the_inflected_infinitive_in_romance.pdf

150

Upload: david-fernandez-sevilla

Post on 02-Dec-2015

50 views

Category:

Documents


7 download

DESCRIPTION

Emily Scida

TRANSCRIPT

  • AdministratorFile Attachment2000c454coverv05b.jpg

  • OUTSTANDING DlSSERTATIONS IN LINGUISTICS

    Edited by

    Laurence Horn Yale University

    A ROUTLEDGE SERIES

  • OUTSTANDING DISSERTATIONS IN LINGUISTICS

    LAURENCE HORN, General Editor

    PRONOUNS AND WORD ORDER IN OLD ENGLISH With Particular Reference to the Indefinite Pronoun Man

    Linda van Bergen

    ELLIPSIS AND WA-MARKING IN JAPANESE CONVERSATION John Fry

    WORKING MEMORY IN SENTENCE COMPREHENSION Processing Hindi Center Embeddings

    Shravan Vasishth

    INPUT-BASED PHONOLOGICAL ACQUISITION Tania S.Zamuner

    VIETNAMESE TONE A New Analysis

    Andrea Hoa Pham

    ORIGINS OF PREDICATES Evidence from Plains Cree

    Tomio Hirose

    CAUSES AND CONSEQUENCES OF WORD STRUCTURE Jennifer Hay

    THE PHONETICS AND PHONOLOGY OF GUTTURALS A Case Study from Juhoansi Amanda Miller-Ockhuizen

    TRUE TO FORM Rising and Falling Declaratives as Questions in English

    Christine Gunlogson

    PHONOLOGICAL AUGMENTATION IN PROMINENT POSITIONS Jennifer Smith

    CATEGORY NEUTRALITY

  • A Type-Logical Investigation Neal Whitman

    THE UPS AND DOWNS OF CHILD LANGUAGE Experimental Studies on Childrens Knowledge of Entailment Relations

    Andrea Gualmini

    MARKEDNESS AND FAITHFULNESS IN VOWEL SYSTEMS Viola Miglio

    THE SYNTAX-INFORMATION STRUCTURE INTERFACE Evidence from Spanish and English

    Eugenia Casielles-Suarez

    LENITION AND CONTRAST The Functional Consequences of Certain Phonetically Conditioned Sound Changes

    Naomi Gurevich

    SYNTACTIC FORM AND DISCOURSE FUNCTION IN NATURAL LANGUAGE GENERATION

    Cassandre Creswell

    THE INFLECTED INFINITIVE IN ROMANCE LANGUAGES Emily Scida

  • THE INFLECTED INFINITIVE IN ROMANCE LANGUAGES

    Emily Scida

    ROUTLEDGE New York & London

  • Published in 2004 by Routledge Taylor & Francis Group 270 Madison Avenue New York, NY 10016 http://www.routledge-ny.com/

    This edition published in the Taylor & Francis e-Library, 2005. To purchase your own copy of this or any of Taylor & Francis or Routledges collection of

    thousands of eBooks please go to http://www.ebookstore.tandf.co.uk/.

    Copyright 2004 by Taylor & Francis Group, a Division of T&F Informa.

    Routledge is an imprint of the Taylor & Francis Group.

    All rights reserved. No part of this book may be reprinted or reproduced or utilized in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including

    photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers.

    Catalog record is available from the Library of Congress

    ISBN 0-203-48876-8 Master e-book ISBN

    ISBN 0-203-58086-9 (Adobe e-Reader Format) ISBN 0-415-97106-3 (Print Edition)

  • For G.

  • Acknowledgments

    I would like to express my sincere thanks to the members of my committee-Linda Waugh, Jay Jasanoff, and especially Carol Rosenfor their guidance and advice as I was completing this project. I am also grateful to the Department of Romance Studies and the Department of Modern Languages at Cornell University for many years of support while in the graduate program. Any errors or shortcomings are entirely my own.

    I would like to thank my family for encouraging me along the way, and especially my parents for instilling in me a love of learning. I am also extremely grateful to Doris Borrelli, Christine Swain, Lisa Welton-Lair, and Julie Dogil for their devoted friendship, constant inspiration, and moral support without whom I would not have survived my many years at Cornell. Special thanks goes to Ti Alkire for reaching out to me during a difficult period, and for his encouragement and tremendously therapeutic sense of humor. I owe the completion of this thesis to the pact we made one October night.

  • Contents

    Chapter 1: Introduction 1 Chapter 2: The Syntactic Distribution of the Inflected Infinitive 6 Chapter 3: Previous Accounts of the Distribution of the Inflected Infinitive 31 Chapter 4: Theories of the Origin of the Portuguese Inflected Infinitive 74 Chapter 5: The Imperfect Subjunctive in Latin 93 Chapter 6: Distribution of the Inflected Infinitive in the Romance Languages 111 Chapter 7: Conclusion 126

    References 129

    Index 136

    THE INFLECTED INFINITIVE IN ROMANCE LANGUAGES OUTSTANDING DISSERTATIONS IN LINGUISTICS

  • CHAPTER 1 Introduction

    The inflected infinitive in Portuguese has een considered an anomaly among the Romance languages. No other major Romance language possesses an infinitive with verbal inflection, whereas in Portuguese the use and distribution of this form is quite extensive. In Portuguese the inflected infinitive exists along with an uninflected infinitive; their forms are similar except for the addition of verb endings which produces the inflected form, but their syntactic distribution is not entirely equivalent.

    amar I love amarmos we love amares you love amardes you love amar he/she loves amarem they love

    Two prominent issues have arisen regarding the inflected infinitive. The first concerns the distribution of the inflected infinitive in modern Portuguese; its occurrence in varied contexts has made a description of its use difficult. The second issue concerns its probable origin and development in Portuguese, which is obscure considering its absence in the other major Romance languages and the completion of its evolution prior to the appearance of Old Portuguese texts.

    Many attempts have been made to provide a description of the use of the inflected infinitive in Portuguese. Previous descriptions have included long lists of rules and examples intended to predict its use in such different contexts as the following:

    (1) O viajante viu entrarem e sarem as abelhas. The traveler saw the bees enter and leave (Corvo, Seleta Nacional, 350) (2) Frade, tu mentes, sem saberes que mentes. Friar, you lie, without knowing that you lie (Garrett, Viagens na minha terra, 17, 89) (3) Eu estou contentssimo de virmos para esta casa. I am very happy that we are coming to this house (Garrett, Frei Lus de Sousa, II, III)

    However, these previous attempts were based on an atheoretical approach that did not provide an explicit structural analysis. Consequently, the rules and norms proposed were inconsistent and contradictory, and allowed for numerous counterexamples. For almost any rule stated, an exception could be found in written and spoken language. For that reason, some have asserted that there are no dependable rules governing the use of the inflected infinitive in Portuguese. We will see that this claim is not entirely accurate.

    Using the framework of Relational Grammar, chapter two examines the syntactic distribution of the inflected infinitive in Portuguese to uncover the environments for its use. First, I provide a background on the theory of Relational Grammar, discussing various types of structures, including monoclausal and biclausal structures, and

  • diagnostics for clause membership. After a careful study of the structure and distribution of the inflected infinitive and related syntactic phenomena, I propose a general condition on its use which correctly predicts the precise range of environments where the inflected infinitive will be acceptable and others where it will be excluded. My analysis differs from many previous descriptions of the inflected infinitive in that it uses a theoretical approach to provide a precise syntactic investigation of the phenomenon. In contrast to other studies which present long lists of rules and norms, my analysis offers one concise condition restricting its use which predicts all possible occurrences of the inflected infinitive in Portuguese.

    Chapter three discusses the previous accounts given for the distribution of the inflected infinitive in Portuguese and Galician and examines the rules of each as they compare to my proposed condition on the use of the inflected infinitive. We will see that many of these previous attempts to explain its occurrence are vague and contradictory and fail to capture the true nature of the inflected infinitive in these languages. Instead, it will be shown that all the rules and examples given by other accounts are entirely predictable under my analysis. Although some of these accounts recognize important characteristics of the inflected infinitive, only my analysis presents a unified account of all the possible syntactic environments for this form in Portuguese and Galician.

    While the first section of this dissertation offers a synchronic study of the distribution of the inflected infinitive, the second section investigates another prominent issue regarding this form, which is its origin and development. Chapter four describes the theories that have been proposed to explain the development of the inflected infinitive in Portuguese. Early theories contended that the inflected infinitive was an innovation unique to Portuguese which arose out of analogy with the future subjunctive, or with the infinitive with nominative subject construction, or with both. After the discovery of an inflected infinitive in other Romance languages, some scholars proposed that the inflected infinitive was a historical development rooted in the Latin imperfect subjunctive. The inflected infinitive has been found to exist in languages other than Portuguese, namely Galician, Sardinian, Old Neapolitan, Old Leonese, and Mirandese.

    GALICIAN (4) Un amigo co que, contrariamente aos que teo agora, raramente falaba, xa que non eran necesarias as palabras pra nos entendermos . A friend with whom, contrary to those I have now, I rarely spoke, since words were not necessary for us to understand each other (A orella, 69)

    SARDINIAN (5) Juanne at tuncatu su barcone pro non s istremparet sa janna. John shut the window so that the door would not slam (Jones 1992:2978)

    OLD NEAPOLITAN

    The inflected infinitive in romance languages 2

  • (6) Lo re Laumedontaavesse negato albiergo a quilli Grieci, chi navegavano in altre parte, de non se potereno reposare. The king Laumedonta had denied shelter to those Greeks, who were travelling in other parts, so that they might not rest (Libro de la destructione de Troya, 18.68)

    OLD LEONESE (7) aqueste aver de suso dicho Recebimos de uos pora pagarmosdebda conoscida que deuiemos. that tribute above mentioned we received from you for us to pay the known debt that we owed (doc. LX, 1267 AD)

    MIRANDESE (8) a f de tener de que lo acusar . in order that they have something to accuse him of (Monteiro, Evangelhos, Revista de educao e ensino, IX, 255)

    In chapter four, I examine each of the theories of the origin of the inflected infinitive and discuss the factors which point to the Latin imperfect subjunctive as its origin.

    Those linguists who support the imperfect subjunctive theory have discussed the possible developments from Latin to Portuguese that produced the inflected infinitive in this language. However, none has provided a detailed comparison of the syntactic environments common to both the imperfect subjunctive and the inflected infinitive. Chapter five provides a close examination of the forms and the syntactic distribution of the imperfect subjunctive and of the inflected infinitive to determine in which contexts they correspond. Proponents have suggested that there is one principal environment common to both, namely volitional clauses, which for them is the probable locus of origin for the inflected infinitive. My study reveals an overall pattern of correspondence between the Latin and Romance forms. The most frequent occurrence of the inflected infinitive is found in adverbial clauses expressing purpose, an environment common to the Latin imperfect subjunctive. However, the distribution of the two forms corresponds in other contexts as well, which provides support for the theory that maintains that the inflected infinitive developed from the Latin imperfect subjunctive.

    The imperfect subjunctive theory offers as evidence: (1) the phonological and syntactic similarities between the imperfect subjunctive and the inflected infinitive, (2) the survival of the imperfect subjunctive in Vulgar Latin and early Romance, and (3) the existence of an inflected infinitive in Romance languages other than Portuguese and Galician. Chapter six studies the distribution of the inflected infinitive in Sardinian, Old Neapolitan, Old Leonese, and Mirandese. In Sardinian, it seems that the Latin imperfect subjunctive survives in finite contexts, with the same paradigm also occurring in contexts that are infinitival. Since there is little evidence of an evolution of the inflected infinitive in the other Romance languages, the distribution in Sardinian sheds light on its development and suggests an intermediate stage in which the imperfect subjunctive was being reanalyzed as an infinitive with inflection. This case offers further support for the imperfect subjunctive theory. This chapter also discusses the existence of other inflected non-finite forms, namely gerunds, present participles, and past participles in Galician, Old Neapolitan, and some European Portuguese dialects. The occurrence of inflection

    Introduction 3

  • with these forms demonstrates that speakers must in fact have reanalyzed the imperfect subjunctive as an infinitive with verbal inflection, creating the conditions for this pattern to spread to other non-finite forms.

    This study explores two important issues concerning the inflected infinitive. First, an examination of the syntactic distribution of this form in Portuguese under the framework of Relational Grammar reveals the exact environments for its use. I propose a general condition restricting the use of the inflected infinitive and, contrary to previous descriptions of the use of this form, my analysis presents a unified account of all possible occurrences of the inflected infinitive in Portuguese. Secondly, I discuss the origin and development of the inflected infinitive in the Romance languages. Taking into account the theories proposed to explain its origin, several facts from Latin and Romance point to the Latin imperfect subjunctive as the origin of the inflected infinitive. This study provides a precise comparison of the uses of the imperfect subjunctive and the inflected infinitive to determine in which contexts they coincide. The results reveal an overall pattern of correspondence, especially in complement and adverbial clauses expressing purpose, which further corroborates the imperfect subjunctive theory. Additional evidence which supports this theory is the existence of an inflected infinitive in other languages throughout the Romance territory, namely Galician, Sardinian, Old Neapolitan, Old Leonese, and Mirandese and the occurrence of inflection on other non-finite forms in some of these languages.

    The inflected infinitive in romance languages 4

  • CHAPTER 2 The Syntactic Distribution of the Inflected

    Infinitive

    The Portuguese inflected infinitive has been the topic of much attention and debate among grammarians and linguists who have attempted to describe its use with long lists of rules and examples. Unfortunately, these attempts have failed to capture the true nature of the inflected infinitive and instead have provided us with inconsistent rules for which counterexamples abound. I propose that an examination of the syntactic distribution of the inflected infinitive in Portuguese will uncover the environments for its use in Portuguese, Galician, and other Romance varieties discussed in chapters three and six. The following discussion is comprised of three sections. In the first section, I provide a description of the theory of Relational Grammar which will serve as a background to my subsequent discussion of Portuguese. Next, I examine the structure and distribution of the inflected infinitive in Portuguese and I propose a general condition on its use. Finally, I offer verifiable predictions regarding the occurrence of the inflected infinitive in Portuguese and other Romance languages.

    2.1 An Overview of Relational Grammar

    Relational Grammar is a general linguistic theory that seeks to characterize the set of syntactic constructions which recur across natural languages. To be able to carry this out, the theory defines three strategic objectives (Raposo 1981:57):

    (1) to characterize the class of grammatical constructions that can be found in natural languages, (2) to formulate linguistic universals which have as a domain of application the constructions referred to in (1), (3) to formulate, based on (1) and (2), grammars of individual natural languages.

    Relational Grammar is based on three basic claims about syntactic structure (Rosen 1984a:38): grammatical relations must be taken as primitives of syntactic theory; the structure of a clause may include a number of successive levels, called strata; and the same set of primitives (the grammatical relations) figures in all levels of structure from the initial through final. This set of grammatical relations is relevant to all grammars, and syntactic rules by and large are sensitive to these relations and not to linear order. The structure of a sentence is represented as a relational network, a graph consisting of arcs which registers all the grammatical relations that exist in a sentence.

  • An arc codes one atomic statement about structure, as in (4) (from Rosen 1988:78). (4)

    The R-sign is the name of a grammatical relation, and the co-ordinate is the level of structure at which the grammatical relation holds. This arc shows that the crowd bears the 2 relation in clause X in the initial stratum. A clause is a maximal set of arcs having the same tail, as in (5b). A stratum is a maximal set of arcs having the same tail and the same co-ordinate. Example (5), shown as both a relational network in (b) and in tabular notation in (c), is a clause consisting of one stratum in which the police bears the 1 relation, the crowd bears the 2 relation, and dispersed is the predicate.

    (5)

    The primitive grammatical relations in this theory include the term relations 1, 2, 3 which correspond to subject, direct object, and indirect object respectively. The nuclear relations are 1, 2 and the object relations are 2, 3. In addition, there are oblique relations (locative, benefactive, instrumental, comitative, etc.), chmeur (Cho), and predicate (P).

    The initial assignment of relations is determined by the relational valence of a predicate, which is registered in the lexicon of each language (Rosen 1988:12). The relational valence of a predicate tells us which grammatical relations must or can be assigned in the P-initial stratum of that predicate. Relational Grammar describes this in the following manner (Davies & Rosen 1988:578): (a) for each predicate, the lexicon supplies a valence which states what arguments the predicate must or can take; (b) the valence is empowered to license the corresponding arcs; (c) the nominals heading these arcs have thematic roles, and a mapping given in the lexical entry links a thematic role with each grammatical relation mentioned in the valence. Examples of different types of valences in English are shown in (6) (Rosen 1988:1214); bring is a transitive verb and its valence assigns obligatory 1 and 2 relations in the predicates initial stratum, illustrated in (7b). The valences of eat and accumulate demonstrate that a predicate can have optional as well as obligatory arguments. For instance, a sentence like Beavis ate

    The syntactic distribution of the inflected infinitive 7

  • exists along with Beavis ate the nachos, and junk accumulates has the alternate I accumulate junk.

    (6) (a) bring: 1 2 (b) eat: 1 (2) (c) accumulate: (1) 2 (7) (a) I brought the money. (b) 1 P 2 I brought the money (8) (a) 1 P (b) 1 P 2 Beavis ate Beavis ate the nachos (9) (a) 2 P (b) 1 P 2 1 P I accumulate junk junk accumulates

    Strata can be divided into three types. A transitive stratum is one that contains a 1-arc and a 2-arc, as in (7). An unergative stratum is one that contains a 1-arc and no 2-arc, illustrated in (8a). And (9a) has an initial unaccusative stratum, which is one that contains a 2-arc and no 1-arc. The last two types of strata fall under the heading of intransitive.

    The notion that a single clause may have more than one predicate, widely accepted now in Relational Grammar (Davies & Rosen 1988), suggests an economical way to represent complex predicates, periphrastics, auxiliation and the like. Each predicate determines a P-sector comprising all the strata in which it heads a P-arc. The first of these is the P-initial stratum for that predicate, and the last is the P-final stratum. Alongside a monostratal clause like (10), the sentence in (11) exemplifies a multipredicate clause in which the auxiliary is the second predicate in the same clause. The passive sentence in (12) is also an example of a multipredicate clause which has two auxiliaries. The clause has three P-sectors; the first is bistratal and the last two are monostratal. We will see ahead that the P-sectors of auxiliaries are always monostratal. In example (12) below, there are three P-sectors. The first stratum is the P-initial stratum of the verb eaten, and the second is its P-final stratum.

    (10) 1 P 2 Beavis ate the nachos (11) 1 P 2 ------------------------------------------------------

    ----- 1 P Cho 2 Beavis has eaten the nachos (12) 2 P 1 1 P Cho ------------------------------------------------------

    -------------- 1 P Cho Cho

    The inflected infinitive in romance languages 8

  • --------------------------------------------------------------------

    1 P Cho Cho Cho The nachos have been eaten by Beavis

    The preceding examples of multipredicate clauses differ from a similar construction called serialization. If we look at the Italian examples in (13) (from Rosen 1997), the main difference is that the serial verb diventare become inherits a 2 relation and the auxiliary essere be inherits a 1 relation.

    (13) (a) Eva diventa furiosa. Eva becomes furious 2 P -----------------------------------------

    --------------

    2 P Cho 1 P Cho Eva diventa furiosa (b) Eva furiosa. Eva is furious 2 P 1 P --------------------------------------- 1 P Cho Eva furiosa (c) Eva diventata furiosa. Eva became furious 2 P --------------------------------------------------

    ----------- 2 P Cho 1 P Cho --------------------------------------------------

    ----------- 1 P Cho Cho Eva diventata furiosa

    Example (13c) shows both auxiliation and serialization in the same clause. The principles of compactness and closure capture the differences between serialization and auxiliation (Rosen 1997:167). The compactness principle states that the P-initial stratum of a serial predicate can be preceded only by strata which are also P-initial. The closure principle maintains that the P-initial stratum of an auxiliary can be followed only by strata which are also P-initial strata of auxiliaries. This entails that P-initial strata of serial Ps must occur in an interrupted series at the beginning of a clause, and that the auxiliation zone is always clause final and always consists of monostratal P-sectors.

    Modal verbs in some Romance languages allow for either a monoclausal or a biclausal structure. Italian allows modal union optionally with the verbs volere want, dovere be

    The syntactic distribution of the inflected infinitive 9

  • obliged, and potere be able, as in the following sentences (Davies & Rosen 1988:5960):

    (14) Ugo wants to/has to/can explain it

    (a) Ugo vuole/deve/pu spiegarlo. (b) Ugo lo vuole/deve/pu spiegare.

    There is ample evidence that the construction in (14a) is biclausal and has the structure in (15a), while (14b) is a union having the monoclausal structure in (15b).

    (15) (a) Ugo deve spiegarlo. Ugo has to explain it

    (b) Ugo lo deve spiegare. Ugo has to explain it 1 2 P 1 2 P Cho Ugo

    deve spiegare

    Various clause-counting tests have been proposed to determine whether a sentence is monoclausal or biclausal. One such test is clitic position: clitics generally cliticize to the final P of their clause in the Romance languages 1 (Davies & Rosen 1988:59; Rosen 1990:431). The position of the clitic attached to the infinitive in (14a) shows that spiegare is its final P and that the sentence must be biclausal. In (14b) the pronoun representing the initial 2 of spiegare cliticizes to the outer predicate deve, identifying it to be the final P of that clause. This effect follows correctly from the claim that (14b) is monoclausal.

    An additional test proposed to determine clause membership is predicate clefting (Rizzi 1978), which produces an acceptable sentence if the cleft does not split a string of predicates belonging to one clause. The acceptability of the predicate clefting of (14a) in (16a) shows that the sentence has a biclausal structure. However, when a monoclausal

    The inflected infinitive in romance languages 10

  • sentence like (14b) undergoes predicate clefting as in (16b), the outcome is ungrammatical.

    (16) (a) Appunto, spiegarlo che Ugo vuole. Exactly, it is to explain it that Ugo

    wants (b) *Appunto, spiegare che Ugo lo vuole. *Exactly, it is to explain that Ugo wants

    it

    Another instance of union is causative clause union, which is proposed to be monoclausal in Italian and many other languages (Davies & Rosen 1988). Causative verbs initialize a new 1 relation, and in Italian the inner 1 generally revalues to 2 if intransitive and to 3 if transitive.

    (17) (a) Il babbo ha fatto parlare Nino. Father made Nino talk P 1 1 P Cho 2 1 P Cho Cho 2 il

    babbo ha fatto parlare Nino

    (b) Il babbo ha fatto accendere il fuoco a Nino.

    Father made Nino light the fire P 2 1 1 P Cho 2 3 1 P Cho Cho 2 3 il

    babbo ha fatto accendere il

    fuoco a

    Nino

    1 An exception to this rule exists in Romanian where the clitic attaches to the participle of the perfect periphrastic, but only if the clitic is feminine singular: Am vzut-o I saw her.

    Causative unions are shown to be monoclausal through certain clause-counting tests such as clitic position, predicate clefting, and passive. For example, in (18a) the cliticization of the clitic on far shows that far is the final P of the same clause where the pronouns originate, which in turn indicates that the whole construction is monoclausal. In addition, the impossibility of predicate clefting in (18b) shows that the causative construction cannot be biclausal.

    (18) (a) Il babbo glielo far accendere. Father will make him light it

    (b) *Appunto, accendere che il babbo glielo far.

    *Exactly, it is to light that Father will make him it

    Causative verbs are serializers because they do not inherit a 1 relation, and therefore according to the principle of compactness described above, auxiliaries, passive, and

    The syntactic distribution of the inflected infinitive 11

  • reflexivization are prohibited inside a causative union. To reiterate, the compactness principle states that the P-initial stratum of a serial predicate can be preceded only by strata which are also P-initial. Constructions with inner auxiliation and passive would be ungrammatical because their inner P-sectors would contain advancements, making them multistratal as in (19), which is disallowed by the compactness principle.

    (19) *Il babbo ha fatto essere acceso il fuoco.

    *Father made the fire be lit P 2 P 1

    ---------------------------------------------------------------------

    P Cho 1 -----------------------------------------------------------

    ---------- 1 P Cho Cho 2

    ---------------------------------------------------------------------

    1 P Cho Cho Cho 2 Il babbo ha fatto essere acceso il fuoco

    This overview of Relational Grammar will serve as a background and basis for my analysis of the distribution of the inflected infinitive in Portuguese which follows.

    2.2 The Distribution of the Portuguese Inflected Infinitive

    In the past many grammarians and linguists (see 3.1) have posited rules or norms to explain the distribution of the Portuguese inflected infinitive in such varied contexts as the following.

    (20) (a) preciso vocs terminarem o trabalho. It is necessary for you to finish the work (b) Tiger viu os colegas assassinarem a

    lagartixa. Tiger saw his colleagues kill the lizard (c) Ao sairmos do cinema, Beatrice deu com

    o seu psiquiatra. As we left the movies, Beatrice ran into her

    psychiatrist

    In many previous studies, rules for the inflected infinitive were based on an atheoretical approach which did not attempt any serious structural analysis. In addition, explanations for the use of the inflected infinitive ended up being contradictory and vague. For almost any rule stated, an exception could be found in written and spoken language. For that reason, some have asserted that there are no dependable rules governing the use of the inflected infinitive in Portuguese. However, after studying the syntactic distribution of

    The inflected infinitive in romance languages 12

  • the inflected infinitive in Portuguese and other Romance languages, we can posit a general condition limiting its use in these languages.

    I propose that the following is the principal condition governing the distribution of the inflected infinitive.

    (21) Condition on the Use of the Inflected Infinitive:

    The inflected infinitive must be the final predicate of its clause.

    In other words, the inflected infinitive occurs in the final stratum of its clause. This observation is very much in harmony with the fact that other types of person inflection are also limited to final predicates. The use of the inflected infinitive is never mandatory. Thus, this condition entails that the inflected infinitive can be used optionally in contexts in which the uninflected infinitive occurs, but only if it is a final predicate. If the infinitive is an inner predicate, as in monoclausal constructions such as causative or modal unions, it must appear in its uninflected form. Under the Government and Binding Theory, others have suggested that the inflected infinitive occurs in a CP in Galician (Longa 1994), Old Neapolitan (Vincent 1996), and Portuguese (Zaring 1985, King 1981, Raposo 1987) without proposing this to be a general condition restricting its use in these languages. 2

    The choice to use the inflected form of the infinitive in environments consistent with the condition in (21) is often influenced by stylistic considerations: the emphasis of the subject of the infinitive and the desire to avoid ambiguity are two frequent reasons to use the infinitive with personal endings. In addition, the presence of an overt subject with the infinitive tends to prompt the use of the inflected form of the infinitive in Portuguese.

    In order to perceive the effects of the above condition on the use of the inflected infinitive, one must be able to recognize the difference between monoclausal and biclausal constructions. The sentences illustrated in (22a) and (22b) both have monoclausal structures; the first example is monostratal and the second is a multipredicate clause with auxiliation in the final P-sector. In (22a) the final predicate of the clause is viu, while in (22b) tinha is the final predicate and visto is the inner predicate. Example (22c) has a biclausal structure in which the two verbs prometeu and ver are final predicates of separate clauses.

    (22) (a) Beatrice viu o seu psiquiatra. Beatrice saw her psychiatrist 1 P 2 Beatrice viu o seu psiquiatra (b) Beatrice tinha visto o seu psiquiatra. Beatrice had seen her psychiatrist 1 P 2 1 P Cho 2 Beatrice tinha visto o seu psiquiatra

    2For interesting discussions of Case, Control, Agreement and related issues under the framework of Government and Binding, see Quicoli (1996a), Quicoli (1996b), Safir (1996), Martins (1999), Mensching (2000), Miller (2000), and Pires (2000).

    The syntactic distribution of the inflected infinitive 13

  • (c) Beatrice prometeu ver o seu psiquiatra. Beatrice promised to see her psychiatrist.

    As presented in the previous section, there are various tests available to determine whether a sentence is monoclausal or biclausal. Two clause-counting tests that Raposo (1981:173202) proposes for Portuguese are clitic position and negation. In the Romance languages, clitic pronouns cliticize to the final P of their clause. If we take the examples in (22) and replace o seu psiquiatra with its corresponding clitic, we get the following results.

    (23) (a) Beatrice o viu. Beatrice saw him (b) Beatrice o tinha visto. Beatrice had seen

    him (c) Beatrice prometeu v-lo. Beatrice

    promised to see him *Beatrice o prometeu ver.

    The position of the clitic representing the 2 relation in (23a) and (23b) indicates that viu and tinha are the final predicates of their clauses, and specifically, of the same clause where the pronoun originates. This effect follows correctly from the claim that these sentences are monoclausal. In (23c) the position of the clitic attached to the infinitive shows that ver is the final predicate of its clause and that the sentence is biclausal. In other words, prometeu is in a superordinate clause and ver is in an embedded clause.

    One other test proposed to determine clause membership in Portuguese is negation. Raposo (1981:177) claims that negation of an inner predicate is disallowed and that only final predicates can be negated.

    (24) (a) Beatrice no viu o seu psiquiatra. Beatrice did not see her psychiatrist (b) Beatrice no tinha visto o seu psiquiatra. Beatrice hadnt seen her psychiatrist *Beatrice tinha no visto o seu psiquiatra. (c) Beatrice prometeu no ver o seu psiquiatra. Beatrice promised not to see her

    The inflected infinitive in romance languages 14

  • psychiatrist

    The position of negation on the verbs viu and tinha in (2 4a) and (24b) indicates that they are final predicates of their clauses and that these sentences are monoclausal. The negation of the infinitive in (24c) indicates that ver is the final predicate of its clause and confirms that the sentence has a biclausal structure.

    Modal verbs in Portuguese allow for either a monoclausal or a biclausal structure, as discussed for Italian in the previous section. 3 Portuguese allows modal union optionally with the verbs querer want, poder can, dever must, ir go, and vir come, as in the following sentences. As demonstrated by the variants below, clitic pronouns in Portuguese can be proclitic or enclitic depending on the speech style and dialect. In both (25a) and (25b), the clitic me cliticizes to the infinitive chamar. However, European Portuguese prefers enclisis as in (25a), while clitic pronouns in Brazilian Portuguese are generally proclitic as in (25b).

    (25) Julie wants to call me (a) Julie quer chamar-me. (b) Julie quer me chamar.

    In both (26a) and (26b) the clitic is cliticized to the modal verb quer. European Portuguese prefers enclisis as in (26a), while proclisis as in (26b) is typical in Brazilian Portuguese.

    (26) Julie wants to call me (a) Julie quer-me chamar. (b) Julie me quer chamar.

    Stylistic requirements dictate that a clitic may not begin a sentence, and that in negative, interrogative, and relative clauses, the mandatory position for a clitic is before the verb. These are tendencies that are often violated, especially in Brazilian Portuguese (Pizzini 1981:404; Galves 1996:227). The sentences in

    3The monoclausal construction with clitic climbing is used more frequently in European Portuguese than in Brazilian Portuguese, where it occurs almost exclusively in the written language. For a discussion of clitic climbing in Portuguese, see Quicoli (1976b), Pizzini (1981), and Comrie (1982).

    (25) are biclausal and have the structure in (27a), while (26) are examples of a monoclausal union construction which has the structure in (27b).

    (27) (a)

    The syntactic distribution of the inflected infinitive 15

  • (b) 1 2 P 1 2 P Cho Julie

    quer charmar

    Clause membership for these modal constructions can be verified using the clause-counting tests proposed for Portuguese. The position of the clitic pronoun on the infinitive in (25) indicates that the infinitive is the final P of its clause and that the sentence must be biclausal. In (26), the clitic is attached to the outer predicate quer identifying it to be the final predicate of the clause, which in turn indicates that the whole construction is monoclausal. Negation also provides us with the same effects.

    (28) (a) Julie quer no chamar-me Julie wants not to call me

    (b) Julie no me quer chamar Julie doesnt want to call me

    (c) *Julie me quer no chamar Julie wants not to call me

    Since only final predicates can be negated in Portuguese, the unacceptability of negation on the infinitive in (28c) follows from the fact that it is not a final predicate and that the sentence is monoclausal. In this sentence, only the outer predicate quer may be negated. The negation of the infinitive in (28a) is acceptable because the infinitive is the final predicate of its clause.

    Raposo recognizes two constructions for the causative verbs fazer make, mandar order, deixar let and the verbs of perception ver see, sentir feel, and ouvir hear in Portuguese (1981:117141). The first construction is parallel to the the causative union structure described in the previous section (2.1) for Italian.

    The inflected infinitive in romance languages 16

  • (29) (a) eu mandei vender os discos aos meus amigos.

    I ordered my friends to sell the records P 2 1 1 P Cho 2 3 eu mandei vender os

    discos aos meus amigos

    (b) Clarice viu cometer vrios erros s crianas.

    Clarice saw the children make several mistakes

    P 2 1 1 P Cho 2 3 Clarice viu cometer vrios

    erros s

    crianas

    In these monoclausal constructions, the causative and perception verbs initialize a new 1 relation and the inner 1 generally revalues to 2 if its P is intransitive, and to 3 if its P is transitive.

    Using the clause counting test of clitic position, we can confirm that these sentences are indeed monoclausal.

    (30) Eu lhos mandei vender. 4 I ordered them to sell them (31) Clarice lhos viu cometer. Clarice saw them make them

    In the above sentences the subject of the infinitive (os meus amigos; as crianas), which revalues to 3 in these causative unions, is realized as a dative clitic (lhes) which attaches to the outer predicates mandei and viu. The direct object of the infinitive (os discos; vrios erros) is realized as an accusative clitic (os) which also cliticizes to the causative and perception verbs. The position of

    4This construction with double clitics is characteristic of European Portuguese and is rare in Brazilian Portuguese (Pizzini 1981:404).

    these clitics shows that mandei and viu are the final predicates of the clauses where the clitics originate, which in turn indicates that the structure of these sentences is monoclausal.

    The inflected infinitive never occurs in this union construction, because the infinitive is the inner predicate of its clause in this context. This effect follows correctly from the condition on the use of the inflected infinitive described in (21) which restricts the occurrence of this form to contexts where it is the final predicate of its clause. In (32) the infinitive is the inner predicate and mandei and viu are the final predicates of their clauses as verified above. In these monoclausal sentences, the infinitive is not the final predicate of its clause and therefore cannot be inflected.

    The syntactic distribution of the inflected infinitive 17

  • (32) (a) *Eu mandei venderem os discos aos meus amigos.

    (b) *Clarice viu cometerem vrios erros s crianas.

    It is recognized that there are two possible constructions available for causative and perception verbs. The second construction in which these verbs can participate is the biclausal 2-control construction illustrated below. In this case, each of the verbs constitutes the final predicate of a separate clause. In addition, the two clauses share an argument, a nominal that bears both the 2 relation in the upstairs clause and the 1 relation in the downstairs clause. Whereas in the monoclausal construction in (29) the subject of the infinitive revalues to 3 and appears after the verb, the subject in the biclausal 2-control construction (os meus amigos, as crianas) appears directly before the infinitive and is in the accusative case.

    (33) (a) Eu mandei os meus amigos venderem os discos. I ordered my friends to sell the records

    (b) Clarice viu as crianas cometerem vrios erros. Clarice saw the children make several mistakes

    Since the infiniive is the final predicate of its clause in these biclausal structures, the inflected form of the infinitive may occur, as is demonstrated in these examples. In fact,

    The inflected infinitive in romance languages 18

  • its use in this environment is quite frequent, due to the strong tendency to use the inflected infinitive in the presence of an overt subject.

    Using the clitic position test for clause membership, we can verify that these constructions are indeed biclausal.

    (34) (a) Eu mandei-os venderem os discos. I ordered them to sell the records (b) Eu mandei-os venderem-nos. I ordered them to sell them (35) (a) Clarice viu-as cometerem vrios erros. Clarice saw them make several

    mistakes (b) Clarice viu-as cometerem-nos. Clarice saw them make them

    The subject of the infinitive in these sentences (os meus amigos; as crianas) is realized as an accusative clitic attached to the upstairs verbs mandei and viu. The direct object of the infinitive (os discos; vrios erros) is also realized as an accusative clitic; however, it cliticizes to the downstairs verb-the infinitive. Since clitics attach to the final predicate of their clause, we can clearly conclude that the infinitive in this case is the final predicate of its own separate clause.

    The occurrence of the inflected infinitive in this context follows correctly from the condition stated in (21) which restricts its use to environments where the infinitive is the final predicate of its clause. In the above biclausal sentences, the infinitive (venderem, cometerem) constitutes the final predicate of the downstairs clause, while the causative and perception verbs (mandei, viu) are the final predicates of the upstairs clause. Since the infinitive in this construction is the final predicate of its clause, it may occur in its inflected form.

    (36) (a) Eu mandei-os vender / venderem os discos.

    (b) Clarice viu-as cometer / cometerem vrios erros.

    Negation is another diagnostic test that distinguishes monoclausal and biclausal constructions. As stated above, only final predicates can be negated in Portuguese. The negation of an inner predicate is disallowed. In the monoclausal union construction proposed for causative and perception verbs, only the final predicates mandei and viu accept negation. Since the infinitives are inner predicates of the same clause, they cannot be negated in this structure.

    (37) (a) Eu no lhes mandei vender os discos, I did not order them to sell the records *Eu lhes mandei no vender os discos. (b) Eu no lhos mandei vender. I did not order them to sell them *Eu lhos mandei no vender. (38) (a) Clarice no lhes viu cometer nem um

    erro.

    The syntactic distribution of the inflected infinitive 19

  • Clarice did not see them make even one mistake

    *Clarice lhes viu no cometer nem um erro.

    (b) Clarice no lhos viu cometer. Clarice did not see them make them *Clarice lhos viu no cometer.

    In the biclausal construction for the same causative and perception verbs, negation is allowed on the upstairs verbs mandei and viu and also on the downstairs infinitives venderem and cometerem. The acceptability of negation on the infinitives in these examples indicates that they are the final predicates of their clauses. This effect follows from the claim that these constructions are biclausal. Negation in examples (39a) and (39b) shows that mandei and venderem are final predicates of two separate clauses.

    (39) (a) Eu no os mandei venderem os discos. I did not order them to sell the records (b) Eu mandei-os no venderem os discos. I ordered them not to sell the records

    The acceptability of negation with the infinitive in (40b) shows that venderem is the final predicate of its clause. In addition, the position of the clitic representing the 2 of the infinitive (os discos) in (40a) and (40b) indicates that the infinitive is a final predicate, since clitics attach to the final predicate of their clause.

    (40) (a) Eu no os mandei venderem-nos. I did not order them to sell them (b) Eu mandei-os no venderem-nos. I ordered them not to sell them

    Negation on viu and cometerem in (41a) and (41b) demonstrates that each is the final predicate of its own separate clause.

    (41) (a) Clarice no as viu cometerem vrios erros.

    Clarice did not see them make several mistakes

    (b) Clarice viu-as no cometerem nem um erro.

    Clarice saw them not make even one mistake

    In (42b), the acceptability of negation with the infinitive cometerem indicates that it is the final predicate of its clause. Additional evidence is the position of the accusative clitic representing vrios erros which is attached to the infinitive in (42a) and (42b); we can conclude from this that the infinitive is the final predicate, because clitics only attach to the final predicates of their clause.

    (42) (a) Clarice no as viu cometerem-nos. Clarice did not see them make them

    The inflected infinitive in romance languages 20

  • (b) Clarice viu-as no cometerem-nos. Clarice saw them not make them

    The position of negation and cliticization in these examples indicates whether a sentence has a monoclausal or biclausal structure. We can use these diagnostics to predict which types of constructions will allow or disallow the use of the inflected infinitive in Portuguese.

    2.3 Biclausal Constructions

    It is recognized that within the class of biclausal constructions there exist various subdivisions. These can be based on whether the subject of the embedded clause is a shared argument of the superordinate clause, and if so whether it is an initial argument of both clauses. In Portuguese, there are two types of biclausal constructions relevant to the distribution of the inflected infinitive: the basic structure and the 2-control structure.

    In the biclausal 2-control construction discussed above for causative and perception verbs, the superordinate and embedded clauses share an argument, a nominal that bears both the initial 2 relation in the superordinate clause and the final 1 relation in the embedded clause. In this construction, the two verbs are not clausemate predicates, but each is a predicate of its own clause. The subject of the embedded clause generally appears before the infinitive and is in the accusative case.

    (43) Nancy viu os seus amigos gastarem muito dinheiro. Nancy saw her friends spend a lot of money

    In this 2-control structure, logical entailment demonstrates that os seus amigos is a semantic argument of both clauses. For instance, the sentence in (43) entails both that Nancy saw her friends and that Nancy saw her friends spend a lot of money.

    In the basic structure, there are no shared arguments between the superordinate and embedded clauses. Since the subject of the embedded clause is clearly only a member of its own clause and is not assigned a semantic role by the superordinate verb, it appears in the nominative case. Verbs that participate in this structure include declarative/epistemic and factive/emotive verbs such as acreditar, afirmar, confirmar, dizer, duvidar, lamentar, pensar, etc.

    The syntactic distribution of the inflected infinitive 21

  • (44) Nancy pensa terem os seus amigos gasto muito dinheiro. Nancy thinks her friends spent a lot of money

    In this example, the subject of the embedded clause (os seus amigos) is not a shared semantic argument of the superordinate verb pensa. This is verified through logical entailment; in other words, it is not logically entailed that Nancy thinks her friends but only that Nancy thinks her friends spent a lot of money. Since os seus amigos is not assigned a semantic role by pensar, it does not occur as a shared argument in this structure.

    In both the basic structure and the 2-control construction, the subject of the embedded infinitival clause can appear as an overt nominal or can have a null realization. However, while in the biclausal 2-control construction the subject of the embedded infinitival clause can be realized as an accusative clitic attached to the superordinate verb, the corresponding realization in the basic structure is disallowed (Rouveret 1980:967; Pizzini 1981:412).

    (45) Nancy viu-os gastarem muito dinheiro. Nancy saw them spend a lot of money (46) *Nancy pensa-os terem gasto muito

    dinheiro. *Nancy thinks them to have spent a lot of

    money

    The example in (45) shows that the subject of the embedded clause (os seus amigos), which functions also as an argument of the superordinate clause, can occur as an accusative clitic attached to the matrix verb viu. However, in the basic structure in (46), raising of the embedded subject os seus amigos to the superordinate clause renders the sentence ungrammatical. Since it is never a 2 of the superordinate clause, the embedded subject cannot be represented by an accusative clitic attached to the matrix verb pensa in this structure.

    The acceptability of the inflected infinitive in these biclausal constructions follows from the condition in (21) that the inflected infinitive must be the final predicate of its clause. In both the basic structure and the 2-control construction, the matrix verb and the embedded infinitive are final predicates of separate clauses. The main difference between these two types of biclausal constructions is that in the 2-control construction, the superordinate and embedded clauses share an argument, a nominal that bears both the

    The inflected infinitive in romance languages 22

  • initial 2 relation in the superordinate clause and the final 1 relation in the embedded clause; on the other hand, in the basic structure there are no shared arguments between the superordinate and embedded clauses.

    2.4 Verifiable Predictions for the Use of the Inflected Infinitive

    Given the condition on the use of the inflected infinitive proposed in (21) and the syntactic effects of related phenomena, we can predict the precise environments where the inflected infinitive will be possible and others where it will be prohibited in Portuguese, Galician, Sardinian, and Old Neapolitan.

    In contexts where the infinitive occurs with an overt nominative subject, the inflected form is always possible. This effect follows correctly from my analysis in which the nominative subject signals a separate clause for the infinitive (examples from Maurer 1968:146; Jones 1992:297; Gondar 1978:64).

    (47) Quando os inglses se rirem de les terem muito dinheiro e ns pouco, toramos a orelha e choremos. When the English laugh about their having a lot of money and us little, lets repent and cry (Herc., Lendas, II, 198) [Portuguese] (48) Non keljo a cantares tue. I do not want you to sing [Sardinian] (49) o millor rxime que se pode maxinar pra iles viviren contentos e ditosos. It is the best method that one can imagine for them to live happy and fortunate (O porco, 112) [Galician]

    These examples are not control constructions; in other words, the embedded subject is not an argument of the superordinate clause. The case of the embedded subject will therefore be nominative, although not overtly marked in the above eles/iles. In these infinitival clauses, the presence of a nominative subject which bears the 1 relation marks a clause boundary and indicates that it is a separate embedded clause in which the infinitive is the final predicate. It follows correctly, then, that this is an environment in which the inflected infinitive can occur in these languages.

    The use of the inflected infinitive in particular types of constructions in these Romance languages follows from the condition in (21). The presence of auxiliation, passive, or reflexivization with an inner predicate automatically signals a separate clause in which the final predicate may be an inflected infinitive. As discussed in the previous section, auxiliation occurs only in the final P-sector(s) of a clause. This explains why the presence of an auxiliary or passive in a complex predicate indicates that it belongs to a separate clause, as in (50). In this example, the auxiliary terem must be the final predicate of the embedded infinitival clause and as such can occur inflected.

    (50) A firmo terem chegado os navios. I declare that the ships have arrived

    The syntactic distribution of the inflected infinitive 23

  • Similar examples in other Romance languages in which the presence of auxiliation signals a separate clause for the infinitive include the following:

    (51) Non credo di sseret ghiratu Juanne. I do not think that John has returned [Sardinian] (52) tenemo secondo la santa f cattolica essereno stati non homicidi li occidituri. We hold according to the holy Catholic faith that the killers were not murdered (Del Tuppo; see Rohlfs 1969: 709) [Old Neapolitan]

    In addition, since clitics cliticize to the final predicate of their clause in the major Romance languages, the presence of a reflexive clitic on an inner predicate signals a separate clause for that predicate, as in (53). These are contexts in which the inflected infinitive is possible, as is predictable under the present analysis since the inner predicate is the final predicate of its clause. In (53), the placement of the reflexive clitic se on the infinitive queixar indicates that the infinitive is the final predicate of its clause, and may occur in its inflected form. The acceptability of either form of the infinitive demonstrates that the use of the inflected form is optional in the contexts consistent with the condition in (21).

    (53) Pasquale ouviu-os queixar-se/queixarem-se. Pasquale heard them complain

    The inflected infinitive in romance languages 24

  • The presence of a reflexive clitic with the infinitive in these examples indicates that the infinitive is the final predicate of its own clause. The optional occurrence of the inflected infinitive in these constructions correctly follows from the condition stated in (21) that the inflected infinitive must be the final predicate of its clause.

    (54) Non se che ocurru nunca pensar que podasquedreste cego, que un da ao espertreste notaras os ollos valeiros? Didnt it occur to you that you could remain blind, that one day on waking up you would notice your empty eyes? (A orella, 55) [Galician] (55) Lo re Laumedontaavesse negato albiergo a quilli Grieci, chi navegavano in altre parte, de non se potereno reposare. The king Laumedonta had denied shelter to those Greeks, who were travelling in other parts, so that they might not rest (Libro de la destructione de Troya, 18.68) [Old Neapolitan] (56) Juanne at tuncatu su barcone pro non s istremparet sa janna. John shut the window so that the door would not slam [Sardinian]

    The acceptability of the inflected form can also be expected in exclamative and interrogative clauses like the following (from Maurer 1968:107; Gondar 1978:137). Since these sentences are independent clauses, there is nothing to which the infinitive is subordinate; therefore, the infinitive is clearly the final predicate, and may appear in its inflected form.

    (57) No comeres, tu que estalavas de fome, e ficares ai como perereca diante da cascavel? You not eat, you who were dying of hunger, and stay there like a tree frog before a rattlesnake? (Monteiro Lobato, Negrinha, 75) [Portuguese] (58) Tambm, irem a Queluz com um dia dstes! Ho de se divertir! Also, they go to Queluz on

    The syntactic distribution of the inflected infinitive 25

  • one of these days! They have to have fun! (Ea de Queirs, Os Maias, I, 328) [Portuguese] (59) Naneeecho, Naneeecho!, matares t unha muller, hom!,-berraba cada vez mais o demo do tonto. Nanecho, Nanecho!, you killed a woman, man!-yelled louder and louder the devil of a fool (Contos da Galiza, 89:197) [Galician]

    A sequence of conjoined infinitives is an additional predictable context for the inflected infinitive. The sequence of infinitives indicates a multiclausal structure wherein each infinitive is the final predicate of a separate clause. This environment would allow for the possible use of the inflected form (examples

    (60) Quem vos mandou agarrar-vos tenazmente, e espalhardes as razes e as mergulhardes no subsolo? Who ordered you to cling stubbornly, and to scatter your roots and to dive into the subsoil? (A.Arinos, Pelo Serto, 92) [Portuguese] (61) Tiabamos que compor cantigas grorificadoras da Pauta e sermos felices na nosa comunidade familiar. We had to compose songs glorifying Pauta and be happy in our familiar community (Elipsis, 59) [Galician] (62) Erano ancora in quella citate, zoy de sopre a le plaze, ordenate e facte multi e diversi portichy, sotto de li quale poteano largamente andare la gente per tiempo plovioso, e recostarenosse in quilli luochy covierti senza poterenosse bagnare n infondere dallacqua plovea. There were ordered and made in that city, that is above the plazas, porticos many and diverse, under which people could walk during rainy weather, and take shelter in those covered places without getting wet or flooding in the rain (79.1923) [Old Neapolitan]

    We should also expect the possible use of the inflected infinitive after verba dicendi, such as dizer, afirmar, proibir, rogar, pedir, etc. The monoclausal union construction is demonstrably not available to these verbs, 5 hence the

    5The verb mandar order is an exception; as we have seen, it may take a monoclausal union structure or a biclausal 2-control structure (see 2.2).

    occurrence of the inflected infinitive with these verbs is always possible because these are biclausal constructions in which the infinitive is the final predicate of its clause (examples from Maurer 1968:113, 184; Gondar 1978:111; Loporcaro 1986:196).

    (63) o instinto dissera-lhe serem esses homens iguais a si e aos seus camaradas. instinct had told him that those men were equal to him and to

    The inflected infinitive in romance languages 26

  • his camarades (Duarte, Vencidos, 80) [Portuguese] (64) Quo profundo e formidvel o mandamento que nos prescreve tomarmos cada qual aos ombros a nossa cruz. How profound and formidable is the commandment that prescribes us to each take on our shoulders our cross (Rui, Prece de Natal, 42) [Portuguese] (65) O que non se pode aceutar como ouxetivas e vlidas interpretacins que denotan claramenteseren productos ideoloxizados. What one cannot do is accept as objective and valid interpretations that express clearly to be ideologicalized products (Conflicto lingstico, 48) [Galician] (66) i poviri malati sensa agiuto o conseglio di medici i quali sensa alchuna carit domandano esserono paghate. the poor sick without help or advice of doctors who without any charity demand to be paid (Cronaca di Partenope, 11.135) [Old Neapolitan]

    When the infinitive functions as the subject of a sentence, the inflected form of the infinitive may occur. Such infinitives frequently occur after impersonal verbs or expressions, or as the argument of noun or adjective predicates such as provvel, pena, imprescindvel, etc. In this context, the infinitive clause is itself an argument of a superordinate predicate, so we know for certain that it forms a separate clause where the infinitive is the final predicate (example from Gondar 1978:93).

    (67) Dese xeito, non che importa saberes o que se fala? In that way, doesnt it matter to you to know what they are saying? (Teatro, 24) [Galician] (68) pena no podermos ir. Its a pity we cant go [Portuguese]

    Lastly, another acceptable environment for the use of the inflected infinitive is after a preposition. Similar to the above case, the infinitive clause is itself an argument of the preposition. We can conclude that in this case the infinitive is the final predicate of its own separate clause. This is a frequent context for the use of the inflected infinitive where it can occur as the complement of a noun, adjective, or adverb, in adverbial clauses, or equivalent to a gerund after a or sem (examples from Maurer 1968:114, 115; Jones 1992:297; Loporcaro 1986:203; Gondar 1978:127).

    (69) E os tropeiros, certos de estarem diante de um fato sobrenatural, falavam baixo. And the cattle drivers, certain of being in the presence of a supernatural occurrence, spoke softly (A. Arinos, Pelo Serto, 37) [Portuguese] (70) Acostumaram-se quilo de verem tudo poltica e de chamarem tudo poltica. They were

    The syntactic distribution of the inflected infinitive 27

  • used to that way of seeing everything as politics and of calling everything politics (Castilho, Felicidade, 7, 140) [Portuguese] (71) Su postinu est colatu prima de arrivaremus. The postman came by before we arrived [Sardinian] (72) se sforzanoper se potereno ben defendere da nuy. they strive to be able to defend themselves well from us (Libro de la destructione de Troya, 44.45) [Old Neapolitan] (73) Chegara o tempo de froitificaren as ideas espalladas pola Revolucin francesa. The time had arrived for the ideas popularized by the French Revolution to bear fruit (Vicente Risco, 72) [Galician]

    Contexts in which the inflected infinitive will always be prohibited are ones in which the infinitive is the inner predicate of a monoclausal construction. An obvious example of such context is the union construction available to modals and causative and perception verbs described in 2.2 above. The monoclausality of a sentence can be confirmed through clitic position and negation. An additional effect is that the inner 1 of a transitive infinitive would be dative in the union construction of causative and perception verbs, whereas it would be accusative in the biclausal 2-control construction. For instance, in (74) the subject of the inner predicate vender occurs with the final predicate mandei as the dative clitic lhes, indicating that the two verbs are clausemates. In this context, the infinitive cannot occur inflected because it is not the final predicate of its clause, as shown in (75).

    (74) Eu mandei-lhes vender os discos. I ordered them to sell the records (75) *Eu mandei-lhes venderem os discos. I ordered them to sell the records

    An additional environment where the inflected form will be excluded is when the infinitive is used in a general sense or with abstract meaning. Since in this case there is no subject for the infinitive to agree with, it cannot take inflection (examples from Maurer 1968:108).

    (76) E resta saber por fim se o estilo no uma disciplina do pensamento. And it remains to find out finally if style is not a discipline of thought (Ea de Queirs, Os Maias, I, 324) [Portuguese] (77) Nos ares andam sempre idias de tdas as idades sem falar nas que vo caindo mortas. In the air ideas from all ages always progress not to mention those which fall away dead (Castilho, Felicidade pela Agricultura, VII) [Portuguese] (78) O traballar proveitoso. Working is profitable [Galician]

    The inflected infinitive in romance languages 28

  • Given the condition in (21) on the use of the inflected infinitive and the discussion of other syntactic phenomena provided in this chapter, we can correctly predict the precise range of environments where the inflected infinitive will be acceptable and others where it will be excluded. This description will serve as a background for my critique of previous accounts of the distribution of the inflected infinitive in Portuguese and Galician provided in the next chapter and as a background for my discussion of the occurrence of the inflected infinitive in other Romance languages provided in chapter six.

    The syntactic distribution of the inflected infinitive 29

  • .

  • CHAPTER 3 Previous Accounts of the Distribution of the

    Inflected Infinitive

    Much has been written concerning the distribution of the inflected infinitive in Portuguese. Many scholars have attempted to predict the occurrence of the inflected infinitive by providing long lists of rules and norms and relevant examples. However, since these descriptions of the inflected infinitive did not attempt any serious structural analysis, rules proposed for its use ended up being inconsistent and imprecise. Using my syntactic analysis of the inflected infinitive from chapter two as a background, I will study the descriptions provided by previous linguists and grammarians for the distribution of the infinitive in Portuguese and Galician and demonstrate how these attempts have failed to capture the necessary condition for the inflected infinitive in these languages.

    3.1 Previous Accounts for Portuguese

    3.1.1 Soares Barbosa (1822)

    Soares Barbosa (1822:2834) was the first grammarian to propose rules for the use of the Portuguese inflected and uninflected infinitives. He states that the uninflected infinitive must be used:

    1. Whenever the subject of the governing verb is the same as the infinitive, as in eu quero fazer I want to do, ns queremos fazer we want to do, etc.

    2. Whenever one wishes to express the action of the infinitive in a general sense, without specifying a subject, as in mentir faltar verdade to lie is to be lacking in truth.

    As support for his first rule, Soares Barbosa claims that the use of an inflected infinitive in the following sentences from Cames is incorrect, because the subjects of the two verbs in each sentence are identical.

    (1) Netuno, lhe disse, no te espantes de Baco nos teus reinos receberes . Oh, Neptune, I said to him, dont be afraid of receiving Bacchus in your kingdoms (Cames, Lus. VI, 15) (2) E folgaras de veres a policia Portuguesa na paz, & na milicia. And you will be at ease by

  • seeing the Portuguese police in peace and in warfare (Lus., VII, 72)

    However, Said Ali (1966) suggests that the inflected infinitive is used in the first example to clarify the subject of the infinitive-that it is not Baco, and in the second to give emphasis to the subject that is performing the action. Stylistic considerations of clarity and emphasis, discussed in the previous section, contribute to the choice to use the inflected infinitive in (1) and (2).

    Others have disagreed with the first rule because they see fault in Soares Barbosas claim that there are two subjects, of the conjugated verb and of the infinitive, in examples like eu quero fazer (Said Ali 1966; Maurer 1968). Instead, they view the construction as an infinitive together with an auxiliary forming one clause with one subject, in other words a monoclausal structure. The pattern observed by Soares Barbosa in his first rule is entailed by my analysis. The occurrence of the uninflected infinitive together with certain verbs like modals and auxiliaries when the subjects of the verbs are the same follows from the observation presented in the previous chapter that modals like querer allow for both a monoclausal and biclausal structure in Portuguese. When the infinitive participates in the monoclausal structure, it must be uninflected because it is not the final predicate of its clause. Although the uninflected infinitive is more common with modal verbs, the inflected infinitive can and does occur in this context since modals also allow a biclausal structure in which the infinitive is a final predicate.

    The rules given by Soares Barbosa for the use of the uninflected infinitive are vague but give us insight into its use. As we have seen in (1) and (2), the inflected infinitive can occur when the subject of the main verb and the infinitive are the same, which contradicts Soares Barbosas first rule above. Therefore, the identity of subjects is not a deciding factor in the choice to use either infinitive. Instead, the pattern that motivates his assertion in the first rule can be stated more accurately by claiming that only the uninflected infinitive can occur in monoclausal structures where it is an inner (non-final) predicate. His second rule states the obvious: when the infinitive expresses an action with no specific subject, the uninflected infinitive must be used. The inflected infinitive is clearly irrelevant in this context, since there is nothing for the inflection to agree with.

    Soares Barbosa next presents the following rules for the purported obligatory use of the inflected infinitive:

    1. When the subject of the infinitive is different from the finite verb, or if there is uncertainty as to the subject of the infinitive, as in creio termos sido enganados I believe we have been deceived.

    2. When the infinitival clause, either as a subject or attribute of another verb, or as a complement of a preposition, is taken in a personal and not an abstract sense, as in louvares-me tu me causa novidade your praising me produces curiosity in me.

    Soares Barbosa claims that the purpose of every instance of the inflected infinitive is to get rid of any uncertainty that there could be over whether or not the subject is the same for both verbs. Here he incorporates a stylistic motive for its usethe need for claritywhich is important but does not always determine the use of the inflected form. For

    The inflected infinitive in romance languages 32

  • example, it does not explain why the inflected infinitive would co-occur with overt subject pronouns.

    The pattern observed by Soares Barbosa in his first rule (use the inflected form when the subject of the infinitive is different from that of the main verb) is true in rough outline. Both this general pattern and the particular departures from it are entailed by my analysis. If the infinitive has a subject different from that of the main verb, the infinitive generally belongs to a separate clause. This would suggest that the infinitive is the final predicate of its clause, which is the environment where the inflected infinitive occurs. In disagreement with this first rule, Said Ali (1966:64) points out that this rule would allow for sentences such as:

    (3) *No nos deixeis cairmos em tentao. Dont let us fall into temptation (4) *Fao-te estudares. I make you study

    Also, this rule does not allow for acceptable sentences such as: (5) Deixai vir a mim os pequeninos. Let the children come to me (6) Fazei-os sentar. I made them sit down

    which would seem to conflict with the first rule because the infinitive does have a different subject but remains uninflected. Therefore, identity or non-identity of subjects does not correctly predict the occurrence of the inflected and uninflected infinitives. A more unified and reliable diagnostic for the use of the inflected infinitive is its status as a final predicate, regardless of whether its subject is the same as that of the verb in a superordinate clause.

    Soares Barbosas second rule for the inflected infinitive is general and obvious: use the inflected infinitive when the action indicated by the infinitive is taken in a personal and not a general sense. In other words, the inflected infinitive is not used when the subject of the infinitive is unspecified. The condition proposed in my analysis provides a more unified and accurate account of the phenomena described in Soares Barbosas rules.

    3.1.2 Diez ([183644] 1973)

    Diez (1973) claims that the inflected infinitive is used only in cases where it is possible to substitute it with a finite verb, whether the infinitive has its own distinct subject or not. He also states that when the infinitive occurs with a modal, it can never be inflected.

    (7) Alegram-se por terem visto o pai. (= porque vieram o pai) They rejoice because they saw their father (8) Acreditando tu no me teres ofendido. (= que no me ofendeste) Believing that you didnt offend me (9) Afirmo terem chegado os navios. (= que chegaram os navios) I assert that the ships had arrived

    Previous accounts of the distribution of the inflected infinitive 33

  • Although imprecise, Diezs rule reveals a valuable linguistic observation. His statement that the inflected infinitive occurs only when it can be substituted by a conjugated verb reflects the fact that the infinitive has lost some of its indefinite value as a nominal by adopting personal endings and therefore approaching a more finite value. By equating the infinitival clause with a finite clause in this context, Diez hints at the fact that when the inflected infinitive is used, it occurs in its own clause and it is the final predicate of that clause. His rule, however, does not predict the occurrence of the inflected infinitive with modal verbs and in other constructions in which the infinitive cannot be substituted by a finite verb. For example, in (10) and (11) the inflected infinitives seres and dizerem cannot be substituted by a clause with a finite verb.

    (10) Possas tu, descendente maldicto de uma tribu de nobres guerreiros, implorando crueis forasteiros, seres presa de vis Aymors. May you, a cursed descendent of a tribe of noble warriors, begging cruel strangers, be the prey of the Aymors (G.Dias, Y Juca Pyrama) (11) Donde vem dizerem as molheres prenhes que s vezes est no ventre to quieta a criana. Whence the pregnant women come to say that sometimes the child is so still in the womb (Nunes, Arr. 569)

    In addition, Diezs assertion that the inflected infinitive never occurs after a modal is incorrect, as we can see in the above sentences, because it can occur optionally in the biclausal structure available to modal verbs described in 2.2. Diezs observations concerning the inflected infinitive are incomplete and do not correctly predict its actual usage in Portuguese.

    3.1.3 Said Ali (1966)

    Said Ali (1966:6366) suggests two cases where an infinitive may be used: (1) when the infinitive unites with a verb of incomplete meaning, forming with it a complete predicate, or (2) when the infinitive has or admits its own subject, mattering little whether or not it is the same as the subject of the other verb. According to him, the uninflected infinitive occurs in the first context, and the inflected infinitive in the second.

    Although it would be difficult to decide a priori which complement-taking verbs are verbs of incomplete meaning, what emerges from Said Alis discussion is that he means to delineate a class which includes modals (poder, saber, dever, haver de, ter de, querer) and causatives (fazer, mandar, deixar) which, according to him, always occur with an uninflected infinitive. There are other verbs and phrases such as estar a, costumar, comear a, continuar a, ousar, desejar, gostar de, vir that for him always take an uninflected infinitive. Since the uninflected infinitive shows no overt subject, he concludes that an infinitive without a subject is the same as an infinitive without inflection. Under my analysis we expect to find in broad outline patterns like those described by Said Ali. From the condition proposed in chapter two, we know that the inflected infinitive is prohibited in monoclausal constructions, because the infinitive in these contexts is an inner predicate and not the final predicate of its clause. The verbs

    The inflected infinitive in romance languages 34

  • listed by Said Ali above are generally interpreted by many scholars to have a monoclausal structure, and therefore would co-occur only with the uninflected infinitive, as he proposes. However, as we have seen in 2.2 and 3.1.2, modals, causatives, and other similar verb phrases in Portuguese allow for both a monoclausal and biclausal structure. When the infinitive participates in the monoclausal structure, it must be uninflected because it is not the final predicate of its clause, but in the biclausal structure available to these verbs, the inflected infinitive may occur because the infinitive is the final predicate of its clause. Therefore, a more accurate explanation would include the status of the infinitive as a final or inner predicate of its clause.

    Said Ali suggests that whenever an overt subject occurs with an infinitive, the infinitive must be inflected. Although this is a strong tendency in Portuguese, this use does not appear to be obligatory. For example in (12), the infinitive occurs with an overt subject (os meninos) but is not inflected (Thomas 1969:193).

    (12) Ouvi os meninos dizer que iam praia, I heard the boys say they were going to the beach

    Said Ali states that the use of the inflected infinitive is obligatory when the action expressed by the infinitive can refer to a person or something different than what we have in mind:

    (13) Deixas criar aas portas o inimigo por yres buscar outro de to longe. You let the enemy set up at the entrance because you went to search for another so far away (Cames, Lus., 4, 101)

    This need for clarity allows for the use of the inflected infinitive in conditions where one normally does not expect it. For example, he claims that it can appear with a modal or causative verb if there is some distance between the conjugated verb and the infinitive. As discussed previously in 2.2, the inflected infinitive can occur in the biclausal construction available to modal, causative, and perception verbs in Portuguese even if there is no distance between verbs.

    (14) Mas a selva j comea a rarear, e os ginetes a resfolegarem com mais violencia. But the woods already start to be scarce, and the horsemen to breathe with more violence (Herculano, Eur., 227) (15) Deviam-n-o trazer todos vocs nas palmas, dar mil graas aos ceos, e acabarem de crer. All of you should bring him in triumph, give many thanks to the heavens, and conclude believing (Castilho Antnio, Tart, 11)

    Said Ali also claims that the inflected infinitive can be used in any environment for emphasis according to the intention of the author. These examples given by the author are instances of an isolated exclamatory infinitive, which is certainly the final predicate of its own clause, and entailed by my analysis.

    (16) Morreres?!Oh, no? Salvaste Hermengarda do opprobrio. You die? Oh,

    Previous accounts of the distribution of the inflected infinitive 35

  • no? You saved Hermengarda from infamy (Herculano, Eur., 289) (17) Assassinares uma fraca mulher, assassinar-te a ti proprio e renegares da vida eterna. You kill a weak woman, you kill yourself and you renounce eternal life (Herc. M. de C. 1, 2)

    Said Ali (1966:7374) believes that we can reduce all uses of the two infinitives to the following five rules.

    Uninflected Infinitive: 1. Whenever the verb indicates action in general, as if it were an

    abstract noun, or when its meaning does not suggest any specific subject. 2. In compound and periphrastic verbs, unless there is a need for

    clarification, for example if the finite verb and infinitive are separated by some distance making the meaning of the clause unclear.

    His first rule is obvious, for the infinitive cannot take inflection if there is nothing for it to agree with, as discussed in 3.1.1. The pattern observed in Said Alis second rule is entailed by my analysis, because some compound and periphrastic verbs occur only in a monoclausal construction in which the infinitive is always a non-final predicate and would therefore never be inflected. But if these verbs occur in a biclausal construction, then the infinitive may be inflected because in this case it would be the final predicate of its own clause. In this second rule, Said Ali does not identify which verbs he considers to be compound verbs and periphrastic verbs; as we have discussed with his explanation of modal and causative verbs above, it is insufficient to list types of verbs with which the uninflected infinitive occurs, because with some verbs there are two structures available for their use: monoclausal and biclausal. The true determining factor, therefore, is the status of the infinitive as an inner or final predicate of its clause.

    Said Ali presents the following rules for the use of the inflected infinitive.

    Inflected Infinitive: 1. Whenever the infinitive is accompanied by an overt subject, noun, or

    pronoun, whether the same or different from the other verb. 2. Whenever it is necessary to underscore the agent, and refer the

    action especially to a subject, either to avoid confusion, or to clarify the thought.

    3. When the author intentionally emphasizes the person to whom the action refers.

    In other words, what determines the use of the inflected infinitive in Said Alis view is: the presence of an overt subject, and therefore simple obligatory agreement; the necessary emphasis of the subject to facilitate comprehension; and the intentional emphasis to underscore the person involved in the action. All of these are elements outlined in my analysis of the inflected infinitive-stylistic concerns such as clarity and emphasis, and the presence of an overt subject with the infinitive. The latter pattern

    The inflected infinitive in romance languages 36

  • entails a biclausal structure (see 2.4), because the presence of an overt subject with the infinitive signals a separate clause for the infinitive in which it is the final predicate and can be inflected. Said Alis second and third rules fit with my analysis; however, there are instances where the inflected infinitive is excluded independently of emphasis or clarity, namely when the infinitive is not the final predicate of its clause. Therefore, the stylistic considerations of emphasis and clarity can only be brought to bear when the syntactic condition for the use of the inflected infinitive is met.

    Said Ali points out other uses of the inflected infinitive, such as in the construction ao+infinitive, which is equivalent to the temporal clause introduced by the conjunction quando. This observation follows from the general use of the inflected infinitive in prepositional phrases (see 2.4). When it occurs after prepositions like ao, the infinitive clearly belongs to a separate embedded clause and therefore can occur inflected. The presence of the reflexive clitic se attached to the infinitive in this example also indicates that it is the final predicate of its own clause (see 2.2).

    (18) Ao aproximarem-se , os dous exercitos de nuvens prolongaramse. On their approaching, the two arrays of clouds extended themselves (Herc. Eur. 53)

    Said Ali also gives examples of an inflected infinitive used with verbs of declaring, showing, and thinking. The inflected infinitive can occur in these contexts because they are biclausal constructions and the infinitive is the final predicate of its clause. The monoclausal union construction is not available to verba dicendi in Portuguese 1 and therefore the possibility of an inflected infinitive in this context is predictable. In addition, this is confirmed if we look at auxiliation in (19) and (20). Auxiliation only occurs in the final P-sector(s) of a clause, and the presence of the auxiliary serem in these examples indicates that it is the final predicate of its own clause and explains the occurrence of the inflected form.

    (19) Para que indinasse a el rey contra elles com algumas razes apparentes, que lhe deram para o caso, affirmando serem verdadeiras. So that he provoked the king against them with some feigned motives, that they had given him for the case, asserting them to be true (Barros, Dec. 1, 4, 9)

    1An exception is mandar command, which accepts the monoclausal construction, as we saw earlier (2.3).

    (20) Dizendo serem aquellas cousas engano. Saying that those things are a mistake (Barros, Dec. 1, 3, 10) (21) E estando elle e os outros contentes, cuidando terem bom posto foram de noite todos chamados. And he and the others being happy, imagining that they had a good position, all were summoned at night (Barros, Dec. 3, 2, 9) (22) Estiveram mui promptos, mostrando terem

    Previous accounts of the distribution of the inflected infinitive 37

  • contentamento na paciencia. They were very ready, showing that they had satisfaction in patience (Barros, Dec. 1, 5, 2)

    Said Ali states that the use of the inflected or uninflected infinitive after a preposition seems arbitrary and depends on the intention of the author and not on a fixed grammatical rule (1923:148). In my analysis, the occurrence of the inflected infinitive after a preposition is always possible, because the prepositional phrase consists of a separate clause in which the infinitive is the final predicate. Its use in these contexts may be determined instead by other factors, such as stylistic considerations or the presence of an overt subject.

    (23) Parece que havia de dizer: os Anjos ouvem a palavra de Deos para a fazerem e nam, os Anjos fazem a palavra de Deos para a ouvirem Pois porque diz que fazem para ouvir , e no ouvem para fazer? It seems that it was necessary to say: the angels hear the word of God in order to do it and not, the angels do the word of God in order to hear it So why say that they do to hear and not they hear to do? (Vieira, Serm. 3, 53) (24) Guardaram a palavra de Christo antes de a ouvir . Executaram a palavra de Christo antes de a ouvirem . They defended the word of Christ before hearing it. They enforced the word of Christ before hearing it (Vieira, Serm. 3, 534)

    Although the inflected infinitive is frequently used in exclamative and interrogative clauses for emphasis (see 1617 above), Said Ali (1923:138) provides examples of the uninflected infinitive used along with an overt subject in similar contexts. The presence of an overt subject with the infinitive normally triggers the use of the inflected form of the infinitive, however these examples taken from Old Portuguese show that the distribution of the inflected and uninflected forms of the infinitive was more fluid at that time. That is, the occurrence of an uninflected infinitive after an overt subject in the following examples perhaps indicates that the use of the inflected form was not as widespread in older stages of the language.

    (25) Alcarac, n poso creer taaes cousas, como me dizes, ca som contra natura; quatro mil cavaleiros mteer lide a tantos e t boos como os meus er! Alcarac, I cannot believe such things, as you tell me, because they are against nature; four