one referent from other possible referents. As in Italian, Portuguese does not require the article with singular kinship terms, as in (49) . 49. Meu pai estci. em casa. My father is in house (home). Inalienable possession includes body parts, physical and intellectual properties in addition to kinship. Tellier Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 169 (1988, 1989) classifies these as inherently relational and as such the possessive is an argument and optional. I propose that this optionality is in actuality the availability of two pronominal structures: one with the possessive and one with pro. The choice between these forms being dependent on pragmatics. Pro would, then, be generated in possessive structures involving inalienable possessive as in (50) . The article is necessary in these cases to license pro. 50. Giannii parlo alia proi mamma. Gianni spoke to the mother. Following Tellier (1988, 1989) ''mamma" being inalienable projects a possessor argument. This argument is coindexed in Italian with a preceding antecedent. This predicts that two possible readings exist for (51) . 51. Giannii parl6 a Mariaj e alia proi/j mamma. Gianni spoke to Maria and the mother Pro could either be coindexed with Gianni or with Maria. since both Gianni and Maria precede pro. And this is in fact the case, both readings are possible. The relation is Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 17 0 one of precedence rather than c-command, as evidenced by the coindexing possible in (52) .26 52. a.Giannij ha presentato Mariaj allai/j cugina. Gianni introduced Maria to the (his/her) cousin. b.Giannii ha presentato allai/*j cugina Mariaj. c.Il ragazzoi che ha veduto Giannij ha presentato Mariaj allai/j/k cugina. The boy who has seen G. presented M. to the cousin 26 Mario Saltarelli (p.c.) suggests that linear vs. hierarchical order is not the exclusive basis for coreferential interpretation. In (i) it is pragmatically clearer that it is the "fidanzata dell’amico"; whereas, in (ii) one is more likely talking about Mario’s fidanzata. i. L'amico di Gianni ha presentato Maria alia fidanzata. ii. L'arnica di Gianni ha presentato Mario alia fidanzata. Then, following the same reasoning the null possessive can on an extremely boarderline sense (*?) refer to Gianni's fidanzata. Note that with "sua" this interpretation improves a bit. iii. ??L'arnica di Gianni ha presentAto Maria alia sua f idanzata. The referentiality of the sentences would be even better in (iv). iv. ?L'arnica di Gianni ha presentato Maria alia fidanzata di lui. Now, even the last sentence would force hierarchical order but the contrast *?-??-? is still recorded. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. In (52a) the cousin could be either Gianni's or Maria's; however, in (52b) the cousin can only be Gianni's. In (52c) the cousin can be the boy's, Gianni's or Maria's. 4.4.1.2 Body parts Body parts, being inalienable possessions, also allow for the pro structure, as seen in (53) . 53. a. Giannii si taglid la proi mano. Gianni cut the hand b. Jei me suis lave les proi mains. I to me have washed the hands. In (53a) la mano (the hand) can only be Gianni's hand.27 French, also, follows this pattern as seen in (53b) . In this context the overt possessive sua is excluded because of the reflexive. The overt possessive sua makes disjoint 27 in Italian the possessive relation holds only as long as the body part is attached (Longobardi, p.c.). In (i) the article can be used, but not in (ii) if the hand has been severed from the arm. i. Gianni si taglid la mano. Gianni cut the hand, ii. *Gianni continud a tagliare la nano anche dopo che si era staccata. Gianni continued to cut the hand after it had been cut off. In (ii) the possessive must be used to get the interpretation that the hand is Gianni's. Note, however, that a reflexive clitic pronoun occurs with the verb in (i) . Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 172 reference possible only in non-reflexive constructions (Zubizarreta, p.c.). It is interesting to note that in Romanian if the possessive is used, as in (54), the hand cannot be mine, but is rather a hand that I had created. 54. M- am ars la mea mina I burnt at my hand If pro is possible in these constructions, then there should be a bound pronoun reading with a quantifier, as in (55) in addition to WCO effects. 55. Everyi son loves hisi mother. While Spanish has no interpretation as inalienable with the definite article28, both Italian and Portuguese do. A bound pronoun reading is allowed in Italian and Portuguese, as in (56). 28 Spanish resorts to the use of an object pronoun with the definite article to show possession without the possessive, as in (i) . i.a. Ricardo se aflojd la corbata. R. himself loosened the tie. R. loosened his (own) tie. b. Vio que ella se ponia las manos sobre la cara.... He saw her putting her hands over her face (Butt and Benjamin, 1988) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 173 56. a. Ognii figlio ama la proi madre. (It) b. Todoi filho ama a proi mae. (Po) French allows for an inalienable reading under rather restricted environments (Vergnaud and Zubizarreta, 1992) . Where an inalienable reading is possible, a bound pronoun reading is also possible, as in (57). 57. Chaquei enfant a lev6 la proi t§te. every child raised the head Every child raised his head. In respect to WCO effects, the Italian example (58a) is questionalbly grammatical, while the Portuguese (58b) is definitely out. 58. a. ?Chii la proi madre ama ti? whoi the proi mother loves ti whoi does the mother love ti b. *Quemi a proi mae ama ti? However, these constructions can be "saved” by stylistic inversion, as shown in (59) ; although, not with the expected meaning. The use here gives the question a rhetorical meaning. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 174 59. a. Chit amaj proi madre tj ti? b. Quemi amaj a proi mae tj ti? French demonstrates an apparent dialectical variation (Vergnaud, p.c.). While (60a) is out, (60b) is marginally acceptable for those who accept (60c). (Note that (60d) is acceptable). 60. a. *A qui sa mire plait-elle? to whom his mother pleases b. ?A quii la proi tete fait-elle mal ti? to whom the head aches c. ?La tete fait mal &. Jean, the head aches to J. d. La tete lui fait mal. the head him aches His head aches. 4.4.1.3 Extensions to English As in Romance languages, possession can be expressed with the definite article under certain conditions in English. Since the definite article refers to something uniquely identified in a context or within the general Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 175 knowledge shared by the speaker and hearer, the definite article can denote possession with body parts, as shown in (61) . 61. a. John banged himself on the forehead. b. John pulled her by the hair. c. The coach patted him on the back. d. Stop digging me in the ribs. e. My aunt complains of a pain in the stomach. In these examples the possessor is indicated by either the subject (61a) and (61e) or the object (61b) - (6ld) . In (61a) it is John's forehead, in (6lb) her hair, in (6lc) his back, and in (61d) my ribs. A further restriction on this construction is that the body part be within a prepositional complement. Should the direct object not be a body part, then the possessive must be used, as shown in (62) . 62. a. John threw the ball with his left hand, b. *John threw the ball with the left hand. The definite article cannot denote possession when the body part is the direct object, as in (63). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 176 63. I cut the hand. in (63) the hand cannot be my own, unlike the case in Romance languages where the hand would be my own and, in fact, must be in Romanian, as in (54), repeated here as (64) . 64. M- am ars la mea mina. I burnt my hand. In (64) the use of the possessive provides the interpretation that the hand is not mine, but one that I had created. This use does not extend to kinship terms as in Italian, shown in (65) . 65. a. Gianni parlo alia mamma. b. *John spoke to the mother. c. John spoke to mother. d. John spoke to his mother. In (65a) the mother is Gianni's; whereas, (65b) does not have this interpretation. In (65c) the interpretation of mother entails both John and the speaker; it cannot be Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 177 only John's mother. If the mother involved is not also the speaker's mother, then the possessive must be used as in (65d) . The English use of the article to denote possessive could be explained under Vergnaud and Zubizarreta' s (1992) proposal for an expletive article in French. However, Vergnaud and Zubizarreta claim that English does not have an expletive article. They relate the occurrence of the article for possession to a class of verbs that license metonymy. I suggest that English does have an expletive article, but the restrictions on its occurrence are more narrow than in French. I suggest that it is licensed by a locative-like preposition.29 If as Belvin (1993) suggests that locative have corresponds to inalienable possession, then the possible occurrence of an inalienable possession with a locative-like prepositional phrase should come as no surprise.30 29 German, also, appears to have an expletive article with inalienable possession. The German counterpart to (6le) is ungrammatical with the full article, as shown in (i) . (Martin Prinshom, p. c.) i. a. Sie hat sich am Kopt vertetet. she has self on+the head hit. She hit herself on the head, b. *Sie hat sich an dem Kopt vertetet. 30 I will leave the structural relation of the PP in (61) to future research. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 178 4.4.2 Articles, Possessives and Predicates The previous section illustrated that in addition to allowing the definite article to function as the possessive, Italian allows for the deletion of the article or a zero article31 with inalienable possession. Italian and Portuguese also allow for the deletion of the definite article or a zero article in certain other constructions: predicates and vocatives. One environment in Italian where the article may be dropped is when the possessive occurs as a predicate after essere (be) , as in (66) . 66. a. Questo denaro & mio. This money is mine, b. 6 tua questa borsa? Is this purse yours? 31 I hesitate to say that the article is deleted. Instead I would like to further explore the possibility of expanding on Longobardi's (class notes-Pall 1994 USC) proposal of a phonologically unrealized, content-filled article as opposed to a phonologically unrealized, contentless article. This gives us four possibilities: 1. phonologically realized and content-filled 2. phonologically realized and contentless (expletive) 3. phonologically unrealized and content-filled (zero article) 4. phonologically unrealized and contentless Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 179 c. Que biglietti sono suoi. Those tickets axe his/hers. This is also the case in Portuguese after ser (be). as in (67) . 67. 0 livro 6 meu. 4.4.2.1 Small Clause Analysis This may be accounted for with a small clause type analysis following Burzio’s (1986) analysis of be as a raising verb taking a small clause. The structure for (66a) would then be something like that shown in (68) The book is mine 68. VP V SC £ DP NP ZZ ZZ questo denaro mio pro Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 180 A small clause analysis would explain the predicative interpretation of the possessive in these examples. However, the meaning changes when the definite article is used, as seen in (69) . The presence of two determiners in (69b,c), but only one in (69a) indicates that distinct thematic structures are involved. 69. a. Questo denaro 4 mio. b. Questo denaro & il mio. c. Questo 4 il denaro mio. In (69a) there is a predication relation where the meaning entails the money being- my property. This would suggest following Montalbetti's discussion of Spanish (see section 4.2.1) that the node for [mio e] or [e mio] in Italian is an x1 in order to be referentially active. (69b) has a contrastive reading where it is my money not someone else’s. And in (69c) the meaning is this and not some other money is mine. While a small clause structure may capture the predicative aspect of these examples it does not provide for a means to explain the difference between (69a) and (69b) . The difference in meaning is derived from the presence/absence of a determiner to theta-bind the possessive. In (69a) the determiner questo theta-binds the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 181 possessive; whereas, in (69b) the possessive is theta- bound by the determiner il allowing for the different interpretations. 4.4.2.2 English Possessives and Predicates English does not allow this use with predicates in general, but does when the complement names a unique role, as in (70) . 70. a. *John was doctor, a'. John was a doctor. b. JFK was president in 1961. b'.JFK was the president in 1961. This use of a zero article in English is also used with restrictive apposition (7la), quasi-locative phrases (7lb) and temporal phrases (7lc). 71. a. USC quarterback Rob Johnson played against UCLA. a1. The USC quarterback, Rob Johnson, ..... b. we studied the Bible at church. b'. We studied the Bible at the church. c. The exam will be on Friday/at 10/at noon. c1. *The exam will be on the Friday/at the 10/at the noon. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 182 Other uses for the zero article involve means of transportation and communication that are preceded by the preposition by, meals, and some illnesses.32 Of particular interest is the use of this zero article after a preposition, since this is the environment that requires an article with the possessive in Spanish, as shown in (72) . 72. No hablo del tuyo sino del nuestro. no talk present + lp of+the your but of+the your I am not talking about mine, but about yours. 4.4.3 Vocatives Finally, the article is dropped in direct address, as seen in (73a) for Italian and (73b) for Portuguese. 73. a. Figlio mio, che fai? My son, what are you doing? b. Venha CcL, meu amigo! Come here, my friend! 32 The use of the zero article with institutions varies dialectically; (i) is possible in British English, but not in American English. i. He went to hospital. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 183 If vocatives are not DPs but NPs, then, as such the possessive in Italian and Portuguese would occur without the article, which is in fact the case. This predicts that the prenominal possessive would not be possible in Spanish and French, since this form is the result of the merger between the article and the possessive. This proves to be the case in the Castilian Spanish , as shown in (74). 74. a. Bueno, hijo mio, me voy. Well, my son, I'm going, b. Te aconsejo que no, amigo mio. I advise you not to, my friend. Latin American Spanish and French; however, prove problematic. Both allow for the prenomial form, as seen in (75a) for Latin American Spanish and in (75b) for French. 75. a. Bueno, mi hijo, me voy. b. Oui, mon amour! In these cases it is questionable as to whether the possessive denotes possession. It seems that the possessive is being used in an epithetical sense, as in (76) . Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 184 76. my dear lady If this is the case, then, they must be accounted for under epithets, which is beyond the scope of this work. 4.4.3.1 Expletive Article Vergnaud and Zubizarreta (1992) propose the existence of an expletive article for French. Their hypothesis states: 77. In French the definite determiner le may function as an expletive from the point of view of denotation. (Vergnaud and Zubizarreta, 1992: 615)33 Stowell (1987) as mentioned in Vergnaud and Zubizarreta (1992) suggests that DPs are referential arguments, while NPs are nonreferential predicates. Since the vocative is referential, it would be expressed as a DP; however, it is not argumental. Following in the Fregean tradition, Longobardi (1993) proposes that it is the determiner that turns a nominal expression into an argument34, then it is 33 Zubizarreta (p.c.) extends this hypothesis to Spanish. 34 cf. Higginbotham, 1983, 1985, 1987a. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 185 possible that the determiner in this case is also an expletive. The position required for the referential aspect of the vocative. The determiner is required to instanciate the referential property of a nominal. 4.4.3.2 English Vocatives As mentioned the vocative construction in Italian and Portuguese occurs without the article; this is also one of only two environments in English where singular count nouns can occur without an article, as in (78) . 78. a. Come here, boy! b. *Come here, the boy! These constructions in some cases permit the use of an article; whereas, in other cases they do not. In those cases where an article is allowed the interpretation changes with an overt article. The zero article indicates only the category of the objects. In (7lb) at church does not refer to a building, but to the institution; however, in (7lb') at the church refers to the building. These constructions can also contain a possessive, as shown in (79) . Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 186 79. a. JFK was our president in 1961. b. Our USC quarterback, Rob Johnson, ... c. We studied the Bible at our church. d. Christmas is during our winter. e. Come here, my son! The addition of the possessive has an effect on the interpretation, in (79b) the appositive reading is no longer available; however, the appositive reading is possible with a possessive, as in (80) . 80. My friend Dong-in completed his degree this year. In (79c) both the building and the institution reading are available. The possessive in (7 9d) has a contrastive interpretation. The winter referred to is that of the Northern hemisphere as opposed to the Southern hemisphere. The possessive with temporal expressions, e.g. times of the day or night, appear to have only this interpretation. In (7 9d) the possessive does not change the interpretation. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 187 4.5 Possessive Placement In Chapter 3, following Bowers (1993), I projected the possessive within a predication structure headed by POSS. This is consistent with Cinque (1993) placing adjectival elements to the left of the noun and then deriving variation as to surface placement via movement of the noun. Because of the nominal adjectival properties of the possessive, it must be in a configuration which allows for agreement. Because of the pronominal properties of the possessive, it must be in a configuration which allows for this feature to be checked. The former requires a position in relation to the noun with which it shares agreement features (i.e. gender, number), while the latter requires a position in relation to the Determiner, which licenses its referentiality. The licensing (or checking) domain of a head, (Chomsky 1992-1995, Marantz 1995)), includes four configurations representing relationships to a head: the Specifier, an adjoined head, an adjunction to the maximal projection of a head, and an adjunction to the Specifier, as shown in (81) . Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 188 81. a. Specifier XP / \ Spec XP' / \ X Complement b. Adjoined Head XP / \ Spec XP1 / \ X Complement / \ Y X c. Adjinaction to Maximal Projection XP / \ Y XP / \ Spec XP• / \ X Complement Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. d. Adjunction to Specifier 189 XP Spec XP’ /\ /\ Spec XP Y Spec X Complement If as I claim the determiner and the possessive merge in Spanish and French, then they must be in a configuration conducive to merger, such as that shown in (19b) , repeated here as (82). Therefore they are represented by the structure shown in (81d) - that of an adjunction to Spec. Structures (81a and c) would not account for the linear ordering of the article and the possessive and are, therefore, not considered. The structure in (81b) would not be available Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 82. a. Incorporation b. Merger mi 190 if following Hudson (1994) the determiner appears in the specifier of its own projection, which is headed by a nonovert referential element. In Italian and Portuguese while the determiner and the possessive form a constituent, they still permit a limited amount of material to intervene. They are projected in an incorporation configuration, as shown in (82a) . I will argue in the next chapter that, at least in respect to Italian, the possessive and determiner are shifting toward a merger configuration. Looking again at the structure proposed in Chapter 3, repeated here as (83) , we can account for the Spanish and French prenominal possessive construction, as well as the Italian and Portuguese prenominal possessive construction. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 191 83. a. DP X X la D' D mia POSSP \ Poss' t NP i casa b. DP la mia D' 4"* POSSP Poss NP casa la+mia=mi The possessive is projected in the Head of POSSP as a predicate. The agreement features are checked in their respective positions via Spec-Head agreement. Finally the possessive moves to the checking domain for its pronominal feature(s) and raises to the Head of DP in Spanish and French and then moves to the spec of DP, following Martin (1995) where I claim it adjoins to the determiner in Spec of DP and merges with the determiner, as shown in (83b) .35 35 Note that if a similar structure is proposed for English, the adjunction to the maximal projection position would still be available for the possible ocurrence of a quantifier with the possessive, as in (i). i. all my books Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 192 In Italian and Portuguese the possessive raises to the head of DP and does not adjoin, as shown in (83a) . In Chapter 5, I will argue for this configuration in French and Spanish. However, through a process of cliticization the article and possessive have mergered morphologically and cliticized to the noun in these languages. Following Marantz (1988), morphological merger will be discussed. Finally the status of the prenominal possessive in Romance as a clitic will be considered, since person agreement, following Aoun (1985), involves a nominal morpheme such as a pronominal clitic or AGR. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 193 Chapter 5 Cliticization and Merger In Chapter four we propose that the Romance languages under discussion have the same proto-structure: article + possessive + noun. We further claim that the variation between Italian-like languages and French-like languages is a result of restructuring due to the cliticization of the prenominal possessive. Through an historical process of cliticization the article and the prenominal possessive have merged morphologically and this merged form has become clitic-like in its behavior. 5.1 Merger Marantz (1988) in looking at verbal clitics observes the problem they present with respect to their bracketing in phonological structure and their assignment of syntactic category labels. Working within a GB framework he observes the parallel between clitic constructions and morphological bracketing paradoxes, as shown in (l) and (2) (his (2) and (3). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 194 The bracketing paradox in (1) results from the mismatch between the syntactic/semantic structures and the phonological structures. In (la) unhappier means more [not happy, not not [more happy]; however, phonologically the un- attaches outside of the -er, since -er does not attach to trisyllabic adjectives. (Marantz (1988: 254) 1. Semantic/syntactic structure Phonological Structure a. [ [un happy] er] [un [happy er] ] b. [ n[adj un [adj grammatical] ity] ] [un [grammaticality] ] 2. a. Mi ni-m-big-ish-ize mwarna ru:hu-y-e I sp-op-hit-cause-t/a child self I made the child hit himself (sp, subject prefix; op, object prefix; t/a, tense/aspect) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 195 b. Semantic/syntactic structure S -ish- mwarna CAUSE child c. Phonological structure -big- ru:hu hit self NP VP V -big - ish- NP NP ru:hu mwarna In this example (2) from Chi-Mwi:ni, a semantically and grammatically independent morpheme (i.e. the causative affix -ish-) becomes part of a derived causative verb. That is to say "a syntactically independent constituent ends up phonologically as part of a derived word." (Marantz 1988: 254) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 196 To resolve this "mismatch", Marantz proposes the mapping principle given in (3) . 3. THE MAPPING PRINCIPLE. An S-structure relation borne by B with respect to A correspondsto one or more of the following phonological relations: A structurely governs B (asymmetric sisterhood in the phonological structure tree) , A is linearly adjacent to B (left/right adjacency) , A morphologically casemarks B (morphology determined by A appears on B or the head of B), A agrees with B (morphology determined by B appears on A or the head of A) . (Marantz 1988: 258) A principle difference between the bracketing paradoxes in (1) and the clitic constructions in (2) involves the linear ordering of the clitics. The position of the clitic is related to its syntactic function in the sentence; whereas, the positioning of the affixes is determined by their status as prefixes or suffixes. To account for the ordering with respect to clitics Marantz proposes the principle of morphological merger as given in (4) . Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 197 4. MORPHOLOGICAL MERGER. At any level of syntactic analysis (D-structure, S-structure, phonological structure), a relation between X and Y may be replaced by (expressed by) the affixation of the lexical head of X to the lexical head of Y. (Marantz, 1988: 261) In the case of the prenominal possessive, the head of the POSSP has moved to the head of the DP resulting in the article + possessive structure. It is within this configuration that the determiner features of definiteness and referentiality are checked. For Italian-like languages this structure is what you see at the level previously referred to in theory as surface structure. For French- like languages the possessive moves further to Spec of DP where it has cliticized resulting in a merged form that fills the Spec position of the DP at what was previously referred to as surface structure. If this is the case, it would suggest that Italian-like languages may also develop a clitic-like definite prenominal possessive. And, in fact, there is some evidence as to such a developement in the dialectical variation seen within Italian. Tekavcic (1972) citing Rohlfs (1968, 1969) reports several dialects Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 198 using reduced forms similar to Spanish: notably Piemontese and Lombardo. 5.1.1 Prosodic influences That Italian may be moving in this direction is suggested by the prosodic conditioning of the prenominal possessive. On the surface there is evidence that the prenominal form is a reduced form. Observe the prosodic contrast between the following given in (5) . In both the independent form (5a) and the postnominal form (5b) the possessive mia is bisyllabic and receives prosodic stress on the penultimate. In (5c) the syllable consisting of the possessive merges into one resulting in a rising diphthong. 5.1.2 Adjective Placement In chapter 4 it was noted that the article and the possessive in Italian form a constituent and material separating the possessive and the determiner is limited or Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. * * * * * * * * * * * ** ★ * * * * * a. [la mia] b. [la kasa mia] the Poss-lp sg the house Poss c. [la mya kasa] the Poss house 199 results in a reduction in grammaticality. This does appear to be the case; however, variation is permitted with respect to word order when there is a numeral quantifier with the article or the article is replaced by a demonstrative, as in (6) and (7), respectively. 6. a. le mie due belle case the my two beautiful houses my two beautiful houses b. le due mie belle case c. ?le due belle mie case d. ??la bella mia casa The position of the numeral with respect to the article and possessive is, I suggest, related to its quantificational scope. That is whether it is quanifying over the possession of beautiful houses or over beautiful houses. Notice that there is a degradation in grammaticality when an adjective intervenes between the article and the possessive. This degradation increases further when the influence of the numeral quantifier is removed. In the case of the demonstrative in (7) . Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 200 7. a. Questa mia bella casa this my beautiful house b. questa bella mia casa c. quelle due belle mie case these two beautiful my houses these my two beautiful houses d. quelle mie due belle casa e. quelle due mie belle case Without the presence of the article, here replaced by the demonstrative, questa (this) or quelle (these), the possessive behaves like an adjective. The demonstrative by virtue of its position in D is able to license the possessive but it is the position here that is important for licensing not the lexical item. Since the D position is filled the possessive is able to check its agreement features as well as the referentiality of the D position. The demonstrative, however, carries with it other features besides definiteness. It has a deictic feature as well which refers to location. Since pronominal elements also carry the deictic feature of person in relation to the situation of the utterance. This would prohibit the adjunction of the possessive to D because of the resulting deixis clash. The determiner and the possessive, not being Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 201 in an adjoined position, would permit elements to intervene between them. 5.1.2.1 Dialectical Variation in Word Order1 Given the relationship between the determiner and the possessive, if as I have argued a process of merger and cliticization is going on, then we would expect that intervening material between the determiner and the possessive would not be possible in dialects where the possessive is a reduced form or, at least, there would be a degradation in grammaticality. In the Italian dialect Tuscan, the possessive has a reduced form, as in (8) . 8. le su case the his house The possessive, su (your, his or her) , in ($8) is reduced in that it does not show gender agreement or number agreement with case (house) which is feminine and plural. According to Cocchi (p.c.) it is impossible to insert material between the article and a preposed possessive, as shown in (9) . 1 I am indebted to Paola Benincci. and Gloria Cocchi for their assistance wrt the Tuscan dialect. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 202 9. a. le su du case2 the his two houses your two houses b. le su du belle case the his two beautiful houses c. *le du su case d. *le belle su case Neither a quantifier, as in (9c) nor an adjective, as in (9d) can intervene between an article and the reduced possessive. 5.1.2.2 Dialectical Variation and Demonstratives Further, given the proposed analysis, we would expect that determiners other than the definite article would not be permitted with the reduced form of the possessive or, at least, there would be a degradation in grammaticality. Again this proves to be the case. The use of the demonstrative with the reduced possessive is judged to be very marginal, as shown in (10). 2 Note the quantified due (two) is also a reduced form. It is reduced phonologically losing its last syllable which is not an inflectional affix. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 203 10. a. questa sua casa (Standard Italian) this his house b. ??questa su casa (Tuscan) 5.2 Possessives as Clitics In the preceding chapters, we have considered the forms that possessive constructions can take concentrating on pre- and postnominal possessive forms. The necessity of considering both the pronominal and adjectival features of these forms which have been traditionally treated as pronouns or adjectives has been shown. To fully acccount for the variation we have been discussing one further classification with respect to the prenominal possessives in Romance needs to be considered. 5.2.1 Previous Works Considering the prenominal possessive as a clitic is not new. Borer (1983) argues that the prenominal possessive in Hebrew is a clitic on the basis of clitic doubling effects, as seen in (11) . 11. a. beit-a house-her her house Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 204 b. beit ha-mora house the-teacher (fm.) the teacher's house c. beit-ai sel ha-morai house-her of the-teacher the teacher's house Arguments for the status of the prenominal possessive as a clitic have been advanced by McCloskey and Hale (1984) for Modern Irish, by Rivero (1986) for Spanish and by Aoun (1985) and Authier (1992) , among others, for French. 5.2.2 The Notion of Clitic Central to the notion of clitic is the idea of dependency. A clitic is dependent, in some way, on another lexical item. This dependency can be phonological, syntactic, morphological, or semantic. Moreover, a clitic may be syntactically/semantically dependent associated with one, while phonologically dependent on another, as is the case with the English possessive 's, as shown in (12a) and the French preposition de (of) in (12b) . 12. a. the porcupine over there's cage b. le porcupine du gargon Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 205 In (12a) the possessive is associated with the entire phrase syntactically and semantically, but is phonologically attached to the last element of the phrase. In (12b) the preposition de takes an NP as its object, but is phonologically combined with the article le resulting in the suppletive form du. (Marantz, 1988) As a result of this dependency, one diagnostic for a clitic is its inability to stand alone. Another diagnostic frequently sited for clitic-hood is related to stress. Kayne (1975) cites the inability of the subject pronouns in French to bear contrastive stress as evidence of their clitic-hood. This lack of stress in addition to their position relative to the item on which they are dependent provide the bases for Halpem's (1995) "rule of thumb diagnostics for clitic-hood." In attempting to define what constitutes a clitic, Halpern proses two basic diagnostics: "a) being (lexically) stressless/accentless; and b) occupying one of a characteristic set of positions." (Halpern 1995: 14)3 Other properties have been utilized to characterise clitics. Basing his analysis of subject pronouns in French on the properties of object clitics, Kayne (197 5) also 3 See Klavans (1985) for a discussion of clitic positions. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 206 cites lack of conjunction, in (13a) (his 64) and fixed order, in (13b) (his 65) . 13. a. *Jean la et le voit. J. object clitic-fm and obj. cl.-ms sees Jean sees her and him. b. Jean me le donnera. J. to me it will give Jean will give it to me. Rizzi (1986) supports the analysis of subject pronouns as clitics in French, but argues that their clitic-hood is essentially a phonological property, citing the compatibility of definite subject clitics with conjunction in (14a) as opposed to the impersonal subject clitic on (one), in (14b). 14. a. Elle chante et dance. She sings and dances. b. *On mange de la viande et boit du bon vin. One eats of the meat and drinks of the good wine. One eats meat and drinks good wine. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 207 In the next sections we will look at the different properties attributed to clitics in an attempt to arrive at a clearer notion of what constitutes a clitic. 5.2.3 Clitic Properties4 Valois (1991) suggests that the French prenominal possessive is a clitic, since there are constructions with the possessive in French that look suspiciously like clitic doubling. These are illustrated in (15). 15. a. sa maison lui Poss. house to him his house b. sa maison k elle Poss. house to her her house Additionally, what appears to be clitic-doubling is present in standard Spanish, as well as in Andean Spanish, as shown in (16) , respectively. 4 I will concern myself here with the classification of an element as a clitic, rather than, the interpretation of what a clitic is. Wrt interpretation I assume Aoun (197 9) in which he proposes that clitics are either the spell-out of case features or AGR features. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 208 16. a. su casa de 61 Standard Spanish Poss. house of he his house b. su casa de Pedro5 Andean Spanish Poss. house of Pedro Pedro's house Clitic pronouns are characterized as "weak" pronouns. As such they are found next to the element that they are an argument of and nothing can intervene between the clitic and that element, as shown in (17) with subject clitics in French. 17. *11, souvent, va au cin6ma. He often goes to the movies. (Jaeggli, 1981) Furthermore as Jaeggli (1981) observes for French subject clitics, the weak form of the pronoun cannot be modified, as shown in (18) . 5 The Andean Spanish examples are from Liliana Paredes. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 209 18. *Ils deux partiront demain. They two will leave tomorrow. Finally, the clitic forms cannot be coordinated with full NPs, as seen in (19) . 19. *Jean et je voulons aller au cinema. Jean and I want to go to the cinema. Tremblay (1991b) also argues for the status of the possessive adjective in French as a clitic employing their cooccurence with adjectives, as shown in (20) . 20. a. Ton beau livre your beautiful book b. *Le beau tien DEF beautiful yours (Tremblay, 1991) However, a similar treatment for Spanish is not possible, as shown by the grammaticality of (2lb) . 21. a. Mi libro azul my book blue Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 210 b. El mio azul the my blue If we are to apply these tests as evidence for clitic-hood, then the possessive in French is not a clitic, nor would it be in Spanish, as elements can intervene between the possessive and the element it is an argument of (22a) ; it can be modified (22b) and it can be coordinated (22c) for the different Romance languages. 22. a. mon gros livre (Fr) my large book b. mi primer gran amor (Sp) my first great love c. il mio libro e quello di Maria (It) the my book and that of Maria This suggests that we cannot maintain the status of the prenominal possessive as a clitic. However, looking more closely at the properties of clitics, it becomes clear that the Spanish and French prenominal possessives are cliticlike; whereas, the Italian possessives are not. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 211 5.2.3.1 Emphasis Perhaps the most often cited property of clitics in the generative tradition (Perlmutter, 1972, Kayne 1975, Jaeggli 1981, Valois 1991 among others) following the Romance tradition (Meyer-Lukke, 1890-1902; etc.) is that they cannot be emphasized. This applies to both prosodic emphasis with stress and situational emphasis through deixis. Emphasis in both these cases requires a stressed, strong, disjunctive pronoun form. However, this does not always hold as an identifying feature. A case in point - Greek clitics can be stressed. Satarelli (p.c.) reports that for some cases in Argentinian Spanish, final clitics can be prosodically stressed, as in (23). 23. a. diimelo give-me-it b. damel6 Furthermore, Saltarelli reports that for Southern Italian dialects they are prosodically stressed, as shown in (24) . 24. a. diimmela (Standard Italian) give-me-it Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 212 b. damm^lla (Southern Italian/Marsian) give-me-it With respect to the prenominal Spanish possessive it is unstressed and this form cannot be used for emphasis. For emphasis in stressing the notion of possession, Spanish uses the full form, as shown in (25) . 25. Es el mio. It's (the) mine. with emphasis, example (25) has the meaning "it's the one which belongs to me." 5.2.3.2 Alternation with strong forms Secondly clitic pronouns alternate with strong forms. Wanner (1987) claims that this alternation need have no necessary morpho-phonological cohesion, in the case of the prenominal Spanish possessives there appears to be a strong morpho-phonological cohesion. The prenominal form is a reduced form of the full form. While the full form is marked for person, number and gender, the reduced prenominal form is marked only for person and number. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 213 except for those cases already mentioned. This is illustrated in (26) . 26. mio/mia -> mi mios/mias -> mis Saltarelli (1986) notes that this loss of gender marking is characteristic of the progression from nominal to clitic object pronouns. 5.2.3.3 Syntactic Independence Thirdly, clitics cannot be used in isolation or after a preposition. This is also the case for the prenominal possessive form. After a preposition and in isolation, the full form is used as shown in (27) respectively. 27. a. En qu6 coche vamos? En el mio (*mi) . Which car are we going in? In (the) mine. b. Coge el mio (*mi). Take (the) mine. c. Qu6 vida tan triste la suya (*su) . What a sad life his/hers/yours/their is. In the case of object clitics, the clitic cannot be separated from the verb. This does not appear to be the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 214 case with the prenominal possessive, since adjectives can intervene between the possessive and the noun, as in (28) . 28. a. mi distinguido colega my distinguished colleague b. sus evasivas respuestas his evasive responses c. tu negra suerte your rotten luck Note that the prenominal position of adjectives is limited and depends on stylistic and other subjective factors; whereas, the Spanish form mi/mio is fixed in pre-/ versus post-position. The presence of an adjective between the possessive and the noun would appear problematic for a clitic analysis; however, if we adopt Stowell's (1981) treatment of prenominal adjectives in English, then we can account for the presence of an adjective between the possessive and the noun. Stowell suggests that prenominal adjectives in English form a complex noun by a rule of word formation. If this is also the case in Romance, then the adjective plus the noun forms a complex word and the possessive cliticizes to this canqplex form. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 215 Extending Raposo’s (1995) Outer Layer Principle6 for verbal clitics to nominal clitics will also account for the apparent intervention of adjectives between the possessive and the nominal head to which they cliticise. According to Raposo, a clitic is always the last element to attach to a given functional category. While it is generally assumed that a clitic which is syntactically associated with its host is also phonologically attached to the same host. This is not necessarily the case as Klavans (1985) proposes. She observes that in addition to the structural notions of precedence and dominance, the phonological notion of liaison must be taken into account, concluding that clitics are phrasal affixes. This is consistent with what we see with respect to possessives. 5.2.3.4 Placement A further property attributed to clitics involves placement. Depending on the form and function of the verb involved, clitic object pronouns are found on the right or the left, as illustrated in (29). 6 The Outer Layer Principle: Clitics may only appear in the outer layer of a functional category. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 216 29. a. Le di la carta. you/him/her (I) gave the letter I gave you/him/her the letter, b. Quiero hacerlo / lo quiero hacer (I) want to do it / it (I) want to do With finite forms as in (29a) the clitic precedes the verb, while with infinitives as in (29b) the clitic follows the verb or precedes by clitic climbing to the finite matrix verb. This fixed linear order between clitic and host is maintained with the possessives. The possessive form that we are considering as a clitic always precedes the noun. When the possessive follows the noun, the full inflected form is used, as in (30). 30. a. mis amigos my friends b. amigos mios (*mis) Since this postnominal form is not a clitic, possessive clitics do not "climb" as object clitics do and the fully inflected form is not found prenominally, nor is the clitic form found postnominally. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 217 5.2.3.5 Coordination Kayne (1975) observes that object clitics cannot-, be coordinated, as in Ola) (Kayne's 64), nor can they cliticise to a coordinated construction, as in (31b) (Kayne1s 111) . 31. a. *jean la et le voit. j. her and him sees J. sees her and him b. *Jean vous parlera et pardonnera. J. to you speak-future and pardon-future J. will speak to you and forgive you In considering subject clitics, recall Rizzi (1986) shows that these clitics can occur with a coordinated structure, as in (14a), repeated her as (32). 32. Elle chante et dance. She sings and dances. However, the subject clitics can not be coordinated, as in (33) . Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 218 33. *11 et elle chantent et dansent. He and she sing and dance Recall the discussion in chapter two, section 2.2.5.2 in which it was shown that these possessive forms can not be coordinated, as shown in (34) for both French and Spanish, respectively. 34.a. *mon et ton livre my and your book b. *mi y tu libro Coordination constructions with possessives, in Spanish, depend on the number of individuals involved; the second occurrence of the possessive can be deleted if the nouns refer to the same individual or aspects of the same thing, as shown in (35). 35. a. mi madre y mi padre my mother and rry father (different people b. my amigo y colega my friend and colleague (same person) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 219 c. su paciencia y valor his patience and courage (aspects of same virtue) (Butt and Benjamin, 1988) In (35b), there is one referent and thus one possessive. This correlates with the use of the determiner, shown in (36) . 36. el misterio o enigma del origen the mystery or enigma of the origin In respect to French, in coordinated structures the use of the possessive follows that of Spanish, where the possessive is repeated when reference is to different individuals or aspects, as shown in (37) . 37.a. ma m&re et mon pdre 1st sg fm mother and 1st sg ms father my mother and my father b. *ma m&re et p&re c. *ma m&re et soeur my mother and sister Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 220 d. mon ami et coll&gue my friend and colleague In (37d), one referent/individual is understood; while in (37a-c) there are two individuals involved. It is possible to have one possessive clitic with a coordinated structure in which separate individuals are involved, as shown in (3 8) for Spanish and French, respectively. 38. a. mis hermanas y hermanos b. mes soeurs et fr^res However, note that in these cases the individuals can be understood as a unique connected group (i.e. a group functioning as a single element) . This also holds for the singular, as in (39) for Spanish. 39. mi hermana y hermano The use of the possessive here differs from the use of the article in Spanish with coordinated structures. A single article can be used with a coordinated structure provided Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 221 both elements being coordinated are of the same gender, as in (40). 40. las aulas y los equipos the-fm classrooms-fm and the-ms equipment-ms the classrooms and equipment I suggest that a single possessive is possible even when nouns of different genders are involved because the lst- 3rd person singular prenominal possessive in Spanish is not marked for gender agreement. When the two elements of the coordinated structure do not form a cohesive group functioning as a single element, then both elements reguire the possessive, as shown in (41) . This is the case even when both elements are the same gender, as in (41b). 41. a. *Mi hermano y amiga estcin en el clase. My bother-ms and friend-fm are in the class, b. *Mi libro y reporte estcin en el piso. My book-ms and report-ms are on the floor Finally, even in the case where the two elements can be interpreted as a cohesive group functioning as a single Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 222 element, if these elements are further modified, both require a possessive, as in (42) . 42. a. mis libros y papeles my books-ms and papers-ms b. *Mis libros que estcin en el librero y paples que estcin en el piso necesitan emparcarse. My books that are on the shelf and papers that are on the floor need to be packed I propose that there are two different coordinated constructions involved: one in which the coordination is between NPs headed by a single DP accounting for the grammaticality of (38), (39) and (42a); and one in which the coordination is between two DPs accounting for the ungrammaticality of (41) and (42b). In the case of (42b) the addition of modification to the elements results in the loss of a cohesive group reading.7 Given this structural difference in coordinated constuctions, then the prenominal possessive clitics behave in the same manner as other clitics with respect to 7 The characterization of coordination is beyond the scope of this dissertation. For problems wrt the characterization of coordination see Goodall (1987) and Camacho (in progress). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 223 coordination. They cannot be coordinated nor do they attach to coordinated structures involving separate DPs. 5.2.3.6 Phonological Adjustment Finally, object clitic pronouns undergo some special phonological adjustments before a verb beginning with a vowel. In Italian they follow the usual vowel contraction rules, as shown in (43). 43. a. l'abbiamo vista (1' her (we) saw we saw her b. 11 amima sua the soul Poss. his/her soul These adjustments are present in French and Italian; however, in Spanish the transition between vowels of contiguous words is unaltered in writing and in slow speech, except with definite articles when the feminine form la is replaced with el if the next vowel begins with a stressed /a/, as in (44) , this is not restricted to clitics. =la) (l'=la) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 224 44. a. la hago [la:£go] it (fm) I do b. la amiga [laimiga] ~ [lamiga] the (fm) friend (fm) c. el (*la) alma bendita the (masc) (*fm) soul (fm) blessed (fm) the blessed soul 5.2.4 French Prenominal Possessives8 Employing the same tests for cliticness that were used for the Spanish possessive, let us consider the French prenominal possessive. At first blush, it would appear that the French possessives are not clitics as they can be emphasized, although the dative form is normally used for the emphatic form, as shown in (45) . 45. a. C'est MON livre. It is my book, b. C'est le livre k MOI. It is the book to me 8 Tremblay (1991a) also argues for the clitic status of the French prenominal possessives. She proposes that they are dative clitics citing, among other arguments, their being doubled by the dative possessive construction, as in (i). (i) mai mdre k moii my mother to me Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 225 In fact, all three forms of the possessive can be used to stress the notion of possession, as illustrated in (46) . 46. a. MON livre b. le livre k MOI le livre est LE MIEN Recall the discussion in section 5.2.3.1, that while lack of stress has been considered characteristic of clitics it can not be considered an identifying feature. Furthermore, Kayne (197 5) cites the lack of constrastive stress as characteristic of a verbal clitic. Just as a verbal clitic cannot bear contrastive stress, as in (47a) (Kayne's 59), the prenominal possessive cannot bear contrastive stress, as in (47b) . 47. a. *Jean LA pref&re. Jean HER prefers Jean prefers her b. *Jean aime SA m&re. Jean loves HIS mother Historically French, also, has both stressed and unstressed forms, as shown in (48); however, only the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 226 unstressed form is used today. The stressed form is an earlier form and now considered archaic (Grevisse, 1969) . 48. a. stressed form: masc. - mien(s) fan. - mienne(s) b. unstressed form: masc. - mon fern. - ma plural - mes Note that, in French, the stressed forms are marked for person, number and gender, while the unstressed forms are marked for person, number and gender only in the singular. In the plural, they are marked for only person and number. with respect to an alternation between weak and strong forms, the French possessive does appear to be clitic-like. It does alternate with a strong form, as given in (49). 49. a. mon livre b. le livre est le mien However, these forms do not show as strong a morpho- phonological cohesion as the Spanish forms do. Furthermore, there is less of a reduction in agreement Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 227 features between the weak and strong forms than evident in Spanish. For French the reduction in agreement is limited to the plural, where agreement is in number only. The singular form shows both number and gender agreement unlike the Spanish form, which does not show gender agreement in the singular. Despite the availability for emphasis and the lack of reduction in agreement, the French prenominal possessive does exhibit clitic-like behavior. These possessive forms cannot be used in isolation or after a preposition. After a preposition and in isolation, the strong or full form is used, as shown in (50) , respectively. 50. a. Dans quelle voiture allons-nous? In which car should we go? Dans la mienne (*ma) . In (the) mine (*my) . b. C'est le mien (*mon). Additionally, the French possessive patterns like the Spanish with respect to placement. While there can be intervening material between the possessive and its noun, this material is limited. In French, as in Spanish, only a limited number of adjectives can occur before the noun and Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 228 these may be accomodated following Stowell (1981) (as discussed for Spanish in section 5.2.3-3). These possessive forms can only be found preceding the noun. In French, there is no form that follows the noun, unlike Spanish which permits the full form to follow the noun. Finally, the prenominal possessive in Old French reflects the same phonological adjustment that the verbal clitics do. Elided forms of these clitics are used before auxiliaries or verbs starting with a vowel, as in (51) . 51. a. elle l'a trouv^e. she it (fern, sg) has found, b. m'aime my friend (fern, sg) In Old French, an elided form of the possessive was used before nouns beginning with a vowel, as shown in (5lb) . While this process is not found in Modem French, there is a denasalization when mon, ton or son are followed by a noun beginning with a mute h or a vowel, as in (52)9. 9 The denasalization described in Judge and Healy (1985) is not uncontroversial. Prunet (1992) argues that nasalization processes wrt mon and bon are not the same. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 229 52. a. mon livre -> /m5 livra/ b. mon habit -> /mO nabi/ c. mon auto -> /mo noto/ (Judge and Healy 1985: 39) In Modern French, instead of contracting, a morpho- phonological syncretism occurs and only the masculine form is used, as in (53). 53. a. mon ami my-ms friend-ms b. mon amie my-ms friend-fm c. *ma amie my-fm friend-fm 5.2.5 Clitic Parameters Since some elements analysed as clitics appear to exhibit a varying number of these properties, Wanner (19 87) proposes four parameters10 to determine the cliticness of an element. These parameters are as follows: 10 These are not parameters in a principles and parameters sense in defining linguistic variation (i.e. null subject parameter). They are more like classifications or the non-GB use of the term parameter to mean a standard of classification. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 230 1. PROSODY PARAMETER11: an element is marked either - clitic having full independent word behavior or + clitic having no stress property. 2. SYNTAX PARAMETER: an element is marked minus clitic having collocational autonomy meaning it exhibits a full range of displacement, e.g. wh-movement, dislocation, extraction. It is marked + clitic having total collocational control by a placement principle if there are coocurrence restrictions and a lack of dislocational potential. 3. PHONOLOGY PARAMETER: an element is marked - clitic when phonological rules are categorically free or + clitic when there are special rules; finally 4. MORPHOLOGY/LEXICON PARAMETER: an element is marked -clitic when there is lexical identity of all word manifestations; the element belongs to an open-ended set and follows normal morphological marking procedures. Or an element is marked + clitic when it belongs to a closed set with special morphological marking or the absence of marking. There is a further disintegration of the concept into clitic and non-clitic types. 11 Saltarelli (p.c.) points out that the prosody of Southern Italian clitics would contradict this. On the other hand, if one spoke of lexical stress it would be okay, but then conjunctions such as e, and, are clitics. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 231 We have seen that both the prenominal French and Spanish possessives exhibit many of the traditional properties of clitics; now in considering these parameters, we can gage the relative cliticness of these elements. These possessives do not exhibit full independent word behavior as they cannot function independently. Further the Spanish prenominal possessives are not stressed. So with respect to prosody Spanish prenominal possessives are + clitic. Since the French prenominal possessives can be stressed for emphasis, but phonologically unstressed, the value of this parameter would be neither plus nor minus. With respect to the syntax parameter, these possessives exhibit a strict linear order in relation to the noun and cannot be extracted or dislocated independent of the noun, as shown in (54), for both Spanish and French respectively. 54. a. *mii, quiero ti libro my, (I) want book b. *de mii, quiero uno ti libro of my (I) want a book c. *moni, je veux ti livre Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 232 Further, these possessives are not accessible to wh- movement. To question ownership the Spanish equivalent to the English "whose1 ' is "de quien" as shown in (55) . 55. a. Este is mi libro. this is my book b. £de quien es este libro? of who is this book c. *cquien libro es este? Likewise, in French, to question ownership you use the equivalent of the full form, k qui, as in (56) . 56. a. Cette est ma voiture. b. k qui (*qui) voiture est cette? However, this is not exactly the case with respect to wh- relative clauses. It would appear that in wh-relative clauses, it is the Poss-relative that needs to move to CP; thus, pied-piping the noun, as shown in (57). 57. Esta es una chica [cuya madre] vimos ti en TV. This is a girl whose mother we saw on TV. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 233 One could infer, however, that this is a characteristic of prosodically dependent clitics. With respect to the syntax parameter, then, these prenominal possessives are + clitic. Phonologically, it is not clear that special adjustments apply to the possessives or to other clitics. Because of the nature of Spanish phonology, in respect to this parameter, the value is neither plus or minus. For French the presence of phonological adjustments has been questioned, so for this parameter, French sould also be considered neither plus or minus. Finally regarding the morphology/lexicon parameter, possessives are a closed set of reduced forms that have non-clitic counterparts. For this parameter both possessives are also + clitic. 5.2.6 Italian Possessives Both the Spanish and the French prenominal possessives have been shown to be clitic-like. Italian prenominal possessives, on the other hand, are not clitics. The forms used for both the possessive pronoun and the possessive adjective are the same in Italian. These forms agree in number and gender with the possessed. Loro Ops pi - their) is the only form which never varies Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. and it is the only form that does not permit the deletion of the article in environments where the other forms permit this deletion.12 Since there is no variation in form, there is no alternation between weak and strong forms; therefore, Italian possessives are morphologically non-clitic. In considering the Italian possessive, the possessive plus the article must be viewed as a unit, as they generally occur together except within certain restricted environments13 (i.e. with kinship terms, in direct address and as a predicate after essere (be) or some other linking verb). As the possessive agrees with the object possessed, there is no distinction made between his and her, both are represented by il suo/la sua. Recall that in French to disambiguate the 3rd person form the dative a lui or a elle was added, giving the appearance of clitic doubling, as shown in (15), repeated here as (58). 12 Loro is the only invariable form, it never changes. I suggest that this might account for why Loro always occurs with the article. The article is necessary for checking of agreement, since only the article would show agreement as to number and gender. 13 These article-less environments were discussed in detail in chapter 4. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 235 58. a. sa maison k lui Poss. house to him his house b. sa maison k elle Poss. house to her her house This option, however, is not available in Italian. While these forms can be disambiguated by the addition of the prepositional phrases di lui (of him) or di lei (of her) , the possessive is not retained, as in (59). 59. a. l'amico di lui b. l'amico di lei c. *il suo amico di lei The possessive in Italian can occur pre- or postnominally; however, the postverbal position of the possessive may provide emphasis or contrast, as in (60) ;although, (60b-c) can be statements without emphasis on the demonstrative. 60. a. £ il mio cane. It's (the) my dog. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 236 b. & il cane mio. It's (the) dog mine! c. Questo d il mio cane. This is my dog. d. Questo & il cane mio. In (60c-d) the emphasis/focus would be on the demonstrative and both orders are possible. (Zubizarreta, p.c.) The postnominal position here would be contrastive with a reading of "this is my dog, not that one". In contrastive forms, Saltarelli (p.c.) reports that the postnominal form is phonologically stressed only when contrastive. In other words, word order is independent of contrastiveness, both pre- and postnominal possessives may be semantically contrastive, in which case you have phonologically contrastive stress. Stress is possible for both pre- and postnominal forms, as shown in (61) . 61. a. il MIO cane b. il cane MIO Italian possessives then are not clitic-like with respect to prosody. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 237 With respect to use in isolation or after a preposition, the Italian possessive can occur in both, as shown in (62). 62. a. & il mio is - 3ps sg the my it is mine b. in il mio -> nel mio in the my in mine The Italian possessives appear, then, not to be a clitic. The extent to which a prenominal possessive is clitic-like is one factor that will be shown to account for the variation in possessives. The following chart summarizes the clitic-like properties of these prenominal possessives. CLITIC. PARAMETERS ____________________Erench_________Spanish________Italian prosody +/- + SYNTAX + + PHONOLOGY + +/- -/+ MORPHOLOGY/LEXICON +______________ ±_______________- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 238 5.2.7 Possessive Clitics and Structure As clitics, the prenominal possessive in French and Spanish would identify a postnominal argument, which would occur as a means of marking possession and would be present in clitic doubling, as seen in (63) for French and (64) for Spanish. 63. a. le livre a moi the book to me = my book b. moni livre k moii my book to me = my book 64. a. el libro de usted the book of you = your book b. sui libro de ustedi 3p-sg Poss book of you = your book If the phrase headed by i . in French and de in Spanish is in an argument position, then the possessive clitic must be in a non-argument position, as argued by Tremblay (1991a) for French and Rivero (1986) for Spanish; otherwise there would be binding theory violations. Principle B would be violated since the pronoun would be bound in its governing category. Since these doubled Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. constructions are grammatical, then, the prenominal possessive must be in a non-argument position (i.e. the Spec of DP) as I proposed in chapter four, with respectto Italian, there is no clitic doubling effect with the possessive, as shown in (65) . This is what would be expected given the non-clitic status of the Italian prenominal possessive. 65. *il suo libro a lui 5.3 Summary In the preceding, I have attempted to provide a unified account for the distribution of the possessive in Romance by proposing the same underlying structure for the possessive in each of these languages under discussion in keeping with the current proposals emphasizing the similarities between languages (Chomsky 1992, Cinque 1993, Kayne 1994). By teasing apart the functional categories expressed in the possessive, we are able to capture both the possessive's adjectival and pronominal features, and to take a first step in accounting for parametric variation in possessive constructions. This proposal has provided an analysis that accounts for the distribution of the possessive without positing Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 240 two types - adjective vs. pronoun. The traditional division of possessives into adjectives and pronouns, while expedient and at times useful, obscures the interaction between these two aspects. Possessives appear to be adjectival only in their requiring agreement with the noun possessed. This agreement also reflects a predication relationship - a relationship that was argued for in chapter 2. With respect to their pronominal nature, possessives vary in their pronoun-like behavior. As seen in chapter 2 the pronominal status of English and Romance possessives is similar with respect to binding by a c- commanding QP and reflexive binding; although, they vary with respect to WCO effects. Following Chierchia's (1985) contention that a syntactic instantiation of semantics exists within the grammar, by projecting the possessive in POSSP as a predicate structure we are able to capture the semantics of possession within the syntactic structure; an endeavor mirrored in Dumitrescu and Saltarelli's (1996) treatment of articulated adjectives in Romanian. Thus two important issues have been raised. First, the validity of standard/traditional classifications has been questioned. This question has been raised outside of the generative tradition in work on fuzzy grammar (see Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 241 previous references), but has largely been ignored in generative work. The nature of the possessive demonstrates the need to carefully assess the classifications we assume. Secondly, the relationship between semantics and grammar, an issue that led to the development of generative semantics, is as yet unresolved. Reflecting the semantics of possession within the syntactic structure allows us to expalain their syntactic properties and behavior. Reference has been made to English possessives, but it remains for future research to determine whether these proposals can be extended crosslinguistically. Also remaining for future study is the discussion of other possessive structures, i.e. prepositional possessive constructions. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 242 Bib1iography Abney, S. 1987. The English Noun Phrase in its Sentential Aspect. PhD dissertation: MIT. Agard, F.B. 1984. A Course in Romance Linguistics: volume 1. Washington: Georgetown University Press. Allen, W.S. 1970. Vox latina. Cambridge: Cambridge University Press. Antrim, N. 1994. AP Possessives in Romance, ms. USC. Antrim, N. 1994. On the definiteness of possessives in Romance. Paper presented at LSA on January 8, 1994. Aoun, J. 1979. A short note on cliticization. ms. Aoun, J. 1985. A Grammar of Anaphora. Cambridge: MIT Press. Arteaga, D. 1995. Strong and weak possessives in Old French, ms. University of Nevada, Las Vegas. Authier, F.-M. 1992. is French a null subject language in the DP? Probus 4:1-16. Barker, C. 1995. Possessive Descriptions. Stanford: CSLI Publications. Belletti, A. 1990. Generalized Verb Movement. Rosenberg and Sellier. Belvin, R. 1993. The two causative haves are the two possessive haves. To appear in CLS 29. Bernstein, J. 1993. Topics in the Syntax of Nominal Structure Across Romance. PhD dissertation, CUNY. Borer, H. 1983. Parametric Syntax. Dordrecht: Foris. Bruneau, C. and F. Brunot. 1969. Precis de grammaire historique de la langue franpaise. Masson, Paris. Burzio, L. 1986. Italian Syntax: a government-binding approach. Dordrecht: D. Reidel Publishing Company. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 243 Butt, J. and C. Benjamin. 1988. A new reference grammar of Modern Spanish. London: Edward Arnold. Carstens, V. 1991. The morphology and syntax of determiner phrases in Kiswahili. Ph.D. dissertation, UCLA. Chomsky, N. 197 0. Remarks on nominalization in Jacobs, R. and P.S. Rosenbaum (eds.), English Transformational Grammar, 184-221. Chomsky, N. 1976. Conditions on rules of grammar. Linguistic Analysis 2:303-351. Chomsky, N. 1988. Lectures on Government and Binding. Dordrecht: Foris Publishing. Chomsky, N. 1992. A Minimalist Program for Linguistic Theory. MIT Occasional Papers in Linguistics 1, Cambridge. Chomsky, N. 1995. The Minimalist Program. Cambridge: MIT Press. Cinque, G. 1993. On the evidence for partial N movement in the Romance DP. ms. Universita degli Studi di Venezia. Camrie, B. 1981. Language Universals and Linguistic Typology. Chicago: University of Chicago Press. Comilescu, A. 1992. Remarks on the determiner system of Romanian: the demonstrative al and cel. Probus 4-3: 189-260. Comilescu, A. 1993. Notes on the structure of Romanian DP and the assignment of the genitive case in Working Papers in Linguistics 3:2. G. Giusti and L. Brug&, eds. University of Venice. Comilescu, A. 1995. Romanian genitive constructions in Advances in Romanian Linguistics. G, Cinque and G. Giusti, eds. Amsterdam: John Benjamins. 1-54. Corrigan, R., F. Eckman and M. Noonan, eds. 1989. Philadelphia: John Benjamins Publishing Company. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 244 Crume, G. 1952. A Grammar of the German Language. New York: Frederick Ungar Publishing Co. Delfitto, D. and J. Schroten. 1991. Bare Plurals and the number affix in DP. Probus 3.2: 155-185. Diesing, M. 1992. indefinites. MIT Press. Dobrovin-Sorin, C. 1993. On the denotation of indefinites. Talk presented at USC, March 9, 1993. Dumitrescu, D. and M. Saltarelli. 1996. Two types of predication modification: evidence from the articulated adjectives of Romanian. Paper presented at the XXVI Linguistic symposium on Romance Languages held March 28-30, 1996 in Mexico, D.F. Einhora, E. 1974. Old French: a concise handbook. Cambridge University Press. Fontanella de Weinberg, B. 1992. El espanol de America. Madrid: Mapfre. Foulet, L. 1930. Petite Syntaxe de L,Ancien Frangais. Freidin, R. 197 8. Cyclicity and the theory of grammar. LI 9:519-549. Giorgi, A. and G. Longobardi. 1991. The Syntax of Noun Phrases. Cambridge University Press. Ghameshi, J. and E. Ritter. 1996. Binding, Possessives, and the structure of DP. To appear in Proceedings of NELS 26. Goodall, G. 1987. Parallel structures in syntax. Cambridge: Cambridge University Press. Grevisse, M. 1969. Le Bon Usage: grammaire frangaise avec des remarques sur la langue frangaise d'aujourd'hui. Paris: Librairie A. Hatier. Grimshaw, J. 1990. Argument Structure. Cambridge: MIT Press. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 245 Gross, G. 1986. Etude Syntaxique du diterminant possessif. In Actes du colloque Diterminants: syntaxe et simantique, pp. 87-111. Paris: Klinksieck. Grosu, A. 1988. On the distribution of genitive phrases in Rumanian. Linguistics 26: 931-949. Halpern, A. 1995. On the placement and morphology of clitics. Stanford: CSLI Publications. Harris, M. and N. Vincent, eds. 1988. The Romance Languages. New York: Oxford University Press. Hawkins, R. 1981. Towards an account of the possessive constructions: NP's N and the N of NP. Journal of Linguistics 17. Higginbotham, J. 1983. Logical form, binding and nominals. Linguistic Inquiry 14:395-420. Higginbotham, J. 1985. On Semantics. Linguistic Inquiry 16, 4: 547-593. Higginbotham, J. 1987a. Indefiniteness and predication in The Representation of (In)definiteness. E.J. Reuland and A.G.B. ter Meulen, eds. Cambridge: MIT Press. 43- 70. Higginbotham, J. 1987b. Elucidations of Meaning. Lexicon Project Working Papers 19. Hudson, w. 1994. Pragmatic Constraints on Binding into Noun Phrases. PhD dissertation, USC. Jaeggli, 0. 1981. Topics in Romance Syntax. Dordrecht: Foris. Jackendoff, R. 1977. X Syntax: a study of phrase structure. Cambridge: MIT Press. Jesperson, 0. 1964. Essentials of English Grammar. University of Alabama Press. Jones, M. 1988. Sardinian in Harris, M. and Vincent, N., eds. The Romance Languages. New York: Oxford University Press. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 246 Judge, A. and F.G. Healey. 1985. A Reference Grammar of Modem French. New York: Hodder & Stoughton. Kayne, R. 1975. French Syntax. Cambridge: MIT Press. Kayne, R. 1993. Toward a Modular Theory of Auxiliary Selection, ms. to appear in Studia Linguistica 47. Kayne, R. 1993. The antisymmetry of syntax, ms. CUNY. Klavans, J. 1985. The independence of syntax and phonology in cliticization. Language 61:1. Koopman, H. 1993. The internal and external distribution of pronominal DPs. ms. UCLA. Koopman, H. and D. Sportiche. 1982. Variable and the bijection principle. The Linguistic Review 2: 365-91. Kratzer, A. 1989. Stage and individual Level Predicates, in Papers on Quantification. NSF Grant Report, Department of Linguistics, University of Massachusetts, Amherst. Lamarche, J. 1991. Problems for N°-movement to Num-P. Probus 3.2: 215-236. Langacker, R. w. 1968. Some observations on French possessives. Language 44. Larson, R. 1988. On the double object construction. LI 19: 335-391. Lausberg, H. 1965. Linguistica Ramcinica. Mhdrid: Biblioteca Rom&nica Hisp&nica. Lepschy, A. and G. Lepschy. 1988. The Italian Language Today. New York: Routledge. Longobardi, G. 1994. Reference and Proper Names: a Theory of N-movement in syntax and logical form. ms. University di Venezia. Lyons, C. 1986. The syntax of English genitive constructions. Journal of Linguistics 22:123-143. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 247 Mallinson, G. 1986. Rumanian. London: Croum Helm. Marantz, A. 1988. Clitics, Morphological Merger, and the Mapping to Phonological Structure in Theoretical Morphology. XXXXXXX, ed. Academic Press. Marantz, A. 1995. The Minimalist Program in Government and Binding Theory and the Minimalist Program. G. Webelhuth, ed. Cambridge: Blackwell. Martin, J. 1995. On the syntactic structure of Spanish noun phrases. PhD dissertation: University of Southern California. Martin, L. 1985. Una mi tacita de caf6: the indefinite article in Guatemalan Spanish. Hispania 68.2: 383- 387. Matthews, P.H. 1981. Syntax. Cambridge: Cambridge University Press. McCloskey, J. and K. Hale. 1984. On the syntax of person- number inflection in Modern Irish. Natural Language and Linguistic Theory l: 487-533. Milsark, G. 1974. Existential Sentences in English. Doctoral dissertation MIT. Cambridge, Mass. Moore, S. and T. Knott. 1965. The elements of Old English. Ann Arbor Michigan: The George Wahr Publishing Co. Napoli, D. 1989. Predication Theory: a case study for indexing theory. Cambridge: Cambridge University Press. Olsen, S. 1989. Das Possessivum: Pronomen, determinans oder adjektiv? Linguistische Berichte 120: 133-153. Parodi, C. 1994. On case and agreement in Spanish and English NPs, in Issues and Theory in Romance Linguistics, ed. M. Mazzola. Georgetown: Georgetown University Press. Partee, B. 1983. Campositionality. ms. Penny, R. 1991. A history of the Spanish language. Cambridge: Cambridge University Press. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 248 Picallo, C. (to appear) Possessive pronouns in Catalan and the Avoid Pronoun Principle. Catalan Working Papers in Linguistics. U.A.B. Picallo, C. 1990. Nominals and nominalization in Catalan. Ms. Universitat Autdnoma de Barcelona. Plank, F. 1992. Possessives and the distinction between determiners and modifiers (with special reference to German). Journal of Linguistics 28: 453-468. Pollock, J.-Y. 1989. Verb movement, UG and the structure of IP. Linguistic Inquiry 20:365-424. Prunet, J.-F. 1992. Spreading and locality domains in phonology. New York: Garland. Quirk, R., Greenbaum, S., Leech, G., and Svartvik, J. 1985. A comprehensive grammar of the English language. London: Longman. Rappaport, M. 1983. On the nature of derived nominals in L. Levin, M. Rappaport, and A. Zaenen, eds. Papers in Lexico-Functional Grammar, Indiana University Linguistics Club, Bloomington, Indiana, 113-142. Raposo, E. 1995. Clitic position and verb movement in European Portuguese, ms. University of California, Santa Barbara. Reid, A. 1991. Gramcitica Totonaca. Mexico, D.F.: Instituto Linguistico de Veranco, A.C. Ritter, E. (to appear) Two functional categories in Noun Phrases: evidence from Modem Hebrew in S. Rothstein (ed.) Syntax and Semantics 26, Perspectives on Phrase Structure: Heads and Licensing. New York: Academic Press. Ritter, E. 1996. Variation in the analysis of pronouns and anaphors. Paper presented at Grammaire et Variation (ACFAS) held at McGill University, May 13, 1996. Rivero, M.-L. 1986. Binding in NPs. In Generative Studies in Spanish Syntax. I. Bordelois, H. Contreras and K. Zagona, eds. Dordrecht: Foris Publications. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 249 Rizzi, L. 1986. On the status of subject clitics in Romance in Studies in Romance Linguistics. 0. Jaeggli and C. Silva Corvalan, eds. Dordrecht: Foris. Rohlfs, G. 1968, 1969. Grammatica storica della lingua italiana e dei suoi dialetti. Torino. Rothstein, S. 1983. The syntactic forms of predication. PhD dissertation: MIT. Saltarelli, M. 1986. Sincretismo funcional, evolucion, y la Castellana. Actas del II Congreso Intemacional Sobre El Espanol de America. Mexico, D.F.: Facultad de Filosofia y Letras. Scinchez, L. 1995. Syntactic structure in nominals: a comparative study of Spanish and Southern Quechua. PhD dissertation: University of Southern California. Siloni, T. 1994. Noun phrases and nominalization. PhD dissertation: Universite de Gen&ve. Stowell, T. 1987. Subjects, Specifers, and X-Bar Theory, ms. UCLA. Szabolcsi, A. 1983. The possessor that ran away from home. Linguistic Review 3. Taylor, J.R. 1989. Linguistic Categorization. Oxford: Clarendon Press. TekavCiC, P. 1972. Grammatica storica dell1 italiano, vol II: morfosintassi. Bologna: Society editrice il Mulino Tellier, C. 1988. Universal Licensing: implications for parasitic gap constructions. Ph.D. dissertation. McGill University. Tellier, C. 1989. Underived nominals and the projection principle: inherent possessors. NELS 20, University of Pittsburgh. Tellier, C. 1994. Copulas and P-Incorporation in French and English. Talk given at USC on March 17, 1994. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 250 Tremblay, M. 1991a. Possessives and datives: binary branching from the lexicon to syntax, dissertation McGill University. Tremblay, M. 1991b. The Syntax of Possession, in Views on Phrase Structure, K. Leffel and D. Bouchard, eds. Kluwer Academic Publishers. Valois, D. 1991. The Internal Syntax of DP. dissertation UCLA. Vergnaud. J.-R. and M. L. Zubizarreta. 1992. The definite determiner and inalienable constructions in French and in English. LI 23:4. Wanner, D. 1987. The Development of Romance Clitic Pronouns: from Latin to Old Romance. Berlin: Mouton de Gruyter. Wheeler, M. 1988. Catalan in Harris, M. and Vincent, N., eds. The Romance Languages. New York: Oxford University Press. Williams, E. 1980. Predication. Linguistic Inquiry 2.1. Williams, E. 1982. The NP Cycle. Linguistic Inquiry 13: 277-95. Williams, E. B. 1969. An Introductory Portuguese Grammar. New York: Dover Publications. woisetschlaeger, E. 1983. On the Question of Definiteness in "An Old Man's Book". LI 14:1. Zamparelli, R. 1993. Pre-nomial modifiers, degree phrases and the structure of AP. University of Venice Working Papers in Linguistics 3:1. Zimmermann, I. 1993. The Syntax of "Possessor" Phrases in The Parametrization of Universal Grammar. G. Fanselow, ed. Philadelphia: John Benjamins Publishing Company. Zubizarreta, M.L. 1987. Levels of Representation in the Lexicon and in the Syntax. Dordrecht: Foris Publications. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

On the representation and licensing of Q and Q-dependents

PDF

In the event of focus

PDF

The syntax of clitic doubling in modern Greek

PDF

On the nature of particles in Japanese and its theoretical implications

PDF

The syntax of focus and topic in Mandarin Chinese

PDF

The syntax of the Chinese BA-constructions and verb compounds: A morphosyntactic analysis

PDF

The life and literary and artistic activities of Robert Baldwin Ross, 1869-1918

PDF

Frequency in sentence comprehension

PDF

Language and context: toward a model of psycholinguistic competence

PDF

The impact of transcendentalism on the novels of herman melville

PDF

Definiteness types in Spanish: A study of natural discourse

PDF

Satirical social criticism in the novels of John Dos Passos

PDF

The puritan funeral sermon in seventeenth century England

PDF

Ernest Hemingway and the doctrine of true emotion

PDF

The phonemics and morphology of Hokkaido Ainu

PDF

A critical edition of the Escorial manuscript of 'historia de los indios de la neuva espana' of fray toribio de benavente (motolinia) (spanish text)

PDF

Syntactic structures in nominals: A comparative study of Spanish and Southern Quechua

PDF

Preview of the vanishing hero: a study of the protagonists in Jacobean drama

PDF

Emotions: Linguistic representation and cultural conceptualization

PDF

The accidence and syntax in john Wallis' 1653 "grammatica linguae anglicanae": a translation and a commentary on its alleged relationship to the 1660 port-royal "grammaire generale et raisonnee"

Asset Metadata

Creator Antrim, Nancy Mae (author)

Core Title On the status of possessives

Degree Doctor of Philosophy

Degree Program Linguistics

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag language, linguistics,Language, Modern,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c17-536951

Unique identifier UC11350066

Identifier 9720176.pdf (filename),usctheses-c17-536951 (legacy record id)

Legacy Identifier 9720176.pdf

Dmrecord 536951

Document Type Dissertation

Rights Antrim, Nancy Mae

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA