Tener +Past Participle Almost any synchronic study of a language uncovers variation, uncertainty and apparent inconsist...
163 downloads
332 Views
1MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Tener +Past Participle Almost any synchronic study of a language uncovers variation, uncertainty and apparent inconsistency. The introduction of an historical perspective to linguistic description can frequently add to our understanding of these phenomena. The so-called periphrastic constructions found in Spanish, and the problems which they pose for linguistic description and classification, provide a useful vehicle for exploring the possibilities for combining synchrony and diachrony. The book centres on an examination of one such construction: tener+past participle. A detailed study of this often neglected construction is justified in its own right, but in addition, because it is of uncertain grammatical status, it encapsulates many of the problems encountered by the synchronic linguist. The considerable variation in usage between speakers, the availability of historical data, and the existence of cognates in other Romance languages all provide useful comparative information which can be brought to bear on the study of modern usage. The work focuses on a study of tener+past participle as used in the modern language and an historical survey of its evolution. The interplay between the two is set against the background of various existing theories concerning auxiliary and periphrastic verbs, providing detailed data against which such theories can be tested. It also places the development of tener+past participle within the wider context of similar developments in other Romance languages. While the study is primarily aimed at linguists specialising in the structure of the Spanish language, it should also provide useful comparative material for those with an interest in theories of linguistic description covering Romance languages as a whole. Catherine Harre studied at Queens’ College, Cambridge where she also completed her Ph.D. She is now a civil servant working for the Department of Trade and Industry. Other books in the Romance Linguistics series include: Structures and Transformations Christopher J.Pountain Studies in the Romance Verb eds. Nigel Vincent and Martin Harris Weakening Processes in the History of Spanish Consonants Raymond Harris-Northall Spanish Word Formation M.F.Lang Tense and Text Dulcie Engel
Variation and Change in French John Green and Wendy Ayres-Bennett Latin Syntax and Semantics Harm Pinkster Thematic Theory in Syntax Robin Clark Tense and Narrativity Suzanne Fleischman Comparative Constructions in Spanish and French Syntax Susan Price Latin and the Romance Languages in the Early Middle Ages Ed. Roger Wright Also of interest: The Romance Languages Martin Harris and Nigel Vincent
Tener+ Past Participle A case study in linguistic description
Catherine E.Harre
London and New York
First published 1991 by Routledge 11 New Fetter Lane, London EC4P 4EE This edition published in the Taylor & Francis e-Library, 2005. “To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to http://www.ebookstore.tandf.co.uk/.” Simultaneously published in the USA and Canada by Routledge a division of Routledge, Chapman and Hall, Inc. 29 West 35th Street, New York, NY 10001 © 1991 Catherine E.Harre 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. British Library Cataloguing in Publication Data Harre, Catherine E. Tener+past participle: a case-study in linguistic description. 1. Spanish language. Linguistics I. Title 460 Library of Congress Cataloging in Publication Data Harre, Catherine E. Tener+past participle: a case-study in linguistic description/Catherine E. Harre. p.cm.—(Romance linguistics) Includes bibliographical references and index. 1. Romance languages—Verb. 2. Romance languages— Syntax. 3. Romance languages—Participle. I. Title. II. Title: Tener plus past participle. III. Series. PC145.H3 1991 440–dc20 ISBN 0-203-99360-8 Master e-book ISBN
ISBN 0-415-05647-0 (Print Edition)
Contents Texts and abbreviations
vi
Introduction
1
1 The question of auxiliary status
10
2 Tener +past participle in modern Castilian
39
3 Historical development of tener+past participle
78
4 Historical development in Portuguese
106
5 Comparison with other Romance languages
127
6 Conclusions
148
Notes
157
References
165
Index
170
Texts and abbreviations In the following lists the abbreviation by which each text is referred to is followed by details of the edition used. Then the date (or approximate date) of the original text is given, and the means by which examples from the texts are referenced is indicated. Spanish Mil
Gonzalo de Berceo, Los milagros de Nuestra Señora, ed. B. Dutton, 1971, London. Early/mid thirteenth century (referenced by verse and line).
PCG
Primera crónica general, ed. R.Menéndez Pidal, 1955, Madrid. Second half of thirteenth century (referenced by page, line, and column).
LBA
Juan Ruiz, Arcipreste de Hita, Libro de buen amor (Salamanca manuscript), ed. M.Criado de Val and E.W.Naylor, 1965, Madrid. 1343 (referenced by verse and line).
Cor
Alfonso Martínez de Toledo, Arcipreste de Talavera o Corbacho, ed. J.González Muela, 1970, Madrid. 1438 (referenced by page).
Cel
Fernando de Rojas, La celestina, ed. D.S.Severin, 1976, Madrid. Post-1499 (referenced by page).
Ter
Santa Teresa de Avila, Epistolario, in Obras completas, ed. E. de la Madre de Dios and O.Steggink, 1962, Madrid. 1546–74 (referenced by letter and paragraph).
Qui
M. de Cervantes Saavedra, El ingenioso hidalgo Don Quijote de la Mancha, Part 2, ed. L.A.Murillo, 1978, Madrid. 1615 (referenced by page).
Fei
B.J.Feijóo y Montenegro, Cartas eruditas, ed. A.Millares Carlo, 1928, Madrid. First half of eighteenth century (referenced by page and line).
Jov
G.M.de Jovellanos, Obras I: Epistolario, ed. J.Caso González, 1970, Barcelona. 1778–91 (referenced by letter and page).
STP
P.A.de Alarcón, El sombrero de tres picos, ed. V.Gaos, 1975, Madrid. 1874 (referenced by page and line).
CHM
M.Delibes, Cinco horas con Mario, 5th ed, 1971, Barcelona. 1966 (referenced by page).
Portuguese Dem
A demanda do Santa Graal, ed. A.Magne, 1944, Rio de Janeiro. Late thirteenth/early fourteenth century (referenced by page).
CDP
Crônica de D.Pedro. First half of fifteenth century.
DSG
Diálogos de São Gregório. 1=Fourteenth-century version; 2= Fifteenth century version.
LRR
Lenda do Rei Rodrigo. Copy from 1410–20 of an early fourteenth-century original and a 1344 copy.
ODE
Orto do esposo, ed. B.Maler, 1956, Rio de Janeiro. Late fourteenth/early fifteenth century (referenced by page and line).
RDJ
Fernão Lopes, Crónica del Rei Dom Joham I, Part 2, ed. W.J. Entwistle, 1968, Lisbon. C. 1440 (referenced by page).
PJ
F.Alvares, Verdadeira informação das terras do Preste João da Indias, ed. A.Reis Machado, 1943, Lisbon. 1540 (referenced by page).
Vie
A.Vieira, Cartas, ed. J.Lúcio d’Azevedo, 1925, Coimbra. 1626–47 (referenced by page).
VME
L.A.Verney, Verdadeiro método de estudar, ed. A.Salgado, 1949, Lisbon. 1746 (referenced by page and line).
CPA
Eça de Queiroz, O crime do Padre Amaro (=Volume 1: Edição do Centenário: Obras de Eça de Quieroz), 1946, Oporto. 1871 (referenced by page).
Vin
M.Torga, Vindima, 3rd ed, 1965, Coimbra. 1945 (referenced by page).
Introduction 0.1 AIMS OF STUDY For most of this century linguistic interest has centred on synchronic rather than diachronic descriptions of language. Linguists have concentrated on providing an accurate account of a language at a given period, without reference to historical considerations. If one’s aim is to formalise the subconscious knowledge which a speaker has of his or her own language and to provide an accurate description of how the language is used, then this is a sensible approach. Most people’s knowledge of their own language is unaffected by historical considerations, and even where speakers do have some knowledge of the history of their language it does not greatly affect the way in which they speak. In a description of the language people use, historical considerations are irrelevant, and their inclusion may be misleading. However, these purely synchronic approaches have failed to account convincingly for some important aspects of language. They have stressed the need to define word classes and to categorise items according to these classes, but they have not come to terms adequately with the fact that it is not always possible to make clear divisions between classes. They have also tended to idealise the situation by ignoring variation—whether in the speech of an individual speaker or between speakers—and borderline cases where speakers are uncertain as to whether an example is or is not part of the language. They have concentrated on the clear-cut cases, leaving the ‘awkward’ cases on one side. Such descriptions ignore an important aspect of languages: the fact that they are constantly changing. The introduction of a diachronic perspective to linguistic description can help to solve problems which prove so difficult for synchronic descriptions. This is a two-way process: the fact that languages change can help to account for variation, uncertainty, and apparent inconsistency; and by looking at such cases, we can often gain useful information about the way in which the language is changing, and about the basic principles of language change. In diachronic linguistics the ‘awkward’ cases shunned by the synchronic linguists are frequently the most useful. They provide clues to past states of the language and indicate the possible course of future development. The Spanish verbal system provides a very appropriate framework in which to study the possibilities for combining synchrony and diachrony. There are a number of periphrastic forms which are on the edge of the auxiliary verb system, having undergone varying degrees of grammaticalisation. These forms have also to varying extents displaced or partially taken over from some simple forms. The ‘system’ is thus not nearly so neat and clearly definable as some writers would have us believe. This study centres on an examination of one of these forms: tener+ past participle. It looks at its use in modern Spanish, traces its historical evolution, and compares it with other reflexes of Latin TENERE+past participle. Tener+past participle has been chosen for two main reasons: firstly, because it is of uncertain grammatical status it encapsulates many of the problems faced by synchronic descriptions; secondly, because some
Tener+Past participle
2
developments undergone by tener+past participle are attested to different extents in the speech of different speakers, it is possible, by examining cross-speaker variation in usage, to find evidence of a large part of the construction’s historical development. There are other good reasons for selecting this construction. There has been no detailed survey of the way in which tener+past participle is used in modern Spanish. Most grammars of the language acknowledge its existence but allocate little space to what is considered to be a marginal construction. Symptomatic of this lack of attention is the fact that probably the best description of modern usage is to be found in the two pages devoted to it in a study designed for foreign learners of Spanish by Fente et al. (1972). 1 While their work aims to list and exemplify the different uses and meanings of the construction it does not deal in any detail with the precise limits on its use. For example, although tener+past participle must usually have a direct object, in some circumstances tener may appear with the past participle of an intransitive verb; but while Fente et al. provide an example with an intransitive verb they give no explanation of the limits of this use. They also indicate that while the participle usually agrees with the direct object this is not always so, without indicating what factors affect agreement. In short, they provide various examples of tener+past participle but provide no criteria which could be used to evaluate examples other than those they list. The historical development of tener+past participle has fared rather better, with surveys by Seifert (1930) and Roca Pons (1958). However, in the light of renewed interest in historical syntax and in particular in the evolution of possible auxiliaries, a fresh look at the history of tener+ past participle is in order. While a study of tener+past participle is of interest in its own right, it has consequences which reach beyond the confines of pure description. It can contribute to discussion of the problems encountered in attempts to define auxiliary verbs. These problems have occupied linguists for many years. Much of the work done in this area has been based on studies of auxiliary verbs in English, which have then been applied to Spanish (see, for example, Klein (1968; 1982); Fontanella de Weinberg (1970); Launay (1980)). More recently, some people have moved away from attempts to fit other languages into theories designed for English, and have looked for more general approaches which can accommodate the variation between languages (see the studies in Harris and Ramat 1987). While some of the work on English auxiliaries is applicable to Spanish, Spanish presents different problems. To take an obvious example, the syntactic criteria which can be applied to English auxiliaries (negation, question formation, ability to take stress etc.—see section 1.5) are of no relevance to Spanish. Much of the work which has been done specifically on Spanish auxiliary constructions has concentrated on constructions involving the infinitive and the gerund. Participial constructions have frequently been ignored or deliberately excluded on the grounds that they present additional problems (for example, the distinction between participle and adjective is not always clear, and word order may be more free). Where participial constructions have been included attention has tended to focus on the auxiliaries haber, ser, and estar; other more marginal auxiliaries have received less attention. 2 This study starts from an examination of one particular participial construction, and then evaluates various theoretical approaches in the light of that examination. It aims to bring detailed information to bear on the more general problem of auxiliary status.
Introduction
3
This study also considers variation among speakers. It demonstrates that even within a restricted sociogeographic group, language is not so homogeneous as many descriptions implicitly assume. By comparing modern Spanish usage with the historical development of tener+past participle, it considers how knowledge of synchronic variation can help us in our examination of diachronic change. As mentioned above, the study includes a cross-linguistic comparison of reflexes of TENERE. The obvious target for such a comparison is Portuguese, where ter is the standard perfect auxiliary, and consequently has received more attention than its Spanish cognate. This study compares the history of the constructions in the two languages, and also examines reflexes of TENERE+past participle in other Romance languages.
0.2 OUTLINE OF WORK The first chapter considers the question of auxiliary status and describes various attempts to define auxiliary verbs. The different criteria are then applied to tener, illustrating the shortcomings of some approaches and, in particular, their underlying assumption that it is possible or desirable to draw up discrete classes of verbs. Chapter 2 provides a detailed description of tener+past participle as used by a number of native-speaker informants. Some speakers consistently accept a wider range of uses of the construction than others do. Using the differences in tolerance among the speakers, various uses of tener+past participle are ordered on a scale, according to the extent to which they are accepted by informants. The results from this survey support the conclusions reached in chapter 1, that hard and fast boundaries between categories are inappropriate: individual examples of the tener+past participle construction may ‘pass’ some tests for auxiliary status and ‘fail’ others; different examples accepted by one speaker may fall into different categories; different speakers’ judgements on the same examples may be broadly similar but differ enough to produce different classifications. These conclusions lead to a discussion of Lehmann’s (1985) framework of graduated continua, which allow a construction to be classified as being relatively more or less grammaticalised in comparison with other similar constructions, and allow for it to exhibit different degrees of grammaticalisation according to different parameters, rather than attempting to assign it to a particular category. These parameters are then used to assess the current position of tener+past participle. The historical survey which forms the basis of chapter 3 is described in the light of the conclusions reached in chapter 2. The same parameters are used to measure the diachronic grammaticalisation of tener+past participle, and the ordered scale of acceptance of tener+past participle established in chapter 2 is compared with the historical evolution of the construction. Chapter 4 is a historical survey of the development of ter+past participle in Portuguese. By comparing Spanish and Portuguese we can discover how much of the evolution from verb of possession to perfect auxiliary they exhibit in common, and can come to some general conclusions about the evolution of the perfect. A comparison of Spanish and Portuguese helps to answer the questions of how and when tener+ past participle and ter+past participle began to differ in use.
Tener+Past participle
4
In chapter 5 we turn our attention to other Romance languages. Use of ter/tener+past participle in Galician and Asturian is examined and the evolution of TENERE+past participle in Catalan and Italian is considered briefly—from secondary sources—as a preliminary to a comparison of the development of reflexes of TENERE+past participle in the Romance languages. The relative grammaticalisation of the construction in the various languages is considered, and the question is asked to what extent they form part of one general trend and to what extent they have to be considered separately. The information about TENERE+past participle in these various languages is then used in a comparison with reflexes of STARE+past participle. Various people have pointed to similarities between these two resultant state constructions. This chapter looks at various parallels between them and also considers the possible interplay between them during their evolution. In the final chapter the various themes are summarised and drawn together. The extent to which synchronic variation can help us in our study of historical change is discussed in the light of the information presented in previous chapters. The information is also used in a discussion of the nature of syntactic change, based on Bennett’s (1981) article on the gradual nature of syntactic change. Finally, the study outlines areas in which work remains to be done, and possible directions for future research.
0.3 TERMINOLOGY Before embarking on the study described above, it is necessary to define some of the terms as they will be used here. 0.3.1 Aspect Aspect has been the subject of much discussion and much disagreement amongst Romance linguists. It is a question which has been approached in various different ways by members of widely differing schools of thought. A multitude of terms and categories have been suggested, some of which are mutually incompatible. It is outside the scope of this work to attempt to review all the major approaches to aspect and the terminology used in them. It is also not necessary for the purposes of this study to define aspect exhaustively: a working definition is all that is needed. I shall therefore outline the way in which aspectual distinctions are used in this work, and describe the terms used, while accepting that there will be those who would disagree with the approach or the terminology (or both). I shall take as my starting point the definition provided by Comrie (1976, 3): ‘aspects are different ways of viewing the internal temporal constituency of a situation’. 3 Comrie (p. 5) describes the difference between aspect and tense as a difference between ‘situation-internal’ and ‘situation-external’ time. 0.3.1.1 Morphological and lexical aspect One of the major causes of confusion and disagreement in the study of aspect has been the question of the distinction between morphological aspect and lexical aspect. These are frequently referred to by the German names Aspekt and Aktionsart respectively,
Introduction
5
although other terms are also used. 4 Although most linguists would accept a broad distinction between morphological and lexical aspect, there are considerable differences of opinion over how this distinction should be made. Coseriu (1980, 18–19) points out that two distinct sets of criteria—both based on Slav languages—have been used to distinguish between Aspekt and Aktionsart: the first set are based on whether a particular way of viewing the verbal action is expressed lexically or by grammatical means; the second set are based on the ‘objective’ or ‘subjective’ way in which the action is viewed. Thus a cet égard on distingue pour les langues slaves…aspect, concernant la façon d’envisager l’action verbale et appartenant a la grammaire, et ‘modes’ ou ‘espèces’ d’action (Aktionsarten), concernant la façon ‘objective’ dont l’action verbale se déroule ou se réalise et appartenant au léxique (par ex. ‘effectif’, ‘perduratif’, ‘cursif’, ‘terminatif’, ‘inchoatif’ etc.). Coseriu suggests that while these two criteria may provide the same results in Slav languages, this is not necessarily the case with other languages. One has to choose one set of criteria or the other, and accept that to use the other criteria would result in a different classification of the means which a particular language uses to express aspectual values. Lyons (1977, 706) also points out various difficulties associated with the term Aktionsart, and suggests the term aspectual character, which he describes as follows: ‘The aspectual character of a verb…will be that part of its meaning whereby it (normally) denotes one kind of situation rather than another.’ He illustrates this with the verbs ‘know’ and ‘recognise’: ‘know’ usually denotes a state, while ‘recognise’ usually denotes an event. Lyons goes on to state that ‘aspect and character are interdependent…because they both rest ultimately upon the same ontological distinctions’. In this discussion of the Spanish verb, a distinction will be drawn between the morphological means which the language has at its disposal to indicate aspect, and the inherent aspectual value (or character) of a verb stem. The first of these categories is based on verbal inflections and auxiliary verbs, which can be applied in a regular way to most verb stems in order to make certain aspectual distinctions. For example, Spanish regularly makes a distinction in the past between imperfective aspect (e.g. cantaba) and perfective aspect (canté). This distinction can be applied to all verbs regardless of the inherent aspectual value of that verb. Two points should be noted. Firstly, although in principle a particular morphological alternation applies to all verbs and is a regular productive device in the language, restrictions may apply to some types of verbs. For example, in general stative verbs such as saber, conocer, pertenecer, do not appear in the progressive forms—e.g. estar+present participle (see Lyons (1968, 315–16); Comrie (1976, 34–9)). The second point concerns the relation between the aspectual value of a verb stem and morphologically indicated aspectual distinctions. Even where the inherent lexical value of a verb does not prevent the use of a particular verb form it can still significantly affect the overall value ascribed to a construction. For the purposes of this study it is not necessary to go into this point in detail or to consider the various different lexical aspects which have been suggested. (For a discussion of these points and an illustration of how morphological and lexical aspect may combine to produce varying
Tener+Past participle
6
aspectual interpretations overall, see Pountain (1983, 50–9).) However, the fundamental notion of the interdependence of morphological aspect and the inherent character of a verb is important when considering the different interpretations of tener+past participle, as will be seen in chapter 2. 0.3.1.2 Perfective and imperfective aspect Use of these terms here broadly follows the description used by Comrie. For Comrie (1976, 21) ‘perfectivity involves lack of explicit reference to the internal temporal constituency of a situation’. In contrast, imperfectivity involves ‘explicit reference to the internal temporal structure of a situation, viewing a situation from within’ (p. 24). 0.3.1.3 Durative, iterative, and punctual aspect The term ‘durative’ is used to indicate that an event is viewed as continuing over a period of time. It is thus the opposite of punctual, which indicates that an event is viewed as having no duration. (Objectively, any event must have duration: the important fact here is that punctual aspect ignores this objective duration.) Iterative aspect indicates repeated action. 5 0.3.1.4 Resultant state As the term suggests, this indicates a state which is the result of a previous action. This notion is particularly important in this study, as one of the principal uses of tener+past participle is to express resultant state. For example, a sentence such as Tenemos resuelto el problema includes both the current situation and the causative action: as a result of our having thought out the problem, that problem is now resolved. In the way the term is used in this study, although the emphasis is on the state, the prior causative action is always also implied. The relative emphasis on state and action may vary from example to example, but both must always be present. 0.3.1.5 The perfect The term ‘perfect’ has been used in a variety of different ways, causing confusion and making it difficult to compare different approaches to the same problem or to compare similar categories in different languages. For some writers the perfect indicates that there is a link between a previous action and a current state, but in this study the term ‘resultant state’ is used to describe such situations (see section 0.3.1.4). However, most writers who follow a ‘current relevance’ approach apply a broader definition in which the perfect indicates ‘the continuing present relevance of a past situation’ (Comrie 1976, 52). Comrie divides uses of the perfect into four categories: perfect of result, where ‘a present state is referred to as being the result of some past situation’; experiential perfect, which ‘indicates that a given situation has held at least once during some time in the past leading up to the present’; perfect of persistent situation which describes ‘a situation that started in the past but continues…into the
Introduction
7
present’; perfect of recent past ‘where the present relevance of the past situation referred to is simply one of temporal closeness’. Another approach is to view the perfect as indicating that the event is regarded as taking place within the ‘extended present’. The extended present (also known as the ‘extended now’) is a period of time stretching back into the past from the present moment and including the present moment. The situation is further complicated by the fact that many writers use the term ‘perfect’ to refer to a particular verb paradigm: so have+past participle is the English perfect, and haber+past participle is the Spanish perfect. If one’s definition of the perfect is such that the category is filled by one particular verb paradigm, then that paradigm is effectively the perfect in that particular language. But there are, of course, differences between ‘the perfect’ as a linguistic category and ‘the perfect’ as a paradigm—even one that embodies the category. For example, if one views it purely as a category, then it is unchanging throughout time and across languages, although the paradigm which expresses it may change, and although different languages (or different stages of the same language) may realise different tense-aspect combinations. If, on the other hand, the perfect is closely linked to a particular paradigm, then it may change during the evolution of the language, and may vary from language to language. In practice, writers tend to combine these two approaches. Their view of what the perfect is as a linguistic category is influenced by the way in which exponents of that category are used in the language under consideration. This is sensible in that, obviously, one wants categories which are actually relevant in the description of a language. But writers are not restricted to viewing one particular paradigm as expressing the perfect. If that paradigm changes considerably in use, then clearly it no longer makes sense to classify it as a perfect paradigm if one wants to maintain a link between the term ‘perfect’ and a particular category. If one does not want to retain such a link, then the perfect becomes little more than a label—a name for a particular paradigm—and as such has very little practical value. 6 The various approaches can be illustrated by looking at different descriptions of Spanish present tense of haber+past participle (e.g. ha hecho). One common description of ha hecho is that it refers to actions in the past which have ‘current relevance’; it thus expresses a relation between a past action and the present moment. Among the exponents of ‘present relevance’ we find Bello and Cuervo (1898, 164–5) who state that: Comparando estas dos proposiciones: ‘Roma se hizo señora del mundo’ y ‘La Inglaterra se ha hecho señora del mar’, se perciben con claridad lo que distingue al pretérito del ante-presente [ha hecho]. En la segunda se indica que aun dura el señorío del mar; en la primera el señorío del mundo se representa como una cosa que ya pasó. La forma compuesta tiene pues relación con algo que todavía existe. …Se dice que una persona ha muerto cuando aun tenemos delante vestigios recientes de la existencia difunta; cuando aquellos a quienes hablamos están creyendo que esa persona vive; en una palabra, siempre que va envuelta en el verbo alguna relación a lo presente.
Tener+Past participle
8
Gili Gaya (1955, 141) suggests that ‘En español moderno significa la acción pasada y perfecta que guarda relación con el momento presente. Esta relación puede ser real, o simplemente pensada o percebida por el que habla.’ Alcina Franch and Blecua (1975, 802) indicate that the present tense of haber+past participle can express various relations between past event and present moment, but one of the main relations they list is that of present relevance: ‘Unas veces se trata del propio significado del verbo cuya acción llega con sus consecuencias hasta el presente; otras, intenciones afectivas del hablante que aproximan la acción remota; otras, la acción pasada habitual.’ Others have rejected ‘present relevance’ as an adequate criterion for use of ha hecho, and have favoured the notion of the extended present (see especially Alarcos Llorach (1947)). Rallides (1971, 29) argues that while this element [present relevance] very often is present, we do not feel convinced that it is justifiable to regard it as the fundamental distinction. We think that what happens subjectively is that when stating an event in the recent past or in the past where the larger context of time is not enclosed within a clear historical period…the speaker does not wish to enclose the event within the past even though he may state the definite occasion of the event. He thus neutralizes the closed nature of the event by using the ‘open context’ morpheme of the verb. However, there are problems with the extended present criterion. Most speakers of peninsular Spanish accept sentences such as (1) Lo he visto hace poco It is by no means clear that hace poco indicates a period of time which includes the present moment. This sort of example might be explained as a peculiarity of the hacer construction. However, Comrie (1985,106–7) claims that some speakers of Spanish accept both of the following sentences: (2) He dicho que Juan se vaya mañana (3) He dicho ayer que Juan se vaya mañana While (2) (the much more widely accepted of the two sentences) can be accounted for by Alarcos’ criterion, (3) cannot. For many speakers, the inclusion of the adverb ayer, which specifically prevents the action from being placed in the extended present, renders the sentence ungrammatical. But for those speakers who accept (3), the ‘extended present’ is not an adequate criterion for use of the ha hecho construction. In an earlier work Comrie (1976, 54) quotes the following example from Stevenson (1970, 62): (4) Gustavo Ferrán ha muerto ayer…se ha estrellado anoche en los montes de nieve 7
Comrie suggests (1985, 85) that the ha hecho construction has two meanings: current relevance and recent past (which frequently coincide, as recent events may easily have current relevance). Examples (3) and (4) would be covered by the criterion of current relevance, while example
Introduction
9
(5)below is covered by the criterion of recent past, even though the act of opening the window no longer has current relevance. (5) Hoy he abierto la ventana a las seis y la he cerrado a las siete For Fleischman (1983,194) the present perfect in English and in Romance languages ‘is used to refer to a situation that began or first occurred at an earlier moment and is still going on, or a situation whose reference period satisfies this criterion (e.g. today, in the past ten years …) or a completed past situation regarded as still relevant at the present moment’. Such a definition subsumes the criteria of Alarcos and Comrie. It can be seen that there are many partially overlapping and partially conflicting views on the ha hecho construction and on the perfect in general. In this study the term ‘perfect’ is used as a cover term which includes both ‘current relevance’ and ‘extended present’ approaches. Where it is necessary to define it more precisely in particular instances this will be done by explicit description. This somewhat vague use is made necessary by the fact that this study involves the comparison of two constructions in Spanish (tener+past participle and haber+past participle) and constructions in other languages (e.g. Portuguese ter+ past participle) which can all be viewed as forms of ‘the perfect’, but which, while displaying various similarities, are not exact equivalents. Applying a broad definition of ‘the perfect’ enables us to group these constructions together and to point to similarities in use. Where appropriate, differences in use will be highlighted by more specific classification.
1 The question of auxiliary status 1.1 CRITERIA The question of what constitutes an auxiliary verb and how to test for auxiliary status is a problem which has been addressed by grammarians of various different languages. This first section examines the principal criteria which have been suggested. Although it concentrates on solutions proposed for the Spanish language, it considers work done on other languages—especially French and English—where this can shed light on the situation in Spanish. First, some traditional views of the problem are considered, and then some of the approaches adopted in more recent years are examined. In addition to a general appraisal of these approaches, the study considers in more detail how the suggested criteria apply to tener+past participle. Some of the traditional reference works do not overtly state the criteria they use to decide what does or does not constitute an auxiliary verb. For example, Bello and Cuervo (1898) have no specific definition of auxiliary status, but merely list auxiliary constructions. They state that haber and estar are auxiliaries, but give no indication as to why these verbs are deemed to be auxiliaries, or why other verbs, such as ser, are not included in their list. Most writers do make some indication of what their classificatory criteria are, although the extent to which they adhere to these criteria and the consistency with which they apply them vary quite considerably. In his work on French auxiliaries, Schogt (1968; 1970) divides the criteria generally used for identifying auxiliaries into three areas: semantic, functional, and formal. This division can also be applied to descriptions of Spanish. However, it must be noted that most linguists make use of more than one type of criterion in identifying auxiliary verbs, and also that to some extent the three areas overlap. 1.1.1 Semantic criteria Semantic criteria, based on the idea of loss of lexical meaning, provide the means most commonly used in traditional accounts of the Spanish verb system for defining and classifying auxiliary verbs and periphrastic constructions. Broadly speaking, where a verb has its full lexical significance it cannot be considered an auxiliary; where this lexical meaning is lost or significantly reduced, the verb is treated as an auxiliary. Thus Gili Gaya (1955, 97–8) affirms that Para distinguir si un verbo está empleado como auxiliar basta fijarse en si ha perdido su significado propio. Cuando decimos voy a contestar a su
The question of auxiliary status
11
carta, el verbo ir es auxiliar, puesto que no conserva su acepción de movimiento de un lugar a otro. …el sentido habrá de decidir, en cada oración en que aparezcan tales perífrasis si su significación se ha perdido o se ha oscurecido en grado suficiente para estimarlos como verbos auxiliares. (This view—and indeed even the wording of it—is followed very closely by the Real Academia Española (1973, 444–5).) Gili Gaya is not the first to use the concept of semantic loss in classifying auxiliary verbs. Lenz (1925, 384) uses this as his main criterion, and argues that tener, ir, seguir, and llevar, in examples such as tengo escrito, tengo que escribir, voy a escribir, sigo enumerando, are auxiliaries. For ‘en todos estos ejemplos los verbos tener, ir, seguir, llevar, han perdido completamente su concepto propio, y sirven sólo para modificar la acción de los verbos principales’. Criado de Val (1958, 129) takes a similar view: ‘Para que un verbo tenga categoría de auxiliar es preciso que el uso haya desgastado su significado original, y se haya gramaticalizado, permitiendo así su fusion más o menos estrecha con participios, infinitivos o gerundios.’ Roca Pons (1958, 11) defines periphrasis as ‘la union de una palabra que pierde total o parcialmente su sentido concreto originario con otra que conserva su valor conceptual, modificado por el valor más o menos funcional que ha pasado a adquirir la primera.’ 1 Particularly in more recent work on auxiliary verbs, linguists have become increasingly critical of this means of defining auxiliary status. Among its major critics as far as Spanish is concerned are Launay (1980), Klein (1982), and Hernanz (1980; 1982). 2 As Launay is at pains to point out, he does not deny the relevance of semantics in considering auxiliary verbs; he merely rejects semantic considerations as a criterion for the definition of auxiliaries: ‘No se trata, que quede claro, de negar tajantemente que el auxiliar “cambie de significado” sino de negar que tal cambio sea un criterio válido [for assigning auxiliary status]’ (p. 45). A similar view is expressed by Rojo (1974, 30): Aceptamos que todos los auxiliares están gramaticalizados, esto es, semánticamente debilitados con relación a su uso como verbos independientes. Ahora bien, no podemos utilizar el criterio semántico—en esta forma al menos—para delimitar entre verbos auxiliares y no auxiliares o entre usos auxiliares y no auxiliares del mismo verbo. A major criticism levelled at the ‘loss of meaning’ approach by both Klein and Launay is that this criterion is difficult to define and to apply consistently, and that it produces widely differing results. Klein criticises the Real Academia Española (1973) for its failure to provide adequate guidelines for distinguishing between ‘loss of meaning’ and ‘differences of meaning’ and for providing conclusions which are seemingly inconsistent with the semantic criteria which have supposedly produced them. Launay shows up the inconsistency of Hamplová’s (1968) approach, for while she uses the ‘loss of meaning’ criterion in her definition of auxiliary verbs, she goes on to include in her list of auxiliaries verbs such as empezar, seguir, terminar, ‘que aún en unión de las formas no personales mantienen su significado sin alteración’. At the other end of the scale, Gili Gaya (1955, 105–6) has such a strict interpretation of his definition of auxiliary verbs that he reaches the conclusion that while estar is an auxiliary in a sentence such as Está
Tener+Past participle
12
pintando la puerta, it is not in sentences such as Está durmiendo tranquilo or Está viviendo con sus padres, as here estar has its full meaning. 3 The question remains whether these inconsistencies are the result of an inherent defect in the criterion used to produce the classification, or whether they arise from inadequacies in the way the criterion is applied. However, it would seem that given the very subjective nature of the criterion and the difficulties in measuring the degree of semantic loss, a certain amount of inconsistency is inevitable. There is another aspect to this question. The fact that such different classifications are produced may be due to the nature of language and of language change. Verbs may not necessarily fall neatly into three categories, as proposed by Launay, or into the two mutually exclusive categories of ‘auxiliary’ and ‘non-auxiliary’ postulated by Klein, or into any other number of discrete classes. It will be suggested later that even with the more clearly defined and consistently applied criteria used by linguists such as Launay and Klein, decisive results are not always achieved. Some of the criticisms hold true even if one is prepared to accept a less rigidly structured view of language than that apparently held by those who seek clear-cut divisions between easily separable categories. Launay’s main criticism—that the ‘loss of meaning’ approach forces one into an arbitrary choice of ‘principal meaning’ for a verb—still applies: ‘nos obliga a atribuir a cada verbo un “significado propio” que no es más que el significado que tiene en otros contextos. Se plantea entonces el problema de saber cuál de los múltiples significados de un verbo merece ser declarado “propio”’(1980, 45). It may not be necessary to describe this in terms of ‘loss of meaning’ or principal meaning. Consider a sentence such as El bedel tiene cerrada la puerta. Here, one may need to decide whether tener has a ‘literal’ meaning—in this case probably a meaning similar to that of mantener or to that expressed by tener in sentences of the type Tengo a mi hijo enfermo—or a grammatical interpretation, with tener acting as an auxiliary. But in order to do this one need not refer to meaning having been lost: one can treat the two possibilities as, to quote Launay, ‘dos valores discursivos sin más de una misma forma lingüística’. However, one still has to distinguish between the different values of a verb, to decide what constitutes ‘the same meaning’ and what ‘change of meaning’. One also has to decide which differences of meaning justify the assignation of auxiliary status and which do not: a verb may have several values, none of which one would want to consider as having an auxiliary function. (On this point, see Klein (1982, 115).) Roca Pons (1979, 315) makes the point that grammaticalisation depends, not on the loss of the original semantic value, but on the lack of any lexical value, and the existence of a primarily grammatical rather than semantic function: Las afirmaciones que suelen encontrarse en muchas gramáticas, según las cuales los verbos auxiliares suponen una pérdida del sentido originario o léxico, son muy confusas. La pérdida de un sentido originario no supone necesariamente una gramaticalización. La llamada pérdida de un sentido léxico tiene más justificación, pero debe partirse de un concepto muy claro sobre las relaciones entre léxico y gramática, de acuerdo con el cual puede pasarse gradualmente de uno a otro aspecto, por lo menos dentro de ciertos límites.
The question of auxiliary status
13
Hernanz (1982, 448–56) argues that while grammaticalisation and loss of meaning may be important diachronic considerations, they have no place in a synchronic description. She prefers to replace these notions with that of polysemy. Discussing grammaticalisation, she states (pp. 450–1) that lo más correcto es entender este concepto en un sentido estrictamente diacrónico, es decir, como el proceso mediante el cual una palabra (ex.g. haber) se va vaciando paulatinamente de significado hasta quedar reducida a un oficio puramente gramatical. Desde un punto de vista sincrónico, en cambio, creemos más operativo…prescindir del citado término y hablar en aquellos casos en los que se ha llegado a un desdoblamiento léxico, no de un valor ‘auxiliar’ (o ‘gramaticalizado’) y otro literal subsumidos en una pieza verbal única dotada de un significado ‘básico’, sino de dos verbos diferentes. Such a view would be compatible with Roca Pons’ proposal of ‘lack of lexical value’ as a more useful criterion than ‘loss of semantic value’. If one were to use any sort of semantic criterion at all, it would not be based on a comparison of different uses of the ‘same’ verb, but would concentrate on the semantic content of the verb as it is used in certain contexts. If in a particular context a verb has no semantic content, then it will be classified as an auxiliary in that context. There is no need for comparison with other uses of the same verb. This would circumvent a problem which, as Launay points out, arises with verbs such as haber and soler, which function solely as auxiliaries. In this case there is no full lexical meaning in the modern language with which to compare auxiliary usage in order to test for loss of meaning. Such a comparison would have to be diachronic rather than synchronic. This could lead to an unacceptable confusion of synchrony and diachrony, in which justification of judgements about the modern language would rest on historical criteria. While most verbs which are used as auxiliaries do also function as lexical verbs, the fact that this does not apply to all auxiliary verbs leads one to look for another criterion which would cover all cases. Launay’s second major criticism of the semantic approach is that ‘este criterio plantea unos problemas insolubles de “graduación”: si se admite que pueden existir “significados enteros” y “significados alterados”, ¿a partir de qué grado de alteración será lícito decir que un verbo es auxiliar? ¿Dónde situar el límite?’ (1980, 45). But this does not necessarily constitute a criticism of the method used to classify the verbs. If one finds that there is a continuum ranging from ‘fully grammaticalised’ to ‘fully lexical’, then if one wants to make an absolute division between auxiliaries and non-auxiliaries, one is faced with the problem of placing an arbitrary dividing line. This is then the ‘fault’ of the language, not of the method used to classify it. A similar answer could be given to Schogt’s (1968) rejection of a class of semi-auxiliaries which would help to accommodate the fact that auxiliary status is not an ‘all or nothing’ matter: Le terme de semi-auxiliaire représente un compromis assez généralement accepté dans les grammaires, mais l’introduction d’un terme intermédiaire a le désavantage de créer deux zones de démarcation au lieu d’une, car il
Tener+Past participle
14
faudra séparer les semi-auxiliaires aussi bien des verbes autonomes que des verbes auxiliaires proprement dits. (p. 6) The problem may lie in the attempt to find strict demarcation lines rather than in the positing of varying degrees of auxiliary status. However, while their criticisms are themselves not exempt from criticism, in general, critics such as Launay show up the weaknesses of the ‘lexical versus grammatical meaning’ method of classification, although they do not deny the relevance of semantic considerations. 1.1.2 Semantic loss as applied to tener +past participle Let us now examine how the criterion of semantic loss can be applied to tener+past participle. Tener appears to possess a range of literal meanings which can be contrasted with any possible auxiliary value. It will be argued later that even auxiliary uses cannot all be classified in the same category. But for the moment let us consider only ‘full’ values of tener. Cano Aguilar (1981, 96–106) puts forward various different values. Although he describes tener as a verb of possession, he suggests that its values far exceed the limits of the concept of ‘possession’. As an example he shows that while Yo tengo un perro expresses possession, one cannot regard Este perro no tiene amo (where the dog appears as the subject and thus as the ‘possessor’) in the same light, as in reality the dog is owned and does not own anything. Cano Aguilar also shows that in many cases tener is what he terms ‘un verbo estatodescriptivo’, indicating a relation between certain characteristics or qualities and the subject: e.g. Las derechas y las izquierdas…tienen cosas buenas y razonables; No tenemos otro móvil que el egoísmo; El túnel de castaños tiene cerca de dos kilómetros en línea recta; Nuestro padre tiene ahora setenta y cinco años. Tener can also indicate availability, as in Tienes todo el día para ir a la capilla or Allí, a mano derecha, tenía la verja de la casa por la que preguntaba. It can also have a value similar to that of sostener (e.g. Es como si tuviera sobre las espaldas una de las pirámides de Egipto) or tomar (e.g. Ten el libro que buscabas). Cano Aguilar also points out the important distinction between tener as an expression of inalienable possession and tener with alienable objects. He shows that with objects of inalienable possession, especially with kinship expressions, there is no true possession, in spite of the fact that one can say mi padre, el padre de Juan etc. in the same way that one can say mi libro, el libro de Juan. With objects of inalienable possession tener either appears without an article, in which case it describes a state (tengo padre ‘I am not an orphan’) or if the object has an article it has to have a complement: Tengo la novia en la puerta; Tiene el padre que se merece (but not *Tengo la novia; *Tiene el padre). Similarly with parts of the body. If the object has no article—El hombre tiene cabeza, tronco y extremidades; La mosca tiene alas—tener is expressing a property of the subject. If the object has an article it must also have a complement: El hombre tiene la cabeza más pequeña que el elefante; La mosca tiene las alas para volar (but *El hombre tiene la/una cabeza; *La mosca tiene las/ unas alas). In a sentence such as Tengo los ojos azules there is no real expression of possession. Although the eyes must be ‘mine’, this is
The question of auxiliary status
15
not the salient point: what is being expressed is the colour of the eyes. It will be seen later that this distinction between alienable and inalienable possession and the question of whether tener expresses primarily possession or a relation between object and complement are questions which play an important role in the study of tener as a possible auxiliary verb. Navas Ruiz (1963, 97–9) distinguishes four major values in current use of tener. These are possession; possession and duration (in which either element may predominate, to such an extent that tener may indicate almost pure duration); expression of a relation; and to consider, as in Tengo por cierto que esta tendencia se halla, or También es cierto que se le tuvo por loco. Roca Pons (1958, 113) does not make such clear distinctions between the different values, claiming that such a classification falls outside his area of study. But while he groups all uses of tener as a main verb under the broad heading of ‘possession’ he uses this term to cover a very wide conceptual area, and recognises various different degrees of possession. Whatever the precise classification one adopts, it seems clear that tener cannot easily be reduced to a single value. One is therefore faced with the problem of deciding which value to take as the principal one when discussing the question of change of meaning. In general, linguists and grammarians who have based their views of auxiliary status on the ‘loss of meaning’ criterion have taken some notion of literal, alienable possession as the main value (although they have often failed to make this choice explicit, and have discussed ‘loss of meaning’ without first clarifying with which meaning they are comparing possible auxiliary usage). While instinctively this may seem the obvious value to choose, it will be seen later that this choice does cause problems in deciding where to draw the line between full and auxiliary usage. If we view semantic criteria as applying to lack of lexicalised meaning rather than to semantic loss then some of the problems are removed. We no longer have to decide which is the principal meaning of tener. Instead we have to decide for any instance of tener+past participle whether tener has a lexical meaning or a grammatical function. But this is not easy. In Cano Aguilar’s illustrations of tener as ‘un verbo estato-descriptivo’ tener has little semantic content. Navas Ruiz suggests that tener can be used to express duration; again there is little apparent semantic content here. Yet in these cases tener is functioning as a main verb and has no possible auxiliary function. If tener as a main verb sometimes has little semantic content, why should tener be classed as an auxiliary in other examples where it apparently has little semantic content, just because it appears with a past participle? The relation between tener as a marker of duration and the possible auxiliary use of tener is discussed in more detail in section 2.5.4. However, we can already see here that even if we choose ‘lack of lexicalised meaning’ rather than ‘semantic loss’ as the main semantic criterion, tener does not lend itself to easy analysis. 1.1.3 Functional criteria Schogt describes this means of classifying auxiliary verbs in the following terms: ‘L’auxiliaire sert à exprimer un élément a) de temps: je mange~j’ai mangé; b) de mode: je travaille~je peux travailler; c) de voix: je transporte~je suis transporté; d) d’aspect: je
Tener+Past participle
16
lis~je suis en train de lire’ (1968, 6). In general, this method of classification is used in conjunction with either semantic criteria or formal criteria. As Schogt points out, it is almost impossible to decide in some cases whether a verbal construction expresses mood, tense, or aspect, and different grammars produce different classifications. Schogt also suggests that such distinctions are not always desirable: ‘On peut d’ailleurs se demander s’il est vraiment nécessaire de séparer les trois categories du temps, du mode et de l’aspect pour les auxiliaires, là où les temps simples les combinent souvent’ (1968, 8). 4 It may be possible to do for auxiliaries what has sometimes been done for simple verb forms, which is to isolate a main value—be it temporal, modal or aspectual—and to classify auxiliaries according to this value. But even if this problem is circumvented there is a second problem: given the difficulty of classifying the different functions of auxiliaries, one is likely to finish by having to accept all the nuances they express. In this case, where does one draw the line between auxiliaries and main verbs? In his treatment of French auxiliaries, Schogt (1968, 9) claims that ‘Si vouloir (travailler) passe dans la catégorie des auxiliaires, il faut également y admettre aimer (à) (travailler); si les constructions se mettre à (travailler), commencer à (travailler) expriment la nuance inchoative, pourquoi ne pas accepter cesser de (travailler), finir de (travailler) comme termes plus ou moins opposés?’ 5 1.1.4 Formal criteria In practice, most linguists make use of formal criteria, either implicitly or explicitly, in classifying verbal periphrases. If they did not, there would be no reason to include only constructions of the type ‘finite verb form+non-finite verb form’ in a list of periphrases. Pountain (1983, 23–4) makes the point that an aspectual or temporal value may be expressed by a verbal periphrasis of this type, or by other lexical means. To illustrate the point he compares French venir de+infinitive and Spanish acabar de+infinitive with use of the adverb appena in Italian to express a similar value. If one does not have recourse to formal criteria, then there would be no immediately obvious reason for including the French and Spanish examples as possible auxiliary constructions, while not considering the Italian functional equivalent. However, few people rely exclusively on formal criteria in delimiting auxiliary verbs, although Pottier (1964a, sec. 12.2) states that ‘Dans un complexe verbal le dernier élément est dit auxilié, les précédents auxiliants’, and that ‘La notion de l’auxiliant doit être prise dans son sens le plus large: tout verbe qui n’est pas le dernier dans une suite verbale.’ This implies that Pottier is relying purely on formal criteria. Such a view has been rejected by most linguists on the grounds that it groups together under the heading of auxiliant verbs which do not share the same functional characteristics, and many of which one would not usually classify as auxiliaries. For example, in his study of Galician periphrastic constructions Rojo (1974, 31) argues that in cases such as vin chegá-lo tren; veu cantando; chegóu canso/cansado, there is no auxiliary, although this formal criterion would clearly classify ver, vir, and chegar as auxiliaries. A similar situation occurs in Spanish (e.g. sale corriendo, fue a comprar pan).
The question of auxiliary status
17
In his treatment of the English verb, Palmer (1965; 1974) distinguishes between auxiliary verbs and catenatives, such as keep, want, like, see, which combine with other verbal forms—and thus superficially appear to have the same structure as auxilaries—but which do not share the characteristics of the class of auxiliary verbs. 7 If one makes such a distinction there is no need to consider all finite verbs in sequences of ‘finite verb form+non finite verb form’ as auxiliaries. While distinguishing between verbs which can appear first in sequences of this type from those which cannot, one can also subcategorise a more restricted set of verbs which have certain characteristics not displayed by other catenatives and which can be considered auxiliary verbs. Palmer’s study suggests that a threeway division of this sort, between auxiliaries, catenatives, and the remaining full verbs is useful, as each group has certain characteristics not shared by the other groups; also, the catenatives behave like other full verbs in some ways, but have certain properties in common with the auxiliaries. While Palmer is dealing exclusively with English verbs—and it does not follow automatically that his conclusions would be applicable to Spanish—it will be seen later that some Spanish verbs which can provide the first element in a ‘finite verb+non-finite verb’ sequence have characteristics not shared by other verbs which can occur in a similar position. This will be considered in greater detail when the syntactic properties of auxiliary verbs are examined. At the moment it is necessary only to point out that by grouping all catenative verbs together as Pottier suggests, some important points about the characteristics of certain of these verbs are missed. In an earlier work Pottier (1961, 325) has a more restricted view of auxiliary verbs: ‘auxiliar es todo verbo que es “incidente” de otro verbo en un mismo sintagma verbal’. For example, when considering auxiliaries which take an intervening preposition, Pottier distinguishes between constructions of the type estoy para decirlo a tu padre and those of the type vengo para decirlo a tu padre. He includes the former in his definition of auxiliary status as the verbal construction shows a ‘functional and semantic unity’ absent in the latter constructions, which are accordingly excluded from his study. Among ‘finite verb+ infinitive’ constructions he restricts auxiliaries to those cases where the two elements have the same subject (quiero hacer, suelo hacer, sé hacer, temo hacer, etc.). Thus verbs in constructions such as permito hacer, veo hacer, oigo hacer, are excluded. He has no such criteria for verbs in constructions with the past participle, but it would seem reasonable to assume a similar requirement here. For example, if Tengo escrita una carta does not entail that the subject of tener wrote the letter, then tener would not be an auxiliary. As for ‘finite verb+gerund’ constructions: Debe considerarse verdadero complejo de auxiliaridad el sintagma que no pueda transformarse en grupo disjunto sin que cambie su significado: ‘está diciendo’ no equivale a ‘está y dice’, en tanto que el significado de ‘habla durmiendo’ no se altera si lo descomponemos en ‘habla y duerme’ o ‘habla mientras duerme’. (p. 328) While these criteria do limit the range of auxiliaries much more than in the 1964 work, Pottier still includes various verbs which many linguists would not consider to be auxiliaries: e.g., quedar por+infinitive, pensar+infinitive, temer+infinitive,
Tener+Past participle
18
seguir+gerund. In his assessment of Pottier’s approach Rojo (1974, 38) also points to a lack of a unified concept of auxiliary status and criteria for defining it. Schogt claims that formal criteria are of little help in considering French auxiliary verbs. He suggests that word order helps to distinguish between auxiliary and nonauxiliary use of avoir (compare j’ai perdu l’argent~j’ai l’ argent perdu) but that this is of no help in the case of être (compare il est ouvert (=‘il est franc’)~il [le musée] est ouvert (actuellement)~il est ouvert (=‘on l’ouvre’)). He points out that ‘La confusion formelle est favorisée par le caractère ambigu du participe qui tient à la fois de l’adjectif et du verbe’ (1968, 10). One finds a similar problem with Spanish tener+past participle. Word order is of little help in deciding whether the past participle is functioning verbally or adjectivally. Both ‘tener+noun+past participle’ and ‘tener+past participle+noun’ are possible (e.g. Tengo mucho escrito~Tengo escrito mucho). Similarly, where tener combines with a noun and adjective, although ‘tener+noun+adjective’ is the usual word order, the sequence ‘tener+adjective+noun’ is also possible (e.g. Tengo limpia la casa; Ten presentes a nosotros). Thus word order does not serve to differentiate clearly between participial and adjectival function, or between main verb and auxiliary. There is another possible formal criterion: clitic pronoun raising (see, for example Lenz 1925, 384–5; Roca Pons 1979, 314–16; Pountain 1983, 25–6). Roca Pons’ use of this criterion, on his own admission, fails to provide clear results. While he groups verbs together on the basis of whether or not they allow clitic pronoun raising, it does not follow that those which do allow raising are auxiliaries. In comparing this group with more fully grammaticalised verbs such as ser and estar he states (1979, 314–15) that Es muy diferente la situación de otros verbos considerados a veces auxiliares, como el grupo de los llamados modales, en los cuales es evidente la falta de un proceso de gramaticalización o de una función gramatical como en los anteriores [ser and estar]. Estos verbos forman, sin embargo, un grupo relacionado, de algún modo, con el concepto de auxiliaridad Por ejemplo, no sólo se combinan con el infinitivo, sino que lo hacen con una mayor cohesión que otros verbos, admitiendo la presencia de una forma pronominal débil… tanto al final como al principio de la construcción formada por los dos verbos. Ciertamente, otros verbos que no suelen considerarse auxiliares se agrupan con los modales desde el punto de vista formal que acabamos de indicar. Given this state of affairs, clitic pronoun raising would appear to be almost totally irrelevant to the question of auxiliary verbs. It could be a very weak filter for sorting verbal constructions: if a verb does not allow raising then it is not an auxiliary; but if it does allow raising then nothing is proved. Pountain suggests that if the criterion is used, a distinction should be made between those constructions where the clitic may come before or after the verb phrase (Lo quiero hacer~Quiero hacerlo) and those which allow only preposed pronouns (Lo he hecho~*Helo hecho~*He hécholo). This would mean that while haber+past participle and tener +past participle would be included, estar+present participle would be excluded. However, most people would intuitively feel that estar+ present participle is more fully grammaticalised than tener+past participle. The fact that the pronoun is obligatorily
The question of auxiliary status
19
placed at the beginning of the tener+past participle construction, and may appear either before or after the estar+present participle construction probably has less to do with grammaticalisation and auxiliary status than with the fact that one can never have the sequence ‘past participle +clitic’, while ‘present participle+clitic’ is possible. (For examples, see Ramsey and Spaulding (1956, 364–5).) 1.1.5 Paradigmaticity A further criterion is that of full paradigmaticity. But this test is not without its problems. Schogt (1968; 1970) argues that auxiliaries are defective and appear only in certain tense forms. Pountain (1983, 27) points out that while constructions such as French aller+infinitive, Spanish ir a+infinitive, and Italian stare+present participle are available only in a limited number of tenses, the more grammaticalised French, Spanish, and Italian perfect auxiliaries are available in every tense. Thus from one point of view, the more grammaticalised a verb is, the less likely it is to have a full paradigm; the other point of view leads to the opposite conclusion. 1.2 POSITION OF TENER +PAST PARTICIPLE What emerges from a consideration of these criteria is that some of them are hard to apply consistently, some produce conflicting results, some are too vague or not selective enough, and none provides a clear, unequivocal definition of auxiliary status. One seems to be left with the somewhat unsatisfying conclusion that there is no clear means of defining auxiliary status, and that all attempts to delimit the field must be based on a rather vague eclectic method which relies heavily on the choice made by the individual as to which criterion should be applied in which circumstances. Alcina and Blecua (1975, 780) reach a similar conclusion: ‘No hay, de hecho, criterios objetivos suficientemente explícitos para fijar el límite más allá de la particular y subjetiva opinión de quien describa el fenómeno.’ Later we shall look at possible solutions to this apparent impasse, but for the moment, having looked at the various criteria and how they seem to apply to tener+past participle, let us now see how this construction is classified in traditional descriptions of modern Spanish. 8 For most writers, tener is an auxiliary verb. The Real Academia Española (1973, 449) states that ‘los verbos llevar, tener, estar y ser, y a veces traer, quedar y dejar, forman perífrasis verbales como verbos auxiliares, desposeídos por lo tanto de su significado propio’. Once again, this view is taken from Gili Gaya (see Gili Gaya 1955, 107). Alcina and Blecua (1975, 781) include tener+past participle in their list of frases verbales (a frase verbal consists of a verbo auxiliar and a verbo conceptual or auxiliado). There appears to be no doubt in Pottier’s mind that tener is an auxiliary. In fact, it is one of the ‘cuatro auxiliares fundamentales’ (1961, 327)—the others being haber, ser, and estar. But other writers are more cautious in their classification. Bello and Cuervo (1898, 161) classify haber and estar as the principal auxiliaries. However, later (p. 187) they have a section entitled ‘Formas compuestas en que entra el auxiliar tener’ But implicit in their
Tener+Past participle
20
description is an assumption that tener is not a ‘full’ auxiliary like haber. They state (p. 116) that reconoceremos, pues, dos especies de participio: el que para diferenciarlo llamaremos participio adjetivo, y el participio sustantivado, que es el que se emplea con el verbo haber. Este segundo es en grado eminente un participio, porque participa de la naturaleza verbal, acomodándose a todas las construcciones del verbo de que nace. But tener, even when functioning as an auxiliary, is deemed to combine with the ‘participio adjetivo’ (p. 187). If the participle has an adjectival function then tener is less of an auxiliary than haber, which combines with a verbal participle. Tener+past participle in a sentence such as Tengo la camisa lavada is a similar construction to that of tener+ adjective in Tengo la camisa limpia, the only difference being that the latter contains an adjective, whereas the former contains a participle with an adjectival function. Similarly Criado de Val (1958) counts tener as a grammaticalised auxiliary, but underlines that it is not so fully grammaticalised as haber. Implicitly at least, Criado de Val recognises the existence of a continuum of grammaticalisation; for him, auxiliary status is not an ‘all or nothing’ classification. Roca Pons (1958) makes a similar point. While tener may be classified as an auxiliary, it is not so fully grammaticalised as haber—the grammaticalised auxiliary par excellence. ‘En los usos auxiliares del verbo haber no puede decirse que ha quedado alguna representación más o menos alejada de su antiguo significado. Con tener, en cambio, es dificil, en algunos casos, decidir si nos encontramos ante un verdadero auxiliar’ (p. 12). This leads to another important point. Not only is there a continuum of auxiliary status on which different verbs can be placed; the uses of any one verb also form a continuum— some examples may be more grammaticalised than others. Roca Pons divides examples of tener +past participle into two broad categories, based on the presence or absence of possession. But he emphasises that ‘no debe olvidarse que el hecho de haber establecido dos grandes grupos…no supone una división tajante y absoluta’ (p. 156). For Gili Gaya (1955, 97) tener is an auxiliary, but he points out in a comment on Seco’s (1930) grammar that tener+past participle is not always a clear auxiliary construction. Nota con razón R.Seco…que las expresiones perifrásticas formadas con tener y un participio pasivo, del tipo yo tengo escrito un drama, no son conjugaciones perifrásticas en el sentido que venimos entendiéndolas…. La frase anterior—dice—no parece tener otro sentido que el que le presta la presencia del participio adjetivo escrito, complemento predicativo en la oración. As can be seen from this quotation, Seco does not accept tener as an auxiliary. He claims that tener has its full semantic value, and that examples where it does have an auxiliary function (e.g. Tengo visto a Pedro=He visto a Pedro) are regional. 9 It is true that there is considerable regional variation: auxiliary usage of tener is far more common in northwestern Spain than in the rest of the country, and this is something which will be
The question of auxiliary status
21
discussed in chapter 5. However, this use of tener is not purely regional, as we shall see in chapter 2. One cannot tell whether or not Alonso and Henríquez Ureña (1945) regard tener as an auxiliary. They make the following comments on auxiliary verbs (p. 120): Pueden ser considerados auxiliares todos los verbos que se combinancon el infinitivo, con el participio o con el gerundio de otro, para expresar aspectos especiales de la significación del segundo. Así, losmodales (poder, querer, saber, deber, soler), los verbos de movimientocon gerundio (ir, venir, andar), quedar, dejar y tener que…Pero los verdaderos auxiliares son haber, ser y estar. They do not state whether this is an exhaustive list of auxiliary verbs, or whether these verbs are merely examples of the different types of auxiliaries. Tener is not mentioned. Hernández Alonso (1984) puts forward a clear exposition of the nature of auxiliary status, but unfortunately deals in detail only with infinitival constructions, restricting himself to very brief comments on periphrases with the gerund and the past participle. He states (p. 391) that in an example such as Te tengo dicho que te calles ‘se percibe claramente la diferente función de este verbo conjugado y del participio, a pesar de la lexicalización del sintagma’. However, he does not state why this is the case, merely going on to conclude that ‘por ello ofrece grandes dificultades interpretar tales grupos como perífrasis verbales’. It can be seen from the above summary that while many writers regard tener as having an auxiliary function, there is disagreement over the question of how fully tener has grammaticalised, and the exact nature of its auxiliary status. Most writers seem to regard tener as being less grammaticalised than haber but more grammaticalised than some other verbs. There is also no clear consensus of opinion as to which examples of tener+past participle should be classified as instances of auxiliary usage, and which should be discounted. Let us now look in more detail at some of the more recent syntactic criteria which have been used in the classification of auxiliaries, to see whether these provide clearer results.
1.3 SYNTACTIC CRITERIA Although Launay’s (1980) study contains much useful work on the question of auxiliary status, it concentrates exclusively on verb phrases with the infinitive and the gerund. There is no mention of how constructions such as tener+past participle should be analysed. In the concluding paragraph Launay mentions the development of haber+ past participle and its passage from lexical verb to auxiliary verb; but there is no syntactic analysis of this construction, and none of his criteria is illustrated by the use of an auxiliary+past participle construction. Although in his introduction he appears to be proposing to consider the general question of auxiliary verbs and periphrastic constructions, he later specifically limits himself to constructions involving the infinitive
Tener+Past participle
22
and the gerund, although no justification is given for the omission of constructions with the past participle. Fontanella de Weinberg (1970) and Klein (1982) do include participial constructions, but even in their articles greater prominence is given to constructions with the infinitive and the gerund. The only participial construction considered by Fontanella de Weinberg is that of haber+past participle. Klein includes the tener+past participle construction as a periphrasis and groups it with haber+past participle. He suggests that the omission of this construction from Fontanella de Weinberg’s list was probably an oversight. What results are arrived at when the suggested syntactic criteria for classification are applied to tener+past participle? Launay’s first major distinction is between what he labels class ‘A’ verbs and class ‘B’ verbs. Class ‘A’ verbs appear in constructions of the type (1)
Estaban contándolo
(2)
Iban a contarlo
and admit transformations producing the following types of constructions: (1a)
¿Qué estaban haciendo?
(2a)
¿Qué iban a hacer?
(1b)
¿Qué estaban contando?
(2b)
¿Qué iban a contar?
(1c)
Lo que estaban contando
(2c)
Lo que iban a contar
(1d)
Lo estaban contando
(2d)
Lo iban a contar
Class ‘B’ verbs do not admit these transformations: (3)
Lloraban contándolo
(4)
Vinieron para contarlo
(3a)
*¿Qué lloraban haciendo?
(4a)
*¿Qué vinieron para hacer?
(3b)
*¿Qué lloraban contando?
(4b)
*¿Qué vinieron para contar?
(3c)
*Lo que lloraban contando
(4c)
*Lo que vinieron para contar
(3d)
*Lo lloraban contando
(4d)
*Lo vinieron para contar
For constructions with the past participle transformation (d) is irrelevant, as the clitic pronoun cannot be attached to the participle and must come before the main verb. However, transformations (a) and (c) can be applied (transformation (b) will be discussed below). Applying them to various sentences we obtain the following results: 10 (5)
Ese chico tiene a su madre preocupada
(5a)
¿A quién tiene preocupado?
(5c)
La que tiene preocupada
(6)
Tengo la casa barrida
(6a)
¿Qué tengo barrido?
(6c)
Lo que tengo barrido
(7)
Tengo pedido el libro
(7a)
¿Qué tengo pedido?
The question of auxiliary status
(7c)
Lo que tengo pedido
(8)
Tienen dicho que mañana no va a haber clase
(8a)
¿Qué tienen dicho?
(8c)
Lo que tienen dicho
23
All the sentences allow these two transformations. This would suggest that in the tener+past participle construction tener is a category ‘A’ verb. But consider a sentence such as: (9)
Tengo un coche prestado
Regardless of whether tener is a full verb expressing possession (‘I have a car that has been lent to me’) or a marker of resultant state or of perfect aspect (‘My car is lent to someone’/‘I have lent my car’) it permits the two transformations: (9a)
¿Qué tengo prestado?
(9c)
Lo que tengo prestado
One can even find examples of tener+adjective which allow the same transformations: (10)
Siempre tiene presente que es un problema muy grave
(10a)
¿Qué tiene presente siempre?
(10c)
Lo que siempre tiene presente
This would lead to the conclusion that tener+adjective is a periphrasis. If one considers transformation (b)—¿Qué tiene hecho?—the situation is complicated further. Some native-speaker informants did not accept this construction at all, except where the original sentence contained tener hecho. So for these speakers tener seems not to be an auxiliary. Some other native speakers did accept the construction, but only in certain very limited temporal and aspectual contexts. For example, one informant gave results along these lines: (5)
Ese chico tiene a su madre preocupada
(5b)
*¿Qué tiene hecho?
(6)
Tengo la casa barrida
(6b)
¿Qué tengo hecho?
(7)
Tengo pedido el libro
(7b)
¿Qué tengo hecho?
(8)
Tienen dicho que mañana no va a haber clase
(8b)
*¿Qué tienen hecho?
Thus for some speakers, under this criterion examples (5) and (8) are excluded from the auxiliary category, while (6) and (7) are included.
Tener+Past participle
24
If one considers the transformations as a group, then sentences (6) and (7) satisfy all the criteria for category ‘A’ status, while sentences (5) and (8) satisfy some but not others. But Launay does not make provision for such cross-categorial status. Constructions of the type found in sentence (5) can be fairly convincingly excluded from consideration as auxiliary constructions by a test which Launay does not use, but which is put forward by Fontanella de Weinberg. She shows that in constructions such as volvió dormido, volver is a full verb, not an auxiliary, because it permits the interrogative transformation ¿Cómo volvió? An auxiliary verb, such as haber in ha dormido, does not permit the transformation *¿Cómo ha? Thus sentences of the type (11)
Ese chico tiene a su madre preocupada
(12)
Tengo irritados los ojos
(13)
Tengo la pierna rota
can be discounted because they do permit this transformation: (11e)
¿Cómo tiene a su madre?
(12e)
¿Cómo tengo los ojos?
(13e)
¿Cómo tengo la pierna?
However, sentences such as (6) above also permit this transformation: (6)
Tengo la casa barrida
(6e)
¿Cómo tengo la casa?
So this type of construction apparently both contains an auxiliary verb—it satisfies the proposed tests—and does not contain an auxiliary verb—it is discounted by another test. More will be said later on the relative degree of grammaticalisation of the constructions found in sentences (5)–(8), and on the relevance of the related interrogative transformations. Suffice it to point out here that Launay’s criteria do not necessarily provide clear-cut distinctions between categories. 11 Let us assume for the moment that—in some cases at least—tener is a class ‘A’ verb. Launay then makes a further distinction between ‘Aa’ and ‘Ab’ verbs. In order to make this division, Launay examines passive constructions. Verbs which have a directly related passive form belong to class ‘Aa’; those which do not, belong to class ‘Ab’. Thus: Aa
Ab
(14)
Los niños solían tomar el ascensor
(14a)
El ascensor solía ser tomado por los niños
(15)
Juan solía patear a Pedro
(15a)
Pedro solía ser pateado por Juan
(16)
Los niños querían tomar el ascensor
(16a)
*El ascensor quería ser tomado por los niños
(17)
Juan quería patear a Pedro
The question of auxiliary status
(17a)
25
Pedro quería ser pateado por Juan
Launay’s analysis maintains that (16a) is ungrammatical, while suggesting that (17a) is grammatical, but not the passive equivalent of the active sentence given. It seems more accurate to describe both (16a) and (17a) as semantically, rather than grammatically, unacceptable. This semantic incongruity stems from the fact that non-auxiliaries, such as querer, impose selectional restrictions on subjects and objects; auxiliaries, such as soler, do not. The question of selectional restrictions will be discussed further below. Tener+past participle does not permit this passive transformation. While with the constructions that Launay considers under ‘Ab’ the passive is semantically unacceptable or not related to the active sentence, with tener+past participle the passive is grammatically unacceptable. Again, this is related to selectional restrictions; in general tener must combine with the past participle of a transitive verb, and thus cannot combine with the past participle of ser to form a passive construction. (18)
Paco tiene escrita la carta
(18a)
*La carta tiene sido escrita (por Paco)
(19)
Juan tiene mandado el libro
(19a)
*El libro tiene sido mandado (por Juan)
This seems to place tener+past participle in class ‘Ab’. Launay limits himself to this test for the delimiting of classes ‘Aa’ and ‘Ab’, on the grounds that one test is adequate and that other tests used by Fontanella de Weinberg produce the same results. The category which Fontanella de Weinberg considers to be that of auxiliary verb corresponds to Launay’s ‘Aa’ class. She draws no distinction between what Launay classifies as ‘Ab’ verbs and ‘B’ verbs, which in her classification would all be nonauxiliaries. As tener+past participle appears to belong to Launay’s ‘Ab’ category, one would expect Fontanella de Weinberg’s classification to exclude tener from the class of auxiliaries. As has already been mentioned, Fontanella de Weinberg fails to consider the case of tener+past participle at all. However, there is no doubt in Klein’s mind that her classification should place it with haber, as an auxiliary. Apart from the passive transformation, Fontanella de Weinberg considers as criteria for classifying verbs: selectional restrictions on subject and objects; 12 the possibility of combining auxiliaries with grammatical elements other than verbal ones; question formation. Auxiliary verbs do not impose their own selectional restrictions on subject and object in the way that full verbs do. While Launay does not include this as a criterion for auxiliary status, he does comment on this loss of selectional restrictions. For Launay, verbs of the class ‘Ab’ no longer govern the selection of the object, and are neither transitive nor intransitive: choice of object is dictated exclusively by the verb in nonfinite form. Verbs of the class ‘Aa’—like Fontanella de Weinberg’s auxiliary class—do not govern the choice of either subject or object. Launay states that La primera condición requerida para ser auxiliar es, pues, no ser ni transitivo ni intransitivo, o, más propiamente dicho, es no tener ningún
Tener+Past participle
26
régimen propio, ya que la frase verbal ha de tener el régimen del auxiliado, sea cual sea. Los verbos que tienen su propio régimen en otros contextos deben, pues, renunciar a él cuando funcionan como auxiliares (1980, 55). This lack of ‘régimen’ is shown by the fact that some verbs, when acting as full verbs, may be purely transitive or purely intransitive; but when they combine with other verbs as auxiliaries, they can be either transitive or intransitive, depending on the nature of the verb with which they combine. For example ir is intransitive, but the ‘ir a+infinitive’ and ‘ir +gerund’ constructions can take a direct object if the infinitive or the gerund is transitive: lo voy a hacer; lo voy haciendo. But even in its putative auxiliary role, tener is still almost entirely a transitive verb. In nearly every case, tener must have a direct object complement. There are occasions when tener+past participle occurs without a direct object; these will be discussed in chapter 2. At the moment it is necessary only to state that occurrences of tener+past participle with no direct object are rare, and are restricted to limited syntactic contexts. In general, tener+past participle cannot appear with intransitive verbs, and sentences of the type *Tienen viajado por el extranjero or *Jorge tiene quedado en casa are ungrammatical. Tener has therefore not lost its ‘régimen’ and thus does not fit into the class of auxiliaries, even of Launay’s ‘Ab’ type, despite the fact that other criteria suggest that it should do so. 13 As far as combination with non-verbal elements is concerned, Fontanella de Weinberg notes that in examples such as (20)
Quiere estudiar la cuestión
(21)
Desea dormir
the non-finite verb can be replaced by a noun phrase or by a que clause: (20a)
Quiere un estudio de la cuestión
(20b)
Quiere que estudiemos la cuestión
(21a)
Desea descanso
(21b)
Desea que duermas
Constructions containing an auxiliary do not permit this replacement. Thus there can be no noun phrase replacement of the infinitive in sentences such as (22) Suele estudiar la cuestión (23) Debe dormir Application of this test to tener+past participle suggest that tener is an auxiliary, although this is not always a clear-cut decision. Sentences such as (24) Tiene prestado el libro (25) Tienen prometido venir a las cinco do not relate in the way discussed above to
The question of auxiliary status
(24a)
Tiene el libro (que ha sido prestado)
(25a)
Tienen la promesa
27
As regards question formation, Fontanella de Weinberg notes that for constructions such as (26) Desea trabajar (27) Volvió dormido (28) Vino caminando there are related interrogatives of the form: (26a)
¿Qué desea?
(27a)
¿Cómo volvió?
(28)
¿Cómo vino?
But for constructions such as (29) Puede trabajar (30) Ha dormido (31) Está caminando there are no such related interrogatives: (29a)
*¿Qué puede?
(30a)
*¿Cómo ha?
(31a)
¿Cómo está?
These questions are either ungrammatical or do not relate to their supposed counterpart. This test would place tener among the auxiliary verbs. Consider the following sentences: (32) Tengo escritas muchas cartas (33) Me tienen dicho que hoy no va a haber clase (34) Tienen viajado mucho por el extranjero If tener is a full verb, these sentences should have the following related interrogatives: (32a)
¿Qué tengo?
(33a)
¿Qué tienen?
(34a)
*¿Cómo tienen?
Here again, one finds the same problem as occurred with Launay’s classification. It is true that just as *¿Qué ha? and *¿Como ha? are ungrammatical, in these cases ¿Qué tengo? and *¿Cómo tienen? are either ungrammatical or do not relate to their supposed counterpart. But while ¿Qué ha hecho? is acceptable, as has already been seen ¿Qué tiene hecho? is not always acceptable (see sentences (5)–(8)). A few comments should be made on Klein’s classification. Klein accepts Fontanella’s tests of passivisation, question formation and the appearance of auxiliaries solely with
Tener+Past participle
28
non-finite verb forms. However, he disagrees with the specific theoretical framework within which Fontanella is working, which is based on Chomsky’s (1965) model. Klein attributes many of what he views as the shortcomings of her analysis not to the analysis itself, but to defects in the model. Two of his criticisms are of a technical nature, one resulting from the fact that in Chomsky’s model auxiliaries are not verbs, and the other arising from the fact that some transformations have to apply to elements which do not form a constituent at any level. The third problem is an empirical one: that the model allows only one modal verb per sentence, and thus fails to generate sentences such as Juan debe poder hacerlo. 14 These criticisms relate to the theoretical model used by Fontanella. The difference between the requirements of a particular theory and the more general requirements of a pretheoretical description will be discussed in sections 1.5 and 1.6. At the moment it is sufficient to point out that while Klein disagrees with some of the theoretical details of Fontanella’s account, he agrees with the fundamental assumptions on which it is based. His analysis does not contradict the tests put forward by Fontanella or Launay: he merely suggests a different way of formulating the results, in accordance with his particular theoretical model. Given that we are not concerned here with the details of the theoretical model, this study will continue to concentrate on the tests for auxiliary status rather than on the theoretical analysis of auxiliaries. What emerges from these tests is that tener does not fit neatly into any one category. For Launay, tener seems to belong to category ‘Ab’. But there are problems with the related interrogatives of the type ¿Qué tiene hecho? and with the question of selectional restrictions. Despite the fact that Klein—and presumably Fontanella—place tener unequivocally in the auxiliary category, once again there are problems with interrogatives and selectional restrictions, and also with the passive. Even if one accepts the findings of each linguist and the conclusions these findings seem to lead to, the results are still not consistent. If the category labelled by Fontanella de Weinberg and Klein as ‘auxiliary’ corresponds to Launay’s category ‘Aa’, then the fact that tener should belong to category ‘Ab’ means that in Fontanella de Weinberg and Klein’s terms, it is not an auxiliary; but Klein categorises it as such.
1.4 COMPARISON OF CRITERIA One of the features of the syntactic criteria is that the tests used treat the periphrasis as a whole rather than concentrating on either the finite verb or the non-finite verb in isolation. According to these criteria auxiliary verbs do not impose selectional restrictions on subject and object, but are subordinate to the restrictions of the non-finite verb; they appear only with a non-finite verb form, and so cannot appear in a similar context as a ‘full’ verb, with a nominal complement, but must be linked to the non-finite verb; they do not allow a related question which does not include the non-finite verbal element; they passivise in such a way that the ‘finite verb form+non-finite verb form’ sequence can be considered as a single unit. All these characteristics suggest that the ‘auxiliary+main verb’ sequence functions as a single unit. In contrast, examples of ‘finite verb form+nonfinite verb form’ which do not satisfy these criteria function as two separate units, each of which retains its own characteristics. Even though the aim is to define and classify
The question of auxiliary status
29
auxiliary verbs, one cannot do this by concentrating exclusively on the finite verb form in the sequence, but must consider the way the sequence functions as a whole. The syntactic approach is not the first one to treat the construction as a whole. People working within other very different frameworks had adopted a similar stance many years earlier. 15 But the advantage of the tests of people such as Fontanella, Klein, and Launay is that they provide a means of testing for this unity of the construction. This method of testing for syntactic unity underlies a major difference in their approach, as compared with previous approaches. The change in the type of criteria used to test for auxiliary status stems from a more general change of attitude towards linguistic analysis. The first set of criteria discussed above—those based on semantic loss, and on functional and formal considerations—arises from a methodology in which the main aim is to break the sentence down into its individual components, which are then to be described in terms of their place in the structure of the sentence. Thus items are largely looked at in isolation, or contrasted paradigmatically with other forms within the system. The following quotation from Imbs (1960, 6–7), referring to the French verbal system, illustrates this basically structuralist approach, based on contrasts within the system. Il est incontestable que les formes composées et surcomposées du verbe sont des périphrases, et que cela ne les a pas empêchées d’entrer dans le tableau des formes étudiées en morphologie. C’est une question de savoir si les périphrases verbales formées à l’aide de semi-auxiliaires (aller, devoir, etc.) ou de locutions semi-auxiliaires (être en passe de…) doivent figurer elles aussi dans le tableau des formes verbales. A notre avis la réponse doit être negative: du point de vue purement formel, ces périphrases sont loin de constituer un système, c’est-à-dire un ensemble d’oppositions et de symétries cohérentes. Mais étant donnée la valeur morphématique des verbes et des locutions semi-auxiliaires (leur signification originelle est très affaiblie), ils occupent une position toute proche des formes integrées au système du verbe. With the rise of generative grammar came a different approach. There was a move away from the study of semantic and formal properties of individual linguistic items, and an increased interest in their syntactic properties, with sentences being analysed in terms of their syntactic structure. Given this emphasis on syntax, it is not surprising that there should be a move away from the earlier criteria for auxiliary status towards an increased interest in the combinatorial properties of elements and the formulation of criteria based on the syntactic behaviour of auxiliaries. Even those who are not principally generativists have often been influenced by this change in attitude. For example, the criteria for auxiliary status suggested by Palmer, and discussed in section 1.5, are largely syntactic in nature, as we shall see. Some of the criticisms levelled at the semantic approach by Launay and Klein appear to recur when their syntactic criteria are used. It is true that the tests are more clearly defined and easier to apply consistently. But there is still some difficulty in drawing clear distinctions between categories, and conflicting results still occur. However, once again, it may be that this is not a ‘fault’ of the classification but rather a ‘fault’ of the language. While it is important to have a means of analysing and classifying different lexical items
Tener+Past participle
30
and constructions, one has to remember that one cannot view these classificatory boundaries as rigid barriers. Language does not respond exhaustively to neat compartmentalising. The polyvalency of many forms means that they cannot be fitted into one slot in the descriptive framework—whatever framework one chooses. Many linguists have attempted to give ‘basic’ or ‘residual’ values for verb forms, and to parallel the supposedly transparent morphological structure of the ‘system’ with a neat functional or semantic system. However, as was pointed out in section 1.1.3, temporal, aspectual, and modal values overlap and are not easily disentangled. A further problem is that the morphological structure is not so neat as it might at first appear. It is not easy to decide to what extent periphrastic forms have grammaticalised and should be considered as a systematic part of the structure of the language. The fact that certain constructions are in the process of grammaticalising leads us to the problem of accommodating diachronic considerations within a synchronic description. Not only are some constructions grammaticalising and changing their functions; some fully grammaticalised forms may be undergoing changes in function too. An outline of the problem will be given in section 1.6; in later chapters its ramifications will be considered through the detailed study of tener+past participle, which exhibits some of these ‘problem’ characteristics.
1.5 THE EXISTENCE OF AUXILIARY VERBS Given the difficulties in defining and delimiting the class of auxiliary verbs, one might be tempted to question whether there is any point in even attempting to define a class which appears to have no possible well-defined limits, and to wonder whether there is any need to treat these verbs—when one has eventually decided which they are—as a class apart from other verbs. If there are very few characteristics shared by all the members of this class, and if there are characteristics which they share with other verbs outside the class, why should one make any major divisions between auxiliaries and other verbs? Why not describe the characteristics of each verb without reference to the question of auxiliary verbs? There has been a considerable amount of discussion on this subject. The greater part of it has dealt primarily with the question of auxiliaries in English, and while the specific syntactic arguments used in the discussions cannot be applied to Spanish, the more general conclusions are very relevant. Palmer (1965; 1974) argues strongly in favour of recognising a class of auxiliary verbs in English. Some of his arguments are based on syntactic criteria (behaviour in negative and interrogative phrases, use of an auxiliary to refer back to a main verb, ability of an auxiliary to take nuclear stress in emphatic affirmation); 16 others are based on the distinction which he makes between simple and complex verb phrases. 17 Palmer’s arguments concerning the formal syntactic properties of auxiliaries are specific to English, and have no bearing on the situation in Spanish. However, we have already seen that the ‘simple versus complex verb phrase’ argument plays a significant part in determining Spanish auxiliaries. A further point which Palmer uses in his arguments for auxiliary verbs—and which could apply in principle to Spanish—is the paradigm test. Palmer argues that the
The question of auxiliary status
31
auxiliaries form a closed class—‘there are no other forms which will fit’—and that each example of auxiliary plus finite verb form in the paradigm is a unified whole, in that it cannot be analysed in terms of the individual elements which comprise it. The first of these arguments is circular. As Palmer admits, the fact that no other form will fit is dependent on the way in which the paradigm is defined. If it is defined so as to include the simple forms, the auxiliaries, and no other forms, then it is not surprising that the simple forms and the auxiliaries form a closed class. As for the second point, this is a consequence of the semantic and syntactic unity of a periphrastic form, but by itself it cannot be used to test for periphrases. To say that something is a unit is not to provide a means of testing for a unitary construction. Before attempting to evaluate the weight of Palmer’s arguments, we must turn to the argument against recognising a separate class of auxiliary verbs. Ross (1969) argues in favour of treating auxiliaries in the same way as main verbs. However, as Palmer (1979a, 1) points out, he puts forward ten arguments to show that auxiliaries are verbs, and only two arguments to show that they are main verbs—one of which comes from German, and the other of which is based on language universals. Ross’s arguments will not be considered in detail here, as other arguments have since been put forward which have greater bearing on the main point under discussion here; but it is worth noting that Ross is arguing within a specifically generativist framework, and even more specifically, against Chomsky’s (1965) analysis of auxiliaries. 18 Later I shall argue—as others have done— that while a particular framework may favour, or even require, a particular analysis, it may be that the conclusions which can be drawn from this analysis apply only within that theoretical framework. The ‘auxiliary versus main verb’ argument has been taken up by various linguists (for references, see Pullum and Wilson 1977). Huddleston (1974; 1976a; 1976b; 1980) provides a more general attack on auxiliaries as a separate category of verbs than that provided by Ross. There is no need to look in detail at Huddleston’s rejection of individual criteria put forward by Palmer. 19 What is of interest to us here are the differing attitudes adopted by Huddleston and Palmer over indeterminate boundaries between classes. Huddleston makes two major assertions: that not all so-called auxiliaries share all the supposed characteristics of auxiliary verbs; and that ‘auxiliaries’ have so much in common with other verbs that there is no justification for treating them as a separate category. He concludes that Palmer’s criteria do not motivate the establishment of a separate category because such a category does not permit a more general description of the phenomena in question than is possible without it. In rejecting Huddleston’s conclusions, Palmer’s principal line of argument is that although the boundary between auxiliaries and main verbs is somewhat blurred there is a basic group of auxiliaries which have certain characteristics in common. While other verbs may share some of these characteristics, they do not exhibit the same combination of characteristics. Although not all auxiliaries have all these characteristics, the basic common syntactic behaviour of these verbs is too consistent to be mere coincidence. On the other hand, it is not surprising that auxiliaries have some characteristics in common with main verbs: they are a subcategory of verbs, not a completely separate category (see Palmer 1974, 40). In the arguments for and against a main verb analysis of auxiliaries we come up against the same question of indistinct class boundaries as arose with the tests of
Tener+Past participle
32
Fontanella, Klein, and Launay. Just as earlier we saw the tendency for generativists to look for clearly defined, mutually exclusive categories, Palmer (1979b, 180) notes the generativist school’s dislike of indeterminate categorisation. He criticises the basis on which much of the ‘anti-auxiliary’ argument is founded: that as no clear line can be drawn between auxiliaries and main verbs they are therefore not distinct. This criticism is found elsewhere in his work (Palmer 1979a, 4–5): It is common both in language and in other kinds of experience to find gradients such that at each extreme it is clear that we have quite different phenomena, but with a whole range of indeterminacy between. If we refuse to make ‘arbitrary’ divisions because of this, we shall fail to make any divisions or classifications at all…. To fail to make any division would obscure the essential difference between the items at the extremes. 20 So far this discussion has centred on English auxiliaries, although comparisons have been drawn with Spanish. Obviously, the precise formal properties of English auxiliaries, and some of the tests used to define the group of auxiliary verbs, are not applicable to Spanish. However, similar questions have arisen in Spanish, and some of the conclusions one can draw from the arguments over English can be applied to Spanish. It should be made clear that most of the argument in Spanish has centred on the socalled ‘modals’ (e.g. poder, deber). While people may disagree over specific points of classification, the existence of auxiliaries per se is not seriously questioned. However, the arguments put forward in this disagreement are relevant to the general question of auxiliary status, and of how distinctions between classes should be drawn. Because the arguments centre on modal auxiliaries and are thus not directly relevant to tener+past participle they will be summarised very briefly here, and more attention will be paid to the more general points they raise. Stockwell et al. (1965, 165) reject the classification of verbs such as poder, querer, and deber as modal auxiliaries, on the grounds that they share almost none of the characteristics of English modal verbs: they are conjugated in the same way as main verbs; they accept objects like other transitive verbs; they can combine with other auxiliaries. They conclude that the ‘modals’ ‘are simply verbs which happen to carry lexical meanings that overlap with the meanings of English modals’. Stockwell et al. are writing a comparative work, aimed at the teaching of Spanish to English speakers. Thus the authors are likely to try to draw comparisons between the two languages. But there is no need to use the same tests in the two languages, nor yet any justification for doing so. We have already seen that it is possible to devise tests for auxiliary status which are designed specifically for Spanish. A more convincing argument against a separate class of modal verbs is provided by Klein (1968). He maintains that a category [+M] in the grammar cannot satisfactorily account for the following facts about Spanish: 1 Multiple occurrence of so-called modals, e.g. debo poder cantar. This can be accounted for in only a very ad hoc fashion. 2 Change in meaning related to position, and to whether other verbs are ‘modals’ or not, e.g. debo poder cantar (‘I must be able to sing’), puedo deber cantar (‘I may have to
The question of auxiliary status
33
sing’); puedo haber cantado (‘I can/may have sung’). These changes in meaning cannot be expressed merely by the feature [+M], as such an analysis allows either poder 1 (possibility) or poder 2 (ability) to be substituted in any context where a modal is possible. A similar situation arises with deber, which can indicate either probability or obligation. 3 Negation: both the following are possible: no puedo cantar, puedo no cantar. But according to the ‘modal’ analysis puedo cantar is one sentence, and the grammar allows only one negative per sentence. Klein suggests a main verb analysis for the so-called modals. He maintains that under this analysis all the rules needed to account for the behaviour of modals are already needed elsewhere in the grammar. By treating modals as main verbs, Klein’s analysis allows for multiple occurrence of these verbs. He deals with change in meaning by positing different underlying structures for the different interpretations. Dual negation is explained by positing two underlying sentences, each of which can be negated. The fact that haber and estar are not main verbs explains why they can appear only once in a verb phrase, and why they can be negated only once. Klein concludes that as restrictions needed to account for ‘modals’ are also needed to account for other main verbs ‘the Spanish grammar which treats modals as “simply verbs” is simpler and more general than one which sets up a category [+M]’. 21 Schroten (1974–5) sets out to prove among other things that ‘no hay un solo argumento válido para considerar los verbos perifrásticos como verbos auxiliares’, but later suggests (p. 43) that ‘haber y estar pueden ser considerados como verbos auxiliares, es decir, miembros de AUX, aunque la prueba no es concluyente’. Later (p. 47) he claims that ‘no hay verbo auxiliar en español, si no es haber’. From this point onwards he deals only with what might be considered as modal periphrases. Haber and estar are excluded from the discussion. It is interesting to note that Schroten in effect deals only with infinitival constructions, other periphrastic constructions being specifically discounted: Hemos hecho caso omiso de algunos verbos que se dicen auxiliares: seguir, andar,…los cuales exigen la forma del gerundio del verbo siguiente, y los que requieren un participio pasado, como tener, dejar, ser, llevar, estar…. En un principio, seguir, etc. sustituye a estar; tener, etc. ofrecen problemas especiales que dejaremos de un lado. (p. 43, footnote 2) It appears then that Schroten’s arguments are intended to be applied only to possible modal constructions of the type ‘modal+infinitive’, and that he is not arguing in favour of the total abolition of the category of auxiliary verb in Spanish. What Schroten actually succeeds in doing, through his tests, is to show that the so-called modals have certain characteristics in common with other verbs and also have certain syntactic properties and restrictions on usage which other verbs do not possess. Such conclusions should come as no surprise, and effectively bring us back to the position adopted by Palmer. Hernanz (1982, 429–67) argues against a separate class of auxiliary verbs. However, as she is working on infinitival constructions, possible auxiliary constructions with the gerund or past participle are not discussed.
Tener+Past participle
34
The argument in Spanish is rather more specific than the discussion in English. Noone appears to argue unequivocally in favour of the total abolition of the auxiliary category in Spanish. The argument centres on the status of the modals, not of auxiliaries in general: it is fairly uncontroversial to talk of auxiliary verbs in Spanish, and to assume that there needs to be some way of distinguishing between main verbs and auxiliary verbs. In section 1.4 it was suggested that syntactic tests of a sort such as those proposed by Fontanella, Klein, or Launay are more appropriate than criteria based on semantic loss, on form, or on function—with the proviso that such a framework must be able to accommodate gradient membership in any class. However, some other points arise from this consideration of modal auxiliaries which are similar to the issues raised by the argument over English auxiliaries. The question of the existence of auxiliaries can be divided into two related areas: whether, within a particular framework, one needs to posit the category of auxiliary verbs (or modals) in order to account for the behaviour of some verbs; and whether or not one can actually delimit a class of auxiliaries, regardless of whether or not the class is needed in the syntactic rules of a grammar. These two points may seem to be two sides of the same problem. If there is no need for the category, because all the rules needed to account for auxiliary verbs are also needed in the description of some main verbs, then there can be no characteristics which uniquely define the class of auxiliaries. Nevertheless, the two points are more independent of each other than this would suggest. Palmer (1979a; 1979b) realises that much of the argument about the status of auxiliaries rests on the theoretical framework used to describe the language. As he points out (1979b, 180), ‘If one adopts a model which must by its nature treat auxiliaries as main verbs, there is little point in arguing whether they are or are not.’ The arguments suggest that a generativist account of English does not need to refer to auxiliary verbs, and that a generativist account of Spanish does not need to refer to modals, as these verbs can be accounted for by rules which also account for main verbs. From this point of view one might want to say that auxiliaries or modals ‘do not exist’. On the other hand, in each language there is a group of verbs which present a common core of shared characteristics. The fact that they possess a combination of characteristics not shown by other groups makes them into a separate subgroup. To ignore this in a description is to miss an important fact about the language. The exponents of the ‘auxiliary approach’ and the exponents of the ‘main verb approach’ have largely been arguing at cross purposes. This is implicitly acknowledged by Huddleston (1984, 140–3) who recognises that the dispute is about the validity of an analysis in which two elements—an auxiliary and a main verb—make up a verb phrase, as compared with the validity of an analysis in which there are two main verbs and no unit which contains both these elements and nothing else: ‘There is no dispute (how could there be?) over the distinction between two subclasses of verbs, one containing verbs like be…, one containing verbs like keep’ (p. 143). Bolinger (1970, 573) makes a similar point in his review of Klein (1968). K, by showing exactly how modals differ from other verbs in their underlying grammar, has simply defined the class of modal verbs, in the same way that the class of transitive verbs is defined in terms of how they
The question of auxiliary status
35
fit into the underlying structure in relation to complement noun phrases. What K. has shown is that it is unnecessary to assume a syntactic FEATURE+M—that poder and deber can be handled as ordinary verbs, restricted by rules already needed in other parts of the grammar. By eliminating the feature, he may give the impression that he has eliminated the need to recognize the class. But…the feature from the standpoint of traditional grammar is a straw man: classes depend on criteria, not features…. We shall…need to continue to talk about modals in Spanish. Similarly with Hernanz’s (1982) claims. Hernanz may have succeeded in demonstrating that within the particular framework which she is using there is no need to include a class of auxiliary verbs. However, in her discussion of the matter she indicates that there is a class of verbs which have certain characteristics in common and which main verbs do not tend to share. This remains true even if she concludes from this that (p. 467) ‘los llamados “auxiliares” se comportan a todos los efectos como “principales’”. This study is not concerned with the specific requirements of any particular theoretical model. As far as this work is concerned, it is largely irrelevant whether or not someone produces a description of Spanish which suggests that auxiliaries ‘do not exist’. As Bolinger suggests, we would still need to be able to talk about them.
1.6 DIACHRONIC CONSIDERATIONS Some earlier treatments of auxiliary verbs, and in particular some of those based on the notion of semantic loss, may have involved an inadmissible confusion of synchronic and diachronic criteria. But in avoiding this confusion, we should not necessarily discount diachronic considerations altogether. Generativists have tended to adopt an entirely synchronic approach to linguistic description. Given that one of the principal aims of generative grammar is to account for the knowledge which a native speaker possesses of his or her own language—a knowledge which does not usually include historical information—this concern with synchrony is not surprising. To say that such accounts ignore diachronic factors is not to criticise them: their aims dictate that their approach must be purely synchronic. It is important to have a clear and accurate description of the state of the language at a particular time, uninfluenced by historical considerations. Not only is this an objective in its own right: knowledge of synchronic language states is also an important part of the study of language change. But if one wants to attempt to explain why certain aspects of the synchronic state are as they are, one may need to bring diachronic information into consideration. Various writers have already drawn attention to this phenomenon. Narbona Jiménez (1981, 176) expresses it in the following terms: Hemos de recordar que la caracterización precisa…de los verbos de un idioma es una pretension loable, siempre que no se olvide que una lengua no es un sistema fijado y cerrado, sino que constituye un instrumento de comunicación en constante evolución y transformación;…está permanentemente abierta a nuevas posibilidades expresivas, al mismo tiempo que sufre desgastes y pérdidas. No estamos abogando por una
Tener+Past participle
36
mezcla indiscriminada y anárquica de los enfoques sincrónico y diacrónico, sino que creemos…en la necesidad de acudir a la historia para una mejor comprensión del funcionamiento sincrónico de las lenguas. It should be made clear here that I am not arguing for the abandonment of synchronic studies or of attempts at categorial definition. Both synchronic and diachronic approaches to description are necessary, as are both attempts to define categories and a willingness to accept fuzzy boundaries and variation in use. Different approaches can be complementary rather than contradictory. As Lass (1976, 220–1) points out: There is no sense in which a non-static orientation produces ‘better’ analyses, or gets us closer to ‘reality’ or ‘the nature of language’ than a static, formalistic one. The ‘nature of language’ is neither a datum nor a discoverable entity: it seems to be a set of complementary aspects or structures, different clusters of which are uncovered or produced by different heuristics and procedures, prompted by different initial assumptions. All analytic strategies are informed by the linguist’s a priori stance, and this to a large extent determines what he ‘discovers’. The static/dynamic dichotomy is thus not a matter of ‘right’ vs ‘wrong’, but of interest, of the linguist’s choice of which of the complementary aspects of language he wants to deal with. But diachronic considerations play a special part in the study of auxiliary verbs in the Romance languages. Green (1987, 257–8) argues that while there are no syntactic reasons for separating auxiliaries from other verbs in Romance, as there are in English, the special treatment of auxiliaries can be justified on semantic and historical grounds: All of the items which in modern Romance are typically classified as ‘auxiliaries’ can be traced without difficulty to Latin etyma whose function was partially or totally different…. Since the items have been in continuous use and have evolved in ways which differentiate them not only from their etyma, but also from any other items of Latin, it is not unreasonable to claim that a new functional category has been created. Whether that category is properly labeled ‘auxiliary’ is ultimately a question for universal grammar: regardless of the answer, there remains a difference between Latin and Romance which requires explanation. 22 Ramat (1987) also comments on the historical dimension: A historical view of the category AUX helps to elucidate the gradual coming-into-being of auxiliaries and their evolution but cannot lead to a categorial definition. It also makes understandable why there are so different views as to what may be considered an example of AUX. The decision, however, depends on the theory the linguist has chosen, not on the historical development of particular examples. (p. 15) In both English and Spanish auxiliaries develop from main verbs. They lose certain characteristics and gain others. This is the case regardless of whether or not one chooses
The question of auxiliary status
37
to recognise a class of auxiliaries in a synchronic language state. For example, Hernanz (1982, 450–1) suggests that the grammaticalisation of haber is a purely diachronic process, and has no place in a synchronic description of Spanish. Given that in the particular case of haber the process has been completed and haber now has no lexical value, her view is almost certainly correct. She suggests that in cases where a verb may— unlike haber—have what would traditionally be considered as both grammaticalised and full verb functions, the various functions should be taken to indicate the presence of entirely separate verbs (see p. 15). So for Hernanz, ir as in Todavía me acuerdo el día en que se me empeñó en ir, con su hermana, a oírme aquel sermoncete, and ir as in Pepe va a conseguir el empleo (examples taken from Hernanz) are two separate verbs. In a purely synchronic description this may be an appropriate decision to take; but if one is looking at the language from a diachronic point of view one needs to be able to show the process by which one develops from the other (if this is indeed the case). A synchronic language state can provide clues as to the nature of this development. Even within a purely synchronic description one needs to be able to account for the ambiguity of sentences such as Viene a cobrar cada mes unas quince mil pesetas en concepto de indemnización (‘Each month he/she comes to collect about fifteen thousand pesetas compensation’/‘It amounts to collecting about fifteen thousand pesetas compensation each month’) and María ha vuelto a buscar el libro (‘Mary has gone back to look for the book’/‘Mary is looking for the book again’) (examples from Hernanz), where both a ‘grammaticalised’ and a ‘literal’ interpretation are possible, and—according to Hernanz—may even both be present. The study of these problems does not rely greatly on any specific notion of ‘auxiliary verb’, but on the idea of the diachronic change in function of a form (in the case of this study tener). It is certainly helpful if the tests used in classifying the different uses of tener can be thought of as tests for auxiliary status, as tener can then be considered in relation to other verbs and verb classes. Also, there is greater justification for using these particular tests if they can be applied with consistent results to other verbs as well: the tests are of greater use if they define a class of verbs than if they merely define the different uses of one verb. However, it does not matter if the tests do not produce a clear separation of the different groups, but suggest ‘full’ members and ‘partial’ members for each class. In summary, there are two types of justification for examining the nature of auxiliary verbs—one synchronic and one diachronic. In a synchronic examination of modern Spanish it is possible to identify a group of auxiliary verbs which share a common core of characteristics, although there is no hard and fast division between auxiliaries and main verbs. Tener+past participle is a particularly interesting case because, as we saw when applying the tests for auxiliary status in sections 1.2 and 1.3, it is apparently more of a hybrid or a transitional construction than many others which are usually classified as auxiliary constructions. Although a description of modern Spanish would not encompass diachronic considerations, once we begin to examine the history of the language, we find in the development of a particular group of verbs a second reason for considering auxiliaries apart from other verbs. Given the nature of this study, questions such as whether one considers the construction to be a sequence of main verbs, or whether one needs extra rules to account for these verbs, do not concern us. As a synchronic investigation, it considers the various
Tener+Past participle
38
different uses of a verb, which will require different analyses regardless of what theoretical framework one chooses. Among the points to be considered here are: what the possible values are for a construction, and how one decides what value it has in a particular context. As a diachronic study, it considers questions such as the relation between the various values of a construction; how one value may develop from another to such an extent that it can be considered a ‘separate’ value; and how a particular usage spreads through the language.
2 Tener +past participle in modern Castilian 2.1 OUTLINE OF STUDY This chapter looks specifically at usage of tener+past participle, discussing in more detail than in chapter 1 the possible uses of the construction. In order to do this it presents results obtained from a survey of usage of the construction by native speakers of peninsular Spanish. Then the various descriptive approaches described in chapter 1 are considered in the light of these results, and a different classificatory approach is discussed.
2.2 SURVEY: METHODOLOGY There has been much discussion on what constitutes the best way of collecting syntactic information. (For a summary, see Carden (1976). See also Labov (1972a); although his discussion is principally concerned with phonetics, it does include a more general discussion of the methodology of data collection and makes some points about the specific problems of syntax.) Carden compares and evaluates four different methods of data collection: forced choice questionnaires, in which informants are asked to give an immediate judgement on the acceptability of test sentences; semi-structured open-ended interviews; direct observation of performance; performance tasks, such as repetition of example sentences or performing operations on a stimulus sentence (e.g. making the sentence negative). After comparing the different methods, Carden concludes that none of these methods is fully satisfactory: Direct observation of performance, while potentially important as a means of validating other methodologies is impractical as a primary technique. Performance tasks seem to be even less reliable than evaluation tasks, and are difficult to adapt to the more interesting syntactic problems. Forcedchoice questionnaires are also difficult to construct, and have at best marginal reliability and very noisy data. Open-ended interviews seem to produce cleaner results, but are very time-consuming and may have bias problems. (p. 103) In view of the problems involved in the different methods of data collection, it was decided to rely on a compromise between forced-choice questionnaire and open-ended interview. Given the rarity of some of the uses of the construction, it would have been impossible to collect enough instances of it through direct observation. Tener is fairly restricted as to which participles it may appear with, and it would have been difficult to
Tener+Past participle
40
discover these restrictions without direct questioning of informants over key examples. Suitable performance tests would have been difficult to construct and, it would appear, are of limited value. The survey was based on a questionnaire which was compiled after a preliminary pilot study. In the final survey each informant was interviewed individually and asked to judge the grammaticality of various sentences containing examples of tener+past participle. Further questions’ were used to discover how particular examples would be interpreted, whether they were ambiguous in any way, what related interrogatives there were (in order to test the criteria used by Fontanella and others) and other grammatical information. Informants were encouraged to volunteer any extra information which they felt might be relevant. Using a combination of questionnaire and discussion provided a means of discovering not only whether or not a test sentence was acceptable, but also what was unacceptable about the ones which were rejected. In this way factors with no direct bearing on tener+past participle could be discounted. This was particularly important where context was concerned: discussion with the informant made it easier to discover cases where a sentence was syntactically acceptable but was rejected because the context (or lack of it) made the sentence seem odd in some way. Context posed some problems in compiling the list of sentences. There is a possibility that the way in which the construction is presented might lead to inaccurate results. 1 For example, it has been suggested that tener+past participle expresses either perfective aspect with present relevance, or an iterative aspect (see, for example, Spaulding 1958, 43; Fente et al. 1972, 43–4). If the construction is presented in such a way that these interpretations are not possible, or seem unlikely, then the sentence may be rejected because of aspectual incompatibility rather than because the construction itself is ungrammatical. Blansitt (1963, 54–5) points out that context and factors such as the presence or absence of temporal or frequentative adverbs can affect the acceptability of a construction. In his survey he notes that pairs of sentences such as Estoy trabajando~Lo que estoy haciendo es trabajar were immediately acceptable, but that pairs of the types Voy trabajando~Lo que voy haciendo es trabajar; Vengo trabajando~Lo que vengo haciendo es trabajar were accepted only after some hesitation. However, in particular contexts these constructions were accepted without hesitation: Voy trabajando cada vez con más gusto~Lo que voy haciendo cada vez con más gusto es trabajar; Vengo trabajando desde hace semanas~Lo que vengo haciendo desde hace semanas es trabajar. It is not that the construction itself is ungrammatical, but rather that lack of context makes it seem ‘odd’ in some way. Discussion with informants made it easier to spot this problem and to adjust sentences accordingly. A written questionnaire which asked only for ‘right or wrong’ judgements could not have done so. It should be made clear that this survey is concerned primarily with what is possible, not what is most common, nor the relative frequency of use by different speakers. While looking for differences in use between speakers, it makes no attempt to quantify such variation, nor to relate it to extralinguistic factors. 2 The purpose of the survey is to find out what is and is not grammatical for each speaker, and to attempt to discover the limits of grammaticality of tener+past participle. Sometimes an informant accepted that a particular example was possible but very uncommon. These examples have been classified as grammatical. There was more of a problem when an informant made a comment such as ‘People do say that, but I
Tener +past participle in modern Castilian
41
wouldn’t.’ On these occasions it was necessary to find out how acceptable the informant found the sentence, and whether they felt that they were ever likely to use it. If they were adamant that they never used it, then it was classified as ungrammatical for that speaker. Although I have tried to be as consistent as possible in my interpretation of such comments, I recognise that such an approach is to some extent subjective and liable to some inconsistency. A survey of this sort relies not only on the consistency of the interviewer but also on the reliability, honesty, and even the linguistic sophistication, of the informants. Sociolinguistic investigations have shown that people’s evaluation of their own speech differs from their actual speech. When relying on native-speaker judgements there seems to be no way round this problem. All informants were told that I was not interested in ‘textbook’ answers, but in what they genuinely felt was possible. Any judgements which seemed ‘odd’ or inconsistent, or which the informant had difficulty in deciding on were rechecked. The survey took place in April and May of 1987. Twelve informants were used. It may seem that twelve is a very small number if we are to gain an accurate overall picture of usage. However, for the type of study undertaken here, detail and comprehensiveness are more important than the number of informants surveyed. Many linguistic descriptions present a generalised account, and merge the usage of different speakers. If what one wants is an overview of the language as a whole, then this is probably the most practicable way to go about it. But this approach does have its problems. Individual patterns are swamped by the need to draw a general outline on a very broad scale. Descriptive grammars may list several variants of the same phenomenon, but give no indication as to whether each individual speaker uses one variant more or less consistently or freely uses all the various forms. For example Fente et al. (1972, 39) maintain that in contemporary spoken Spanish agreement of the participle with the direct object in the tener+past participle construction is vacillating. On one level this is true: although the participle almost always agrees with the direct object there are cases where it does not. But if one looks at usage in more detail one finds that for many speakers there is no such vacillation: they consistently make the agreement. Some other speakers usually make the agreement but have a few exceptions where the participle is always invariant. For some speakers there are a few cases where the participle may or may not agree, and in these limited cases there is free variation. It is not my intention to criticise descriptions such as that of Fente et al. In many cases the sort of overview which they provide is precisely what is needed. But there are times when a more detailed survey is appropriate. One of the questions addressed by this study is how changes filter through the language and appear to affect certain parts of the language before others. For this type of investigation a detailed knowledge of grammars of various speakers is needed. Once the situation for various individuals has been ascertained, these results can be compared, and an attempt can be made to find the principles which lie behind the variation between speakers. It may be that this variation can then be related to language change, with differences between speakers being attributed to different rates of change. It is for these reasons that the study concentrates on a few speakers rather than undertaking a large-scale survey. The twelve informants came from three different cities in Spain: Oviedo, Valladolid, and Zaragoza. Different locations were used to make sure that the results were not biased towards one particular geographical area. It was also
Tener+Past participle
42
hoped that if any regional variation was found it would provide additional material for work on comparative degrees of gram-maticalisation. In the event, it was found that the results from Oviedo differed considerably from those of Valladolid and Zaragoza, which did not differ significantly along regional lines. The difference in usage shown by the Oviedo speakers is due to regional factors. As these results betoken a very different system, rather than slight variation from ‘standard’ usage, they will be dealt with separately in chapter 5. The informants in all three locations were educated middle-class speakers. All had been educated at least to the level of a first university degree or were in the course of studying for their degree. Although it would have been interesting to investigate sociolinguistic class-based variation, this was beyond the scope of the survey, which was limited to the study of ‘standard’ educated speech. 3 The age of informants ranged from twenty to forty-eight.
2.3 DESCRIPTION OF QUESTIONNAIRE The sentences were presented to informants in random order. They were devised to provide examples of the construction in a range of different contexts, to see how this affects grammaticalisation and interpretation. In the survey, tener appears with the participles of various different types of verbs in similar syntactic contexts. Some of the verbs are consistent with the notion of possession and others are not. Verbs expressing different degrees of possession are included. For example tener dado would suggest that the subject of tener does not possess the object; tener prestado implies that the subject still possesses the object but does not have it at his disposal; tener perdido suggests that the subject does not have the object, but may in some contexts still be considered to own it. The same verb is included with both literal and figurative uses: e.g. prestar el coche, prestar ayuda; perder la cartera, perder kilos, perder la costumbre, perder la cabeza. Various different lexical values of tener are included—full possession, inalienable possession, etc.—to attempt to discover whether certain values are more prone to grammaticalisation than others. Syntactic differences as well as semantic differences have to be taken into account, and so there are examples of various different classes of verbs: intransitive, transitive, reflexive, etc. There are cases of tener+past participle with a sentential complement and with an infinitival complement. For example, prometer+noun, decir+noun are contrasted with prometer que…, decir que…; and proyectar+noun is contrasted with proyectar+infinitive. Sentences are included which test factors such as word order; separation of object and participle, or auxiliary and participle; agreement of the past participle with the direct object. In all the sentences tener occurs in the present tense. It was felt that the examination of which tenses of tener can occur in the construction was better left until some idea of the general scope of the construction had been obtained. The present tense of tener+past participle is also the most interesting of the tense combinations in that it raises questions about the use and development of the present perfect in Spanish.
Tener +past participle in modern Castilian
43
2.4 ANALYSIS OF SPEAKERS’ JUDGEMENTS Let us now look at the information gained from the questionnaire. At this stage I shall do little more than describe and summarise results. Once this has been done for the eight speakers from Valladolid and Zaragoza, I shall go on to discuss these results and draw some conclusions from them. The aim here is to outline major types of use which are significant in the development of the construction. The acceptability and interpretation of a particular example of tener+past participle may depend on a number of factors, not all of which are of major importance in the way the construction gains in acceptance or changes its interpretation. More detailed comments will be made where they are particularly relevant to an understanding of this development. 4 First the results obtained from two speakers will be considered. Then the results of other informants will be described inasmuch as they differ from those of these first two speakers. 2.4.1 Informants 1 and 2: outline
Examples which were accepted by the speakers can be divided into twomajor groups, within which various subdivisions can be made. Thepredominant interpretation of tener+past participle is that it refers to animperfective state. This first major category can be further broken down. One subset relates to inalienable possession. Both informants accepted sentences such as: (1) Tengo los ojos irritados (2) Tengo rota la pierna In this type of example tener+past participle refers purely to the current situation; there is no mention of prior causative action. The speaker states that his eyes are currently sore or his leg is broken. Sentence (1) is similar to sentence (3): (3) Tengo los ojos azules Both refer to a current state. There is a difference between them, in that in the situation described in (1) something happened to cause the state, while (3) denotes an inherent characteristic. But although in (1) there must pragmatically be a causative action, this action is not referred to in the grammatical construction of the sentence. Just as in (3) azules is an adjective, in (1) the participle irritados functions as an adjective. 5 A further indication that irritados is an adjectival participle rather than a verbal one is that it combines with muy rather than mucho: (1a)
Tengo los ojos muy irritados
(1b)
*Tengo los ojos irritados mucho
This criterion cannot always be applied. For example (2a) is not possible: (2a)
*Tengo muy rota la pierna
Tener+Past participle
44
This is because (2) is something which does not admit of degrees. The leg is either broken or it is not: it cannot be very broken. But although the impossibility of inserting muy does not necessarily indicate that the participle is verbal rather than adjectival, where the insertion is possible it does indicate that the participle is adjectival. The next set of examples also refer entirely to a current state. It includes sentences such as: (4) Este chico tiene preocupada a su madre (5) Las últimas noticias les tienen muy angustiados Although, like (1) and (2), these sentences have a purely ‘current state’ interpretation, they differ from the earlier examples. In sentences (1) and (2) the action causing the state is over. In (1) the speaker’s eyes are sore, but there is no indication that they are necessarily still being made sore. In (4) and (5) the action continues: the boy still worries his mother; the news still upsets the people. Here, tener+past participle can be replaced by the simple present: (4a)
Este chico preocupa a su madre
(5a)
Las últimas noticias les angustian
Fente et al. (1972, 44) point out that verbs which behave in this way are ones which express a mental state. In (6)–(9) we have a further set of possible sentences. (6) Mis padres me tienen prohibido que salga los sábados por la noche (7) Me tiene prometido el regalo (8) Tenemos proyectado ir a la playa este fin de semana (9) Tengo pedido el libro Sentences (6)–(9) differ from (4) and (5) in some respects. As was pointed out, in (4) and (5) tener+past participle can be replaced by the simple present. In (6)–(9) the informants offered as alternatives to tener +past participle the haber+past participle construction, not the simple present. Although the emphasis is still on a current state, there is some indication of the action which led up to that state. The reason why haber+past participle rather than the simple present is substituted is because of differences in the verbs used. In examples (4) and (5) both the action expressed by the verb and the participle are durative. In examples (6)–(9) although the participial construction expresses a durative resultant state the verbal action is punctual. In examples (4) and (5) there is not a resultant state so much as the continuation of the action expressed by the verb. In (6)–(9) a distinction can be drawn between the punctual action and the state resulting from it. This explains why in (4) and (5) tener+past participle can be replaced by the simple present, whereas in (6)–(9) it cannot. It also explains the difference in the way the examples behave with haber+past participle. We can replace tener+past participle with haber+past participle in (4) and (5): (4b)
Este chico ha preocupado a su madre
(5b)
Las últimas noticias les han angustiado
Tener +past participle in modern Castilian
45
But if we do so, the implication is that the situation no longer applies: that the boy has in the past worried his mother but no longer does so; that the news has upset the people but does not necessarily upset them now. This implication can of course be cancelled, e.g.: (4c)
Este chico ha preocupado a su madre y sigue dándole muchas preocupaciones
But as (4b) and (5b) stand by themselves the implication is one of past time reference. The use of haber+past participle implies that the action took place in the past. Because in (4) and (5) there is no resultant state but rather a continuation of the action, if this action is described as being in the past it is assumed that any continuing action is also in the past, and there is no separate resultant state which still applies. 6 This is not the case when tener+past participle is replaced by haber +past participle in (6)–(9). (6a)
Mis padres me han prohibido que salga los sábados por la noche
(7a)
Me ha prometido el regalo
(8a)
Hemos proyectado ir a la playa este fin de semana
(9a)
He pedido el libro
In these examples there is a distinction between past action and resultant state. And while the action is in the past, the resultant state may continue into the present. So why make any distinction between these groups of sentences when it appears that the difference is due to the particular verb used and not to the construction itself? If we make a distinction between these two uses of tener+past participle then we should also distinguish various uses of haber+past participle, for example, depending on the type of verb with which it appears. This seems unnecessary when the differences can be attributed to the aspect or the semantics of the verbs rather than to any difference in the construction itself. However, when describing a construction which is not completely grammaticalised, which does have considerable selectional restrictions, and which is limited to cooccurrence with certain types of verbs, it is worth at least pointing out that the construction can be used with different types of verbs. In this case tener can be used with the past participle of verbs which do not purely indicate a current state but which also imply the action which leads up to that state. Secondly, informants felt that there was a difference in interpretation between the two groups. Even given the difference in the behaviour of the two sets of verbs, it would still be possible for sentences (6)–(9) to refer only to the current situation. But the informants did feel that in these sentences there was at least some indication of the previous action: they were ‘more past’ in nature than the other sentences so far discussed. This view is supported by information offered by one of the two informants, who maintained that in certain cases, tener+past participle could be used to refer primarily to actions in the past which still applied in the present, rather than to states caused by previous actions. The question ¿Qué tienes hecho? could be used to refer to a series of events which had happened over a period of time which included the present moment (today, this year, a lifetime, etc.). This informant was able to find contexts in which (6)–(9) were possible answers to the question ¿Qué tienes hecho? but could not think of contexts which would make any of sentences (l)–(5) possible answers. The reason he gave was that (l)–(5) were all purely present and did not indicate any past
Tener+Past participle
46
action. 7 A further reason for pointing to this difference in interpretation is that it may provide an insight into the diachronic development of the construction, as it gradually widens its range of cooccurrence and comes to be used with different types of verbs. A slight shift in interpretation, linked to the aspect and semantics of the particular participle, may provide the base for further developments away from the primarily current state usage. This argument will be taken up more fully at a later stage. The various examples we have considered so far differ in some ways: in cases of inalienable possession and where the action is durative the construction refers exclusively to the current state; if the action is punctual there is a resultant state interpretation, with some indication of previous causative action. In some cases the subject of tener is also the agent of the participle, in other cases this is not so. But all these examples have in common the fact that they refer primarily to current durative states. It is this factor which separates this group from the second major group of examples of tener+past participle. In this second group tener +past participle is much more similar to haber+past participle. Possible sentences included: (10) Ya te tengo dicho no sé cuántas veces que no hagas eso (11) Se tienen contadas muchas historias el uno al otro (12) Tengo oído que mañana no va a haber clase This seems to represent the greatest degree of grammaticalisation of tener in the speech of these two informants, and the greatest move away from any original lexical value. One of the informants was doubtful about (12) and said that it sounded much less natural than (13) Tengo entendido que mañana no va a haber clase If there is a general tendency to ascribe a durative resultant state interpretation wherever possible, rather than an interpretation which focuses on punctual past action, then this informant’s judgements are unsurprising. Entender is stative, or covers a punctual action which marks the start of a durative state. Oír refers primarily to a punctual action; it contains less emphasis on a resultant state than do the verbs in sentences (6)–(9). Thus it is not so readily accepted. It seems that the imperfective, current state interpretation takes precedence over other possible interpretations. This is the interpretation which is ascribed wherever possible: for example, both informants accepted (14) Tengo el libro olvidado For both of them (14) implied that the speaker could not remember anything about the contents of the book. It did not mean, for example, that the speaker had forgotten to bring the book with him. The two informants agreed on the following results: (15) *Antes estudiaba francés, pero lo tengo dejado ya (16) Antes estudiaba francés, pero lo tengo abandonado ya It appears that dejado puts more emphasis on the action than does abandonado, which is frequently used adjectivally. Sentence (16) refers to the current situation: the speaker no longer bothers about French. Similarly with
Tener +past participle in modern Castilian
47
(17) *Tengo despertado al niño (18) Tengo despierto al niño Example (18), with the adjectival participle, referring to the state rather than the action, is possible. The verbal participle in (17) has too much emphasis on the past action and is unacceptable. 2.4.1.1 Informants 1 and 2: syntactic considerations For these speakers the tener+past participle construction can be used only where there is a direct object, with which the participle agrees (or occasionally where there is an implicit direct object—see below). The direct object is usually a noun. Both the order ‘tener+participle+ noun’ and ‘tener+noun+participle’ are possible. In general, the interpretation of the sentence is not affected by whether the participle comes before or after the noun. Exceptions will be discussed in section 2.5.1. ‘Tener+past participle+infinitive’ and ‘tener+past participle+ sentential complement introduced by que’ are also found; e.g. (19) Tenemos pensado ir al cine esta tarde (20) Ya te tengo dicho que no hagas eso For one of the speakers tener+past participle must always have an overt direct object, but the other speaker would accept examples with no direct object provided that a direct object was implicitly understood. For example, he accepted (21) Tengo escrito sobre ese asunto For him it implied (21a)
Tengo algo escrito sobre ese asunto
Tener+past participle is not generally possible with intransitive verbs. 2.4.1.2 Informants 1 and 2: reflexive constructions Certain reflexive constructions are possible, and others are not. The following judgements were given by both speakers. (22) *El niño se tiene lavado (23) *El niño se tiene lavadas las manos (24) * Jorge se tiene quedado en casa (25) *Se tienen mirados con odio el uno al otro (26) *Se tienen peleados el uno con el otro (27) Se tienen dicho lo mismo el uno al otro muchas veces (28) Se tiene mucho dinero invertido en ese negocio ‘True’ reflexives as in (22) are not possible. Neither are indirect object reflexives (23) or ‘inherent’ reflexives (24). Reciprocal reflexives are possible where the reflexive pronoun is an indirect object (27), but not where it functions as a direct object (25), (26). Impersonal reflexives are accepted (28).
Tener+Past participle
48
2.4.1.3 Informants 1 and 2: aspectual constraints The following results were obtained from both speakers: (29) Tengo barrida la casa (30) *Tengo barrida la casa varias veces (31) Tengo prestado el coche (32) *Tengo prestado el coche varias veces One of the values of tener+past participle which is most generally recognised is that of repeated action (see, for example, Criado de Val 1948, 118–19; Fente et al. 1972, 43–4). So at first sight it might seem odd that these sentences with the frequentative adverbial phrase muchas veces are unacceptable. However, for these two speakers, tener+past participle principally indicates a durative state. Thus (29) implies that (having been swept) the house is now clean. Similarly (31) indicates that the car has been lent to someone and is still lent. For these informants the durative state interpretation is so dominant that it cannot be cancelled, even explicitly by use of phrases such as muchas veces. One of the informants did allow (33) Tengo pedido el libro varias veces This seems to contradict the above explanation. It may be that this is indicative of a first move towards accepting this type of construction. Example (33) is also accepted by other speakers who do not accept examples such as (30) and (32). It regularly seems to be one of the first of this type of example to be accepted. The regularity of this occurrence is presumably not coincidental. Although it is impossible to draw definitive conclusions from one example, there is a possible explanation as to why (33) is permissible while (30) and (32) are not. In (30) and (32) the repetition of the action implicitly cancels the resultant state. In order for someone to lend their car a second time they must presumably have got it back in their possession after lending it the first time. The fact that they can lend it again automatically cancels the resultant state caused by lending it the first time. Similarly with (30): presumably someone does not clean the house again until it is no longer still clean as a result of the previous cleaning. But with (33) the repetition of the action does not cancel the resultant state. However many times the book is reordered it can remain on order throughout. It seems that this is the first step towards being able to cancel the imperfective state interpretation. This does not mean that those who suggest that tener+past participle frequently indicates repeated action are mistaken. Where the durative state interpretation is unlikely, there is likely to be an iterative aspect. See, for example, sentences (10) and (11) above. 8 If one accepts that where a current durative state interpretation is not possible tener+past participle is likely to indicate repeated action, this explains the following results: (34) Me tiene dicho repetidas veces que no piensa casarse con íl (35) *Me tienen dicho que ayer estuvieron en casa (36) Mis padres me tienen prohibido que salga los sábados por la noche (37) *Mis padres me tienen prohibido que salga esta noche
Tener +past participle in modern Castilian
49
Example (34) is grammatical because it indicates repeated action; (35) is ungrammatical because context makes repeated action highly unlikely. Sentence (36) does not necessarily indicate repeated action, but does indicate a durative state which applies on several occasions. The context in (37) implies prohibition on an individual occasion, which conflicts with both iterative and durative interpretations. 2.4.2 Informant 3 The judgements given by this speaker are very similar to those given by the first two speakers, but do exhibit some important differences. The main outline is the same in terms of interpretation and constraints imposed by syntax and the aspect of the participle, but there was a tendency for this speaker to move further away from the current state interpretation. For example, like the previous two informants, this speaker accepted (38) Tengo vistos a esos hombres However, for the previous speakers this was purely a set phrase equivalent to Los tengo muy vistos (i.e. I know those men very well and I know all about them). For this third speaker this construction could have this interpretation, but it could also appear in a context such as (39) Esos hombres, yo los tengo vistos de antes This sentence did not imply Los tengo muy vistos, but rather that the speaker has actually seen the men somewhere before and so they seem familiar to him. This informant also accepted sentences such as (40) Tengo vistas muchas películas suyas (41) Tengo vistas muchas cosas raras Here the emphasis is more on the iterative past action than on the resultant state. This type of sentence was not accepted by the first two informants discussed. Informant 3 accepted (42) Tengo recorridos veinte kilómetros One of the first two informants accepted this sentence but the other did not. Sentence (42) is only appropriate if the speaker plans to continue travelling; it is not appropriate, for example, on the completion of a walk. But the third informant would also accept (43) Tengo recorrida mucha Francia de cuando yo era militar As with (39), sentence (43) still has a resultant state interpretation and the past action is not referred to directly; but with the phrase de cuando yo era militar the action is situated further back in the past. This sort of example was not accepted by either of the first two informants. It seems that although this speaker’s use of tener+past participle is much the same as that of the first two speakers, there is a slight decrease in the importance of the current state interpretation. This trend is developed a little further in the grammar of informant 4.
Tener+Past participle
50
2.4.3 Informant 4 Again, this informant’s judgements do not differ vastly from those of the informants already discussed. He accepted sentences (38)–(43), and also accepted some sentences of this type which were deemed ungrammatical by the previous informant. For example (44) Tiene conocidos muchos países durante sus viajes This is a similar case to sentence (43). The presence of durante sus viajes puts the emphasis more on the moment of the action than on any possible resultant state. Also, this speaker accepted sentences such as (45) Tiene viajado mucho por el extranjero This is a first move towards acceptance of tener+past participle with intransitive verbs. Various points should be noted. Firstly, that this type of example with an intransitive verb is grammatical only where mucho is inserted; without this quantitative adverb the sentence is ungrammatical. Secondly, mucho must imply quantity rather than frequency. Where it must imply muchas veces the sentence is not grammatical: (46) *Tienen ido mucho a Madrid Thirdly, at this stage, mucho appears to be the only adverb which makes sentences with intransitive verbs grammatical: (47) *Tiene vivido mucho tiempo en España, y por eso habla tan bien el español (48) *Tengo comido muchas veces en ese restaurante There are some cases where even with mucho the sentence is ungrammatical: (49) *Tenemos hablado mucho sobre ese asunto (50) *Tengo pensado mucho en ese problema There seems to be no immediately obvious answer as to why (49) and (50) are unacceptable whereas (45) is acceptable. It may just be that at this early stage of grammaticalisation, only certain examples have grammaticalised. Or judgements on (49) and (50) may be influenced by the fact that in each case there is a very similar transitive construction which is grammatical, and to which (49) and (50) were immediately changed: (51) Tenemos hablado ese asunto (52) Tenemos pensado ese problema
Tener +past participle in modern Castilian
51
2.4.4 Informant 5 As compared with informant 4, there is further grammaticalisation. In addition to accepting the various types of examples accepted by the informants discussed above, he accepted more examples containing intransitive verbs and quantitative adverbs: (47) Tiene vivido mucho tiempo en España y por eso habla tan bien el español (49) Tenemos hablado mucho sobre ese asunto (53) Tienen viajado muchas veces por el extranjero Other than this, the degree of grammaticalisation appears to be much as for informant 4. 2.4.5 Informant 6 This is the first informant to be discussed in whose grammar the durative current state or resultant state can be explicitly cancelled, and an iterative interpretation substituted: (54) Tengo perdida la cartera varias veces (55) Tengo castigado al niño muchas veces (56) Les tengo prestado el coche muchas veces Without the phrase muchas veces (54)–(56) each refers to a current state: the wallet is lost at the moment; the child is currently being punished; the car is lent to someone at the moment. But the phrase muchas veces cancels this interpretation. Example (54) states that the speaker has lost the wallet several times, but it does not imply that the wallet is lost at the moment of utterance. Likewise with (55): there is nothing to indicate whether or not the child is currently being punished. And similarly (56) does not imply that the car is lent at the moment of utterance. 2.4.6 Informant 7 This informant showed a greater tolerance of the construction. As well as accepting sentences such as (54)–(56), where the current state interpretation has been cancelled, she also accepted examples such as (57) Tengo despertado al niño un montón de veces For all previous speakers, where a verb has two participles, one verbal and one adjectival, the adjectival participle is the only one which can be used in conjunction with tener. For this speaker the verbal participle can be used provided that there is an explicit indication of repeated action. Without the phrase un montón de veces (57) was immediately changed to (58) Tengo despierto al niño Example (58) indicates a current durative state, not a past action. For other speakers, (58) was the only possibility; not only was the verbal participle not acceptable, but also, unsurprisingly, the adjectival participle could not be used in any way that indicated past action.
Tener+Past participle
52
This speaker also showed increased tolerance of tener+past participle in reflexive constructions. Here are examples of the judgements she gave: (59) *El niño se tiene lavado solo muchas veces (60) ?Jorge se tiene quedado en casa cantidad de veces (61) ¡Cuántas veces me tengo levantado por la noche! (62) ?Se tiene lavado el pelo muchas veces (63) Me tengo torcida la muñeca varias veces (64) Se tiene invertido mucho dinero en ese negocio (65) Se tienen escritas varias cartas (66) *Se tienen peleados muchas veces Examples (60)–(63) are grammatical only where they contain a frequentative adverbial phrase indicating repeated action rather than current state. This is the only speaker who allows examples of tener+ past participle with ‘true’ reflexives (61), ‘inherent’ reflexives (60), or indirect object reflexives with objects of inalienable possession (62), (63). The results are far from conclusive and indicate some uncertainty by the speaker. But they do at least indicate a move towards greater acceptance of reflexive constructions. The results with impersonal and reciprocal reflexives are the same as for previous speakers. 2.4.7 Informant 8 So far, although there have been some apparent inconsistencies in results, all the speakers considered have fitted into the same overall pattern of development. However, there was one informant whose results did not seem to fit into the same pattern. Some categories are clearly acceptable, such as those illustrated by sentences (l)–(5) (inalienable possession and the verbs expressing mental states). But these categories are the only ones which were regularly accepted. 9 Some categories are clearly impossible: intransitive verbs are not possible, even where there is a quantitative adverb; tener+past participle is not possible with any reflexive construction, including the impersonal and reciprocal reflexives accepted by the other speakers. But even discounting these examples, there remain a fair number of examples whose ungrammaticality has not been accounted for. Some of the speaker’s judgements are surprising, both because they show a rejection of constructions which are fairly common in Spanish, and because they do not seem to be internally consistent. For example: (67) Tenemos pensada la excursion para la semana que viene (68) *Tenemos pensado ir al cine (69) *Tiene proyectada una máquina que haga el trabajo de veinte hombres (70) Te tengo dicho que no hagas eso (71) *Me tiene dicho repetidas veces que no piensa casarse con él No other speaker accepts ‘tener pensado+noun’ but not ‘tener pensado +infinitive’. It may be that for this speaker ‘tener+past participle+ infinitive’ is always impossible. This would be more plausible if ‘tener+ past participle+noun’ were the only possible combination; but ‘tener +past participle+que’ is also acceptable in some cases. Also, it is odd that (67) should be accepted when (69) is not. Similarly, it is odd that (70) is
Tener +past participle in modern Castilian
53
grammatical and (71) is not: both contain tener dicho que with an iterative aspect; the only difference is that one is followed by the indicative and one by the subjunctive. No other speaker accepts one but not the other. It seems that in this informant’s speech, tener+past participle is much less grammaticalised than it is in the speech of any of the other informants. It may be that these results reflect the sort of fluctuation which one might expect in the very early stages of the development of a construction. Certain examples of the construction are picked up and assimilated into the speaker’s grammar, but at this stage they remain isolated examples. 10 They are almost like set phrases, and have not yet led to the formation of any analogical constructions or to the restructuring of the grammar.
2.5 COMPARISON OF RESULTS Having looked at each speaker individually, and plotted a general development of increased tolerance of the construction, it is now time to compare the judgements. 2.5.1 Word order In general the grammaticality and interpretation of a sentence are not affected by the relative ordering of the past participle and a nominal direct object. Informant 8 never differentiated in interpretation between examples with preposed and postposed participles; but for all other speakers there were some cases where the position of the participle did make a difference. Although the participle may appear before or after the noun, there was a tendency to prepose it where the construction had a more grammaticalised use. Postposition tended to lead to a more lexical interpretation: for example, all informants apart from informant 8 accepted both (72) and (73): (72) Tengo prestado el coche (73) Tengo el coche prestado But there is a difference between the two sentences. With a preposed participle the interpretation is unequivocally that ‘My car is lent to someone else’; with a postposed participle there is some ambiguity: (73) might mean either the same as (72) or ‘I have a car which has been lent to me.’ All speakers who accepted Tengo comida mucha carne agreed that Tengo mucha carne comida is ungrammatical unless it means, somewhat implausibly, ‘My flesh has been eaten away.’ Again, while the preposed participle leads to a more ‘auxiliary’ interpretation for tener, a postposed participle imposes a more literal possessive interpretation. Another example where the ordering of noun and participle is significant is in the pair of sentences (74) Nos tiene contadas muchas historias (75) */?Nos tiene muchas historias contadas All speakers who accepted (74) agreed that (75) was either ungrammatical or at least much clumsier than (74). If we assume that the order ‘tener+past participle+direct object’
Tener+Past participle
54
is more grammaticalised than ‘tener+direct object+past participle’ it is not surprising that (74) is preferred to (75) in a case such as this, where the interpretation is more or less Nos ha contado muchas historias and tener+past participle is very similar to haber+past participle. It is generally assumed that the greater the cohesion between the elements of a periphrasis the greater the degree of grammaticalisation (see the summary of Lehmann’s work, section 2.8). Thus there might be a tendency to regard a preposed participle as indicative of greater grammaticalisation than a postposed participle where the two elements of the construction are separated. 2.5.2 Agreement of direct object and past participle For five of the eight speakers, whatever the value of the construction and whatever the word order, the past participle must agree with the direct object. Even for the other three speakers there were only very occasional examples of a participle not agreeing with the object. Macpherson (1967) notes that in the development of haber+past participle, nonagreement first tends to occur where one or more of the following conditions are met: the past participle is immediately adjacent to haber, especially when the participle precedes the direct object; the past participle is separated from the direct object by one or more grammatical elements; the past participle is that of a verb whose meaning is incompatible with the notion of possession (e.g. dar). However, although examples of these conditions were included in the questionnaire, so few examples of non-agreement were found that it is not possible to draw any firm conclusions as to whether a similar pattern is to be found for tener+past participle. Informant 3 accepted both (76) No me creo las cosas que me tienen dichas (77) No me creo las cosas que me tienen dicho However, he was not totally convinced of the acceptability of (77). Where the participle was not in a relative clause, it showed obligatory agreement: (78) Me tienen dichas muchas cosas (79) *Me tienen dicho muchas cosas This would support the view that cases where the participle is separated from the direct object are more likely to exhibit non-agreement. However, there was only one example of this, as the other two speakers who showed some cases of non-agreement accepted only (76) and not (77). This was the only example of non-agreement from informant 3. All other cases came from the two speakers with the greatest tolerance of tener+past participle. Informant 6 accepted both (80) Durante estos últimos años se tienen construídas muchas fábricas por esta zona (81) Durante estos últimos años se tienen construído muchas fábricas por esta zona Without the phrase durante estos últimos años only an agreeing participle was possible. This phrase puts the emphasis more on the past action and is an indication of a move
Tener +past participle in modern Castilian
55
away from the current state interpretation (cf. the comments on informant 3’s acceptance of example (43)). This provides us with a very tentative indication that lack of agreement tends to arise in the examples which are semantically furthest removed from lexical tener. With this informant there was only one more case of non-agreement. The relevant results are as follows: (82) Tengo vistos a los hombres (83) Tengo vistos los hombres (84) Tengo visto a los hombres (85) *Tengo visto los hombres (86) Tengo a los hombres vistos (87) Tengo los hombres vistos (88) *Tengo a los hombres visto Here agreement and position of the participle are linked to interpretation. Where there is an agreeing participle, the interpretation is Los tengo muy vistos (i.e. ‘I know all about them’). Where the participle does not agree the interpretation is similar to Los he visto. Where there is a postposed participle only the first interpretation is possible, indicating that the sequence ‘tener+past participle+direct object’ is more liable than ‘tener+direct object+past participle’ to receive a perfect tense rather than a lexical interpretation. There is a further indication that the non-agreeing participle is part of a more grammaticalised construction than the agreeing one is. In section 1.3 we saw how Fontanella and others used the cómo and qué questions to test for auxiliary status. When asked about related interrogatives for this set of sentences the informant stated that the question ¿Cómo tienes a los hombres? (indicating a nongrammaticalised construction) required the participle vistos; the question ¿Qué tienes hecho? (indicating an auxiliary construction) required the answer Tengo visto a los hombres, with a non-agreeing participle. Note that even when preposed, a non-agreeing participle is possible only where there is a ‘personal a’. However, although the ‘personal a’ may affect agreement to some extent, its presence does not automatically enable a non-agreeing participle to be used: (89) *Tiene conocido a muchas chicas durante sus viajes Informant 7 gave similar results to informant 6 over the Tengo visto a los hombres set of examples. The only other example of a non-agreeing participle she provided was in (90) ¡Cuántas veces me tengo levantado por la noche! The participle remained levantado even if the subject was a woman. The fact that this is one of the more obviously grammaticalised examples, with emphasis on the action not the resultant state, may influence agreement. The examples of non-agreement are so few and far between that very little can be concluded from them. Although they are not conclusive enough to offer any firm evidence that agreement in the development of tener+past participle behaves in the same way as Macpherson’s account of haber+past participle, they certainly do not contradict his findings. It is worth pointing out that the few cases of non-agreement that occur appear among the more grammaticalised examples. Firstly, they occur in the judgements of the speakers with the highest tolerance of tener+ past participle (apart from the one rather tentative
Tener+Past participle
56
judgement from informant 3). Secondly, they occur in examples where tener+past participle has a more grammaticalised interpretation. This is seen most clearly in the set of examples (82)–(88). This would suggest that lack of agreement is linked to grammaticalisation. 11 2.5.3 Auxiliary and participle: identity of subject Where there is purely an imperfective state there need be no relation between the subject of tener and the agent of the participle. As mentioned earlier, Tengo rota la pierna does not imply that the subject broke his own leg. If there is a resultant state with some indication of the causative action there is generally held to be a relation between the two, even if they are not identical: the subject of tener must in some way be responsible for the action even if he did not carry it out himself. But if the emphasis is to move from resultant state to causative action then the two must be identical. If the interpretation moves from (a) ‘I have a letter which is written’ to (b) ‘I have written a letter’, then in (a) the subject must be the person doing the writing. Vincent (1982) considers this link as it applies to the development of habere+past participle, and suggests that tener+past participle is now following a similar line of development. Vincent’s account of the grammaticalistion of habere+past participle makes use of Case Grammar, and analyses the construction in its early stages in the following terms:
In this example, according to Vincent, the LOCATIVE-subject of habere is expressed in the inflection, but the AGENT-subject of collocare is not expressed at all, allowing for two different interpretations of the sentence, depending on whether or not the investors of the money are taken to be the people who currently possess it. Vincent describes the NEUTRAL as the ‘hinge’ between habere and the accompanying verb in participial form. Although the LOCATIVE of habere and the AGENT of the participle are not necessarily identical, if in most cases circumstances dictate the identification of these two elements, this habitual identification may become an obligatory one. Thus ‘habere no longer selects its own LOCATIVE but fills that position in the structure by promotion of the unexpressed AGENT of the verb to which it is attached by the NEUTRAL “hinge”’ (Vincent 1982, 84). A similar analysis can be performed for modern Spanish tener. If one considers a similar Spanish sentence—Tienen mucho dinero invertido en ese negocio—one can analyse it as One can assume that, as was the case with habere, the frequent identification of the LOCATIVE of tener with the AGENT of the verb in participial form gradually leads to
Tener +past participle in modern Castilian
57
the obligatory identification of the two, and perhaps from there to the grammaticalisation of the construction as a compound perfect form.
Vincent claims that an account such as his makes certain predictions about the way in which the grammaticalised construction spreads through different classes of verbs in the language. One of the main predictions is that the first verbs to appear in the grammaticalised construction will be those verbs where the LOCATIVE-subject of habere and the AGENT-subject of the participle must automatically be identical. In this category are verbs with what Vincent terms EXPERIENCER-subjects. In defining EXPERIENCER Vincent follows Fillmore (1971, 42), who states that ‘Where there is a genuine psychological event or mental state verb, we have the Experiencer.’ It applies to subjects of verbs such as know, believe, think, persuade, etc. Examples of tener+past participle containing this type of verb are among those most readily and most widely accepted by informants. But it seems that there are now other examples which imply identity of LOCATIVE and AGENT. And it is those cases where tener+past participle has more of a past interpretation, concentrating on the action, that this identity is most evident. For example, there may or may not be identity of LOCATIVE and AGENT in (91) Tengo castigado al niño There are also two interpretations of (92) Tengo castigado al niño muchas veces If it refers to a habitual state, again there may or may not be identity of the two roles. But if it has a ‘past’ interpretation, then the two must automatically be identical. To look at another example: unless there is evidence to the contrary, identity is assumed in a sentence such as (93) Tengo perdida la cartera But the identity is felt more sharply in (94) Tengo perdida la cartera varias veces Resultant state involves both the action leading to the state (or perhaps more precisely the beginning of the state or moment of change) and the state resulting from this. The question of who performed the action may not be particularly relevant. The important point is that the subject of tener has something in a particular state, regardless of whether or not the subject performed the action which led to this state. Given that the question of causative action is also considered, there must be some link between the subject of tener and the agent of the participle—i.e. between the ‘maintainer’ of the state and the cause of
Tener+Past participle
58
the state—but while this link may lead one to assume that the subject and agent are identical, their identity is not especially important and may not even be considered. Where attention focuses on the action, the resultant state element is less important. If one is no longer concerned with the resultant state, then the question of who is responsible for maintaining the state is irrelevant. Given that Spanish automatically expresses the category of ‘person’ in finite verb forms, then person is indicated in the form of tener, but not in the accompanying participle. But in most cases where tener+past participle refers to the action, it is the opposite which is wanted. The subject of tener—the person involved in the ‘resultant state’ value—is not needed; the unexpressed agent of the participle—the person responsible for the action or change of state—is more important. This problem is resolved if the overt subject of tener is always the same as the unexpressed agent of the participle. In these cases the salient person can be expressed by automatic identity of subject and agent. Where tener+ past participle is imperfective and there is less focus on the causative action and more on the resultant state, the subject of tener retains its own identity, which may or may not be the same as that of the participle. Vincent’s analysis also accounts for the fact that ‘two-place’ verbs (i.e. ones that require two noun phrases or arguments) will grammaticalise before ‘one-place’ verbs. With tener, as we have seen, the first verbs to grammaticalise are transitive (‘two-place’) verbs. But some intransitive verbs are accepted in the construction before all transitive verbs have been accepted. Given an analysis such as Vincent’s, in which there is identity between the LOCATIVE-subject of habere and the AGENT-subject of the non-finite verb, with a NEUTRAL-object as a necessary ‘hinge’, then only transitive verbs, which have a NEUTRAL-object, will occur in the construction. Intransitive verbs, with no NEUTRALobject, will not fulfil the syntactic requirements of the construction. Applying this reasoning to tener, one can analyse a sentence such as Tengo escritas muchas cartas as follows:
And one can analyse a sentence such as Tengo escrito mucho as follows:
Tener +past participle in modern Castilian
59
Here mucho functions as the NEUTRAL-object ‘hinge’. Although a sentence such as Tienen viajado mucho cannot be analysed in the same way, and mucho here is not a NEUTRAL-object, the formal similarity of sentences of the type Tengo escrito mucho, with a transitive verb, and the type Tienen viajado mucho, with an intransitive verb, might help to account for the spread of the construction to intransitive verbs. In the description of individual informants’ judgements we saw that mucho was the first quantitative adverb to appear with intransitive verbs in the tener+past participle construction. But then other adverbial phrases such as muchas veces and mucho tiempo also start to occur. This suggests a further slight move away from usage with transitive verbs only. Although there may be a formal similarity between Tengo escrito mucho and Tengo viajado mucho, there is no such similarity when muchas veces or mucho tiempo are used, as these cannot be used to form transitive constructions. Presumably, once usage of the tener+past participle construction spreads to include intransitive verbs with mucho, it then spreads further, by analogy, to include intransitive verbs with other quantitative adverbial phrases. Use with intransitive verbs is still very limited. Even where mucho is used, not all verbs are accepted: for example, all informants rejected (95) *Tengo ido mucho a Madrid (96) *Tengo estado mucho en Madrid Barrera-Vidal (1972, 138) states that any intransitive verb which can be used with quantitative adverbs can occur in the tener+past participle construction. According to Barrera-Vidal, the only verbs which cannot occur are verbs such as ser and estar which do not combine with these adverbs. However, although it seems that the first move towards the incorporation of intransitive verbs into the construction is the occurrence of tener+intransitive verb+mucho (where mucho implies quantity rather than frequency), later intransitive verbs can occur with phrases such as muchas veces. But even then it seems that this usage is limited to intransitive verbs which occur with the ‘quantitative mucho’ rather than the ‘frequentative mucho’. So one finds results such as: (97) Tienen viajado mucho por el extranjero (98) *Tienen ido mucho a Madrid (99) Tienen viajado muchas veces por el extranjero (100) *Tienen ido muchas veces a Madrid It is not wholly accurate to say, as Barrera-Vidal does, that verbs such as estar do not appear with mucho. The important point seems to be that if they do so then mucho implies muchas veces, and this is the factor which prevents them appearing in the tener+past participle construction. 12 2.5.4 Verb classes and possession There are further restrictions on the use of this construction. While certain verbs are readily accepted in the tener+past participle construction, others are not. For example, all eight speakers accepted (101) Tengo irritados los ojos
Tener+Past participle
60
(102) Mis padres me tienen prohibidas muchas cosas (103) El bedel tiene cerrada la puerta (104) Tenemos pensada la excursion para la semana que viene (105) Tiene mucho dinero invertido en ese negocio But only four speakers accepted (106) Tengo comida mucha carne And only one speaker accepted (107) Tiene dado el coche a Juan To some extent it is possible to draw up a hierarchy of verbs in order of acceptability. Examples of speakers’ judgements are shown in Table 2.1.
Table 2.1 Examples of informant judgements, ordered according to degree of acceptability Informants Example
8
1
2
3
4
5
6
7
Tengo irritados los ojos
√
√
√
√
√
√
√
√
Ese chico tiene preocupada a su madre
√
√
√
√
√
√
√
√
Mis padres me tienen prohibidas muchas cosas
√
√
√
√
√
√
√
√
Tengo escritas varias cartas
√
√
√
√
√
√
√
√
Tengo pedido el libro
√
√
√
√
√
√
√
√
Tengo olvidado el libro
?
√
√
√
√
√
√
√
Tengo recorridos veinte kilómetros
x
x
√
√
√
√
√
√
Tengo perdida la cartera
x
x
√
x
√
√
√
√
Tengo comida mucha carne
x
x
√
x
x
√
√
√
Tengo enviada la carta
x
x
x
x
√
x
√
√
Tiene dado el coche a Juan
x
x
x
x
x
x
x
√
Notes: √=accepted; x=not accepted; ?=uncertain. Informant 8 has been placed first in the table because, although she was the last informant to be discussed, her results show the least grammaticalisation of tener+past participle. In the table the informants are approximately ordered according to the degree of grammaticalisation shown by their results.
Although there are some exceptions, in general the greater the degree of grammaticalisation of tener+past participle the greater the tolerance of different verbs in the construction. So it seems that grammaticalisation is not just linked to tolerance of syntactic phenomena (use with reflexives, with intransitive verbs, etc.) or to aspectual factors such as the possibility of replacing a resultant state by an iterative aspect; it is also linked to the semantics of particular verbs which may appear in the construction.
Tener +past participle in modern Castilian
61
What factors might influence the likelihood of a particular verb appearing in the tener+past participle construction? A major candidate is possession: examples where possession is explicitly ruled out are less readily accepted. It seems that the more obvious the clash between possession and the semantics of the past participle the less acceptable the sentence. So tener dado was accepted by only one speaker; it was not accepted by the other informants even if dar was thought of as almost equivalent to prestar, and one assumed that the object would be returned. However, in many widely accepted examples of tener+past participle the notion of possession is largely absent. Although speakers tend to reject sentences involving a direct contradiction of possession, interpretation of a sentence is not greatly affected by whether or not the subject of tener actually possesses the object. For example, the grammaticality of Tengo escritas varias cartas is not affected by whether or not the speaker still has the letters. All speakers accepted Tengo pedido el libro even though this sentence implies that the speaker has not yet got the book. Half the speakers accepted Tengo comida mucha carne, which implies that the meat no longer even exists. Tengo enviada la carta and Tiene dado el coche a Juan both imply that the subject no longer possesses the object. But three informants accepted the first of these sentences while only one accepted the second. In an example such as Mis padres me tienen prohibidas muchas cosas the question of possession does not even arise. Perhaps the major factor in the grammaticalisation of tener is not the notion of tener as poseer but rather of tener as mantener. Thus in the following sentences the idea of maintaining something in a particular state is more important than possession. (108) Tiene preocupada a su madre (109) Tengo la casa barrida (110) Tengo prestado el coche (111) Tengo pedido el libro In the section on lexical values of tener (see section 1.1.2) it was suggested that it was not easy to select a single value for tener, and in particular that possession was not the dominant factor. But all its values are aspectually imperfective. To quote Navas Ruiz (1963, 98–9): Prescindiendo del valor de considerar como…son tres fundamentalmente sus significados de atributivos: posesión, duración, relación. No es posible fijar una frontera segura entre los tres, señalar cuándo comienza uno y cuándo acaba otro. Pero hay algo que los une, que se da en los tres y que, en definitiva, permite clasificar a tener entre los atributivos de permanencia: el carácter imperfectivo del verbo que tiñe en esta idea, con mayor o menor nitidez, sus tres matices. It is not really surprising that it should be this imperfective value, persisting throughout the various meanings of tener, which appears as the central value for lexical tener and proves to be more important than the more restricted possessive interpretation. Roca Pons (1971, 19 and 22) puts forward a copular function for tener. He suggests that the transitive verb tener corresponds to some extent to the intransitive estar. Siempre
Tener+Past participle
62
tiene la casa abandonada is equivalent to Su casa siempre está abandonada. This copular use does not imply possession: ‘cuando digo, por ejemplo, tengo a mi padre enfermo, el uso de tener está justificada no por expresar ninguna idea de posesión, sino simplemente por el hecho de que el padre a quien se alude lo es repecto a mí’ (p. 19). Thus tener essentially expresses an imperfective state, and at the same time implies some relation between the subject and object of tener. This does not mean that possession plays no part in grammaticalisation. But if one takes the mantener or the copular value of tener as central, why should one then need to consider its possessive value? Before attempting to answer this, let us consider another question: how does one explain the fact that where possession is explicitly ruled out the sentence is frequently unacceptable? The answer to both these questions lies in the fact that the degree of possession entailed by a construction may help to determine the relative emphasis placed on resultant state and on previous action. In grammars where tener+past participle has not grammaticalised very far, it seems that where there is a nominal direct object, the subject must be keeping that object in a particular state, even if it is not in the subject’s actual possession. However, in order to keep something in a state one must have it in one’s charge in some way. Let us consider how this might work. We will take as an example (112) Tengo perdida la cartera If one thinks of the wallet as still being in one’s possession, then a resultant state interpretation, while not inevitable, is possible. If one no longer considers the wallet as being in one’s possession, one is less likely to consider the resultant state, because if the subject no longer has the wallet, he can no longer maintain it in any state. In a sense, if the owner no longer possesses the wallet, it is no longer in a state of being lost by that owner. And so emphasis shifts from the state of the wallet to the act of losing it: from ‘It is lost’ to ‘I have lost it.’ For speakers who show a low degree of grammaticalisation of tener, where the durative state is stressed, this interpretation is not possible, and the sentence is ungrammatical. It seems that initially it is not enough for the object to be in a resultant state; the subject must still be closely related to that state. Here are some judgements from the first informant: (113) Tengo prestado el coche (114) Tengo pedido el libro (115) *Tengo enviada la carta (116) *Tengo comida mucha carne Once the subject has sent the letter it is no longer anything to do with him; therefore one does not think of it as having been sent by the subject. Once the meat is eaten it is no longer in a resultant state as it does not exist. But if the subject has lent the car to someone then it is still his responsibility, and if he has a book on order then he has an interest in that book. This may seem like a disguised form of the notion of possession. Certainly, possession is involved, but only as part of the more general semantic properties of verbs. By thinking in terms of these general semantic properties examples such as (114), which all speakers accepted, and in which there is no expression of possession, can be more easily accounted for. Here the imperfective copular function of tener as described by Roca Pons is clearly
Tener +past participle in modern Castilian
63
in evidence. A further indication that this function is more important than possession is to be found in judgements on sets of sentences such as the following, on which half of the speakers agreed. (117) El humo me tiene irritados los ojos (118) *El chico me tiene rota la pierna (119) Ese chico me tiene roto el corazón While (117) can refer to a durative state—the smoke is currently irritating the speaker’s eyes—(118) cannot do so: the boy may have broken the speaker’s leg, but now that it is broken he cannot keep on breaking it. There is no difference in the implications regarding possession, but grammaticality is decided by whether or not some kind of durative value is present. This would explain why (119) is grammatical while (118)—which contains the same verb—is not. In (119) the boy can be thought of as continuing to make the speaker unhappy. The four speakers who differed from the results given above did so because they did not accept sentence (117). It seems that speakers differ over how long a situation must last in order for tener+past participle to be grammatical, or that some speakers consider that as the smoke is not actively keeping the speaker’s eyes sore, tener+past participle cannot be used. Example (117) was replaced by (117a)
El humo me irrita los ojos
In contrast, sentences such as (119) above or (120) below are possible, even though in (120) tener+past participle could also be replaced by the simple present. (120) Ese chico tiene preocupada a su madre In some cases a sentence is ungrammatical even where an imperfective interpretation would be possible. For example, all speakers rejected (121) *Tiene querido ese coche (122) *Para su cumpleaños tiene querido ese regalo (123) *Tiene deseadas tantas cosas In each case tener+past participle was replaced by the simple present. Stative verbs do not generally appear in the construction. Sentences (121)-(123) seem to be tautologous: tener+past participle generally implies that the subject is keeping something in a particular state; but a verb such as querer or desear also implies this, and so tener+past participle is redundant. This can be illustrated by a verb such as pensar, which can be either stative or non-stative. Tener pensado is acceptable where it indicates both punctual action and durative resultant state (cf. English ‘I’ve got it thought out’). There is no such process with *tener querido. Where pensar is stative, it is unacceptable, like querer. 13 There are occasional examples of tener with the past participle of stative verbs. The following one is taken from El Jarama by Sánchez Ferlosio: ‘¡En toda mi puta vida no me vuelvo a bañar en este río! ¡Lo tengo aborrecido para siempre!’ But in an example such as this tener+ past participle does not apparently express a simple present, as in the ungrammatical examples (121)–(123) above. It seems to stress the change in state resulting from a previous event (i.e. the speaker hates the river because of what has
Tener+Past participle
64
happened). It is also a particularly emphatic way of expressing the continuing existence of the imperfective state, almost implying the active intention of the speaker to maintain the state. Various factors suggest that the grammaticalisation of tener+past participle involves a move away from the imperfective aspectual value inherent in lexical tener, and away from the copular function of tener as described by Roca Pons. In general, a durative resultant state interpretation takes precedence over an iterative or a punctual one. Verbs which are less likely to admit a resultant state interpretation are less readily accepted in the construction. For speakers with less tolerance of tener+past participle the resultant state interpretation cannot be cancelled and replaced by an iterative interpretation. It seems that there is a scale ranging from examples which focus purely on imperfective state (e.g. Tengo rota la pierna), through examples of resultant state with emphasis on the state of the object, but with some suggestion of the causative action (e.g. Tengo pedido el libro) to examples where there is greater emphasis on the causative action. But there are other examples which have not yet been accounted for. So far, we have considered only cases with a nominal direct object. Some of the examples with sentential complements have the same durative value as those examples already discussed. For example: (124) Tiene prometido comprarme un regalo (125) Tenemos pensado ir al cine esta tarde (126) Mis padres me tienen prohibido que salga These examples are similar to (127) Me tienen prometido el regalo (128) Tenemos pensada la excursion para la semana que viene (129) Mis padres me tienen prohibidas muchas cosas All speakers apart from informant 8 accepted all these examples. So even where there is a relatively low degree of grammaticalisation, tener has grammaticalised to the extent of taking direct objects which are not nouns. Perhaps more interesting is the fact that all speakers (except informant 8, who did not accept (131) or (132)) accept sentences such as (130) Ya te tengo dicho que no hagas eso (131) Me tiene dicho repetidas veces que no piensa casarse con él (132) Nos tiene contadas muchas historias (133) Tengo oído que mañana no va a haber clase In these cases there is more of a punctual or an iterative interpretation. These sentences are regarded as perfectly grammatical by people who do not accept sentences with nominal direct objects which suggest a punctual or iterative interpretation rather than a resultant-state one. Where speakers had a relatively low tolerance of tener+past participle there were very few examples of this type, and it may be that the sentences which were accepted are only isolated unproductive examples. However, it should be pointed out that all speakers accept cases where not only is there no nominal direct object but also where there is no durative state.
Tener +past participle in modern Castilian
65
2.6 CÓMO AND QUÉ QUESTIONS In section 1.3 I discussed briefly the use which some linguists make of interrogatives in deciding which verbs to classify as auxiliaries. According to these writers, with auxiliary verbs questions of the following type can be formed. (134)
Estaban contándolo
(134a)
¿Qué estaban haciendo?
(135)
Iban a contarlo
(135a)
¿Qué iban a hacer?
Verbs which are not auxiliaries do not allow this type of related interrogative. (136)
Lloraban contándolo
(136a)
*¿Qué lloraban haciendo?
(137)
Vinieron para contarlo
(137a)
*¿Qué vinieron para hacer?
Auxiliaries do not have a similar related interrogative formed with cómo. (134b)
¿Cómo estaban?
(135b)
¿Cómo iban?
(These are grammatical but not related to (134) and (135).) So if tener is an auxiliary we should be able to form interrogatives with qué and not with cómo. If it is not an auxiliary then we would expect the reverse. But as was suggested in chapter 1, it is not so simple as that. For six of the eight informants the situation is fairly straightforward: interrogatives of the qué type are not possible. This suggests that tener is not an auxiliary. But not all examples of tener behave in the same way: while some have a related cómo question, others do not. Although the informants did not agree completely as to which examples did or did not allow a cómo question, the results are broadly the same. Obviously, where the direct object is not a noun, the cómo question is impossible. (138)
Te tengo dicho que no hagas eso
(138a)
*¿Cómo tienes?/¿Cómo tienes que no hagas eso?
(139)
Tiene prometido comprarme un vestido
(139a)
*¿Cómo tiene?/¿Cómo tiene comprarme un vestido?
But not all examples with a nominal direct object have a cómo question. In general, the further the construction has moved away from the imperfective copular or mantener value of tener, the less likely it is to permit a cómo question. The question is possible with examples (140)–(142). (140)
Tengo rota la pierna
Tener+Past participle
(140a)
¿Cómo tienes la pierna?
(141)
El bedel tiene cerrada la puerta
(141a)
¿Cómo tiene la puerta?
(142)
Tengo la casa barrida
(142a)
¿Cómo tienes la casa?
66
As was seen earlier, it is not always easy to disentangle the questions of possession and more general semantic properties. In cases where there is no direct object of the type which can be possessed—either because of the nature of the object or because the verb entails an action which destroys the object—there is no possible cómo question. (143)
Tiene dichas muchas tonterías
(143a)
*¿Cómo tiene las tonterías?
(144)
Tengo comida mucha carne
(144a)
*¿Cómo tienes la carne?
It seems that where the verb involves a direct negation of possession there is no cómo question. (145)
Tiene dado el coche a Juan
(145a)
*¿Cómo tiene el coche?
But it appears that possession in a strict sense is not always necessary for the cómo question to be possible. There was disagreement among the informants over whether or not a cómo question is acceptable where the subject does not have the object in his possession or control. Some speakers accepted it in sentences (146)–(148), others did not. (146)
Tengo prestado el coche a los alumnos
(146a)
¿Cómo tienes el coche?
(147)
Tengo enviada la carta
(147a)
¿Cómo tienes la carta?
(148)
Tengo perdida la cartera
(148a)
¿Cómo tienes la cartera?
Where a speaker accepts cases where a resultant state interpretation is cancelled by the addition of an iterative adverbial phrase, no related cómo question is possible, even if one is possible where there is no such iterative phrase. (149)
Tengo castigado al niño
(149a)
¿Cómo tienes al niño?
(150)
Tengo castigado al niño muchas veces
Tener +past participle in modern Castilian
(150a)
67
*¿Cómo tienes al niño?
This is as one might expect, given that the cómo question is supposed to relate to the less grammaticalised uses of the construction, and also that the iterative examples do not imply that the object is in a particular state at the moment of utterance. The replies of these six informants largely fit in with the interrogative tests put forward by people such as Fontanella. The use of the cómo question indicates that tener is not fully grammaticalised. The question is not always possible, and the cases where it is not appropriate are those where tener+past participle is apparently more grammaticalised. Even where the construction seems to represent something close to a present perfect construction similar to haber+past participle, a qué question is not possible. Although the interrogative tests seem compatible with the conclusions reached here as to the relative degree of grammaticalisation of examples of tener+past participle, they do not provide a means of differentiating clearly between auxiliary and non-auxiliary use of tener. Even though for these speakers the qué question can be discounted, there are still two types of construction: those which permit a cómo question and those which do not. Two informants remain to be discussed. One of these (informant 2) produced results like the ones discussed above except in one important respect: he did sometimes accept the question ¿Qué tienes hecho? However, the relevance of this fact is not to be found by just stating which sentences the question applies to. The informant explained that the question cannot refer to a single action but has to refer to what has been achieved during a period of time: a day, a year, one’s lifetime, etc. So for any individual sentence the question ¿Qué tienes hecho? is inappropriate. Many of the sentences could appear in a list of things done, given in a reply to such a question; but the question does not relate directly to any particular example. It is a general question, which could easily provoke an answer containing haber+past participle or the preterite, for example. The informant paraphrased ¿Qué tienes hecho? as ¿Qué es lo que durante tu vida/este año/etc. has ido haciendo? or, more simply as ¿Qué llevas hecho? Sentences such as Tengo barrida la casa; Tengo comprada la comida; Tiene escritos muchos libros can be related to a qué question. But a sentence such as Te tengo dicho que no hagas eso cannot, as there is no idea of achievement. So for this speaker the qué question is of little help in distinguishing between auxiliary and nonauxiliary uses of tener. Some of the examples which might be considered to be among the most grammaticalised examples do not permit the qué question; other apparently less grammaticalised examples do permit it. It is interesting to see that this informant is not one who shows a relatively high level of grammaticalisation of tener+past participle, as might have been expected. The final informant to be discussed is informant 6, who does show a fairly high level of grammaticalisation. She too accepts the qué question, but in a way rather different from informant 2. Her judgements may be thought to indicate a development of the situation indicated by informant 2’s judgements. The judgements of informant 6 divide the sentences into three main categories: those sentences which take only a cómo question; those which take both a cómo and a qué question; those which take only a qué question. The sentences which allow only a cómo question are the ones where tener has a lexical value. (151) Tengo rota la pierna
Tener+Past participle
68
(152) Ese chico tiene preocupada a su madre (153) Con tanta conmoción todo el mundo tiene perdida la cabeza Turning to the other two categories, we find that here tener has grammaticalised to some degree at least. Examples (154)–(156) have both a cómo and a qué question. (154) Tengo mandado el libro (155) Se tiene invertido mucho dinero en ese negocio (156) Tiene proyectada una máquina que haga el trabajo de veinte hombres Examples (157)–(164) have only a qué question. (157) Ya te tengo dicho que no hagas eso (158) Tengo comido muchas veces en ese restaurante (159) Tienen viajado mucho por el extranjero (160) Tengo comida mucha carne (161) Tiene perdidos cinco kilos (162) Tengo vistas muchas películas suyas (163) Nos tiene contadas muchas historias (164) Tiene conocidas muchas chicas durante sus viajes The lack of a cómo question for some of these examples can be explained, as before, by the fact that the direct object is not a noun (157)–(l 59) or no longer exists (160)–(161). Examples (162)–(164) are ones where there is less of a resultant state value, backing up the view that the cómo question indicates a lesser degree of grammaticalisation, if any at all. Part of the reason for these examples not taking a cómo question could be that there is no object which can be literally possessed or in the charge of the subject. But cómo questions are possible where the object noun is not possessed, as in example (154). Even though the qué question presupposes a certain degree of grammaticalisation, it is not necessary for the subject of tener and the agent of the past participle to be identical in order for the question to apply. However, the subject of tener must be felt to be in some way responsible for the action. For example, No he traído al niño conmigo porque le tengo castigado allows both questions, even if the subject has not meted out the punishment him or herself. But where it is made clear that the subject and agent are different—as in No he traído al niño conmigo porque le tengo castigado por su padre— the qué question is not applicable. Comparing the results from all the speakers, we can conclude that where tener+past participle is least grammaticalised it allows either a cómo question or neither of the questions; then a qué question becomes possible, in certain restricted contexts; the qué question widens its range of application, and the cómo question gradually becomes less appropriate, especially where the construction is most grammaticalised. We see from this that use of the cómo and qué questions is applicable to the grammaticalisation of tener. But although the questions can be used to help plot the progress of the construction, they do not provide clearcut answers to the question of whether or not tener is an auxiliary— as certain writers seem to assume they should do. A note of warning should be sounded about the reliability of the tests. In some cases it was quite difficult to discover whether or not a cómo question was possible. Informants would sometimes reject it but then say that grammatically it was possible, although it was a very odd question to ask. A recurrent example is Tengo castigado al niño. Several
Tener +past participle in modern Castilian
69
informants said that while they supposed the question ¿Cómo tienes al niño? was possible in this context, the usual reply one would expect to this question would be bien, enfermo, etc. If it is viewed as part of a grammatical exercise, the applicability of the cómo question can be used in the study of auxiliary verbs. But it presupposes a certain degree of linguistic sophistication on the part of the informant. Although almost any tests are going to seem slightly forced and removed from normal use of language, one should be wary of moving too far away from linguistic reality: the more theoretical and linguistically unnatural the tests become, the harder it is for informants to give natural replies or honest judgements. 2.7 CLASSIFICATION OF TENER +PAST PARTICIPLE Bello and Cuervo (1898, 187) imply that tener+past participle is equivalent to haber+past participle. However, most linguists maintain that the two constructions are far from equivalent. For some, tengo+ past participle is entirely a present tense, expressing the current state resulting from a past action. Alonso and Henríquez (1945, 115) state that ‘haber con participio indica la acción pasada; es pues un pretérito ocurrido con anterioridad al tiempo del auxiliar haber…. Tener con participio es un presente en el que se acumula una acción pasada; indica el resultado actual de la acción pasada.’ Similar views are expressed by Hanssen (1913, 233–4), Lenz (1925, 401), Gili Gaya (1955, 140– 1), Ramsey and Spaulding (1956, 330). This resultant state value is indeed the most frequent one for tener+ past participle. It is also the one which tends to take precedence over other possible interpretations. But other possible interpretations do exist. Something of this sort is suggested by Keniston (1937, 195), who points out that periphrases containing the past participle usually denote ‘the attainment of a present state as a result of the action’—the same view as that discussed above. But after the example he gives of tener+past participle—Tengo oído que es una alhaja—he comments that ‘this is almost a true perfect’. (For Keniston the perfect indicates ‘an action or state as completed at the time expressed by the auxiliary verb of the tense.’) Here is some indication that the value can vary: tener+past participle can refer to the resultant state or to the causative action. Spaulding (1958, 43) also suggests that tener+past participle may either substitute for haber+past participle or represent a resultant state. These various uses of tener+past participle are perhaps best recognised by Criado de Val (1948, 117–20). He rejects the view that tener+past participle expresses only a current resultant state: Aun reconociendo que el perfecto formado con tener no es en ninguna manera comparable al formado con haber ni en frecuencia, ni en significación, ni en la extension de su zona lingüística, sí creemos en cambio que posee un sentido claro de perfecto, y que conserva con precision el doble valor ‘terminativo’ e ‘iterativo’. (p. 118)
Tener+Past participle
70
He goes on to say (pp. 118–19) that ‘el doble aspecto “terminativo” e “iterativo” propios de esta forma no siempre se presentan juntos. Puede existir solamente el primero, pero no así el caso contrario, es decir, el perfecto con tener con sólo valor iterativo.’ Perhaps the most detailed description of the various interpretations of tener+past participle is provided by Fente et al. (1972, 43–5). But as they are writing a descriptive account for foreign students of Spanish they limit themselves to listing possible uses, and do not attempt to find any overall pattern in usage. They classify tener+past participle as having, like haber+past participle, a primarily terminative aspect, but as differing from haber+past participle because it expresses nuances of duration, repetition or insistence, or accumulation. They point out that nuances overlap and that frequently context decides which is to be the dominant value. Although their view of overlapping values is undoubtedly correct, at times the importance they place on context leads them to make unnecessary divisions in their classification and to miss broader generalisations. For example, under the heading of terminación they give (165) Ya tenía proyectado ir a Suecia antes de que me lo sugirieras And under the heading of duración (166) Me tienen prohibido que salga However (165) does not just refer to a completed action. It is not an end point but a durative situation. The only thing that really implies terminación in (165) is the antes clause. But presumably the speaker still planned to go to Sweden after the suggestion was made. The important element is not so much the fact that the decision-making is in a state of completion, but the resulting situation. There is no major difference between (165) and (166). In (165) the emphasis may be slightly less on the action and more on the resultant state, but not enough to warrant separate classification. It seems more advantageous to think of the various examples as ranging along a graduated scale. This is more or less what Fente et al. imply by their comments on the overlapping of interpretations. But in their attempt to pinpoint particular uses for the foreign learner they then lose this general outline. Fente et al. also state that in this construction tener cannot be used in compound forms. But informants readily accepted sentences such as Me han tenido prohibido que salga; He tenido varias veces castigado al niño; He tenido perdida la cartera varias veces. It seems that this use of haber tenido is possible only where tener has a mantener interpretation. It is not possible where tener +past participle is more grammaticalised, as in *Te he tenido dicho que no hagas eso; *He tenido comida mucha carne. (Compare the similarly ungrammatical *Te he habido dicho…; *He habido comido….) Haber tenido+past participle stresses the state more than the equivalent sentence with the simple present form of tener and the same participle. Although Tengo perdida la cartera principally indicates current state, there is some indication of the causative action. This is not the case with He tenido perdida la cartera varias veces. Haber tenido also appears to be possible only where there is an iterative aspect. So, for example, while He tenido perdida la cartera varias veces is possible, if the speaker wishes to indicate that the wallet was lost only once (and has since been found) the sentence would be Tenía perdida la cartera.
Tener +past participle in modern Castilian
71
2.8 ACCOMMODATING GRADUATED CLASS MEMBERSHIP The examination of various classificatory criteria for auxiliary verbs and the survey of native-speaker judgements suggest that classifications based on rigidly defined mutually exclusive categories are inadequate. There is a need for a framework which can accommodate the fact that a particular construction may not fit neatly into any one traditional category; that when attempts are made to classify a construction, examination of different instances of that construction may produce different results; and that the usage of different speakers calls for different categorisation. 14 The framework should also be able to combine synchronic and diachronic considerations, so that synchronic variation can, where relevant, be accounted for as the effect of diachronic change, and variation can be ordered according to the extent of the diachronic development it denotes. The idea of variation being the result of change is of course nothing new, but traditional descriptions frequently have difficulty in incorporating such a view. In recent years attempts have been made to find ways of formalising this variation and uniting synchronic and diachronic descriptions. Lehmann (1985) puts forward a framework which allows for graduated class membership and can be used to describe both synchronic states and diachronic developments. He starts from the view that grammaticalisation has both synchronic and diachronic implications. Under the diachronic aspect, grammaticalization is a process which turns lexemes into grammatical formatives and makes grammatical formatives still more grammatical…. From the synchronic point of view, grammaticalization provides a principle according to which subcategories of a given grammatical category may be ordered. (p. 303) He observes that grammatical devices ‘do not fall into neatly distinct classes, but differ only gradually so that they may be ordered on a scale’ (p. 305). This scale he calls a grammaticalisation scale, which is formulated according to a set of criteria which measure the autonomy of that device. The more autonomy a device (or language sign) has, the lesser the extent to which it has grammaticalised. He suggests three factors which should be taken into account when assessing the autonomy of a sign. These are: weight— ‘a property which renders it [the language sign] distinct from the members of its class and endows it with prominence in the syntagm’; cohesion—the extent to which a sign ‘systematically contracts certain relations with other signs’; variability—‘a momentary mobility or shiftability with respect to other signs’. Each of these factors has a paradigmatic and a syntagmatic aspect which relate to the sign’s selectional and combinatorial properties respectively. Paradigmatically, the weight of a sign is its integrity. ‘its substantial size, both on the semantic and the phonological sides’. Syntagmatically, it is its scope: ‘the extent of the construction which it enters or helps to form’. The cohesion of a sign is divided into paradigmaticity—‘the degree to which it enters a paradigm, is integrated into it and depends on it’—and bondedness—‘the degree to which it depends on, or attaches to, such other signs’. Its paradigmatic variability is ‘the possibility of using other signs in its stead or of omitting it altogether’; its syntagmatic variability is ‘the possibility of shifting it around in its construction’. So
Tener+Past participle
72
Lehmann’s theory produces six parameters by which to measure the degree to which a construction has grammaticalised in any synchronic state of the language. These synchronic parameters can be reformulated as diachronic processes (i.e. the further a sign develops historically via one of these processes, the more grammaticalised it will be when measured by the corresponding synchronic parameter). The process which decreases a sign’s integrity Lehmann calls attrition, which entails ‘the gradual loss of semantic and phonological substance’. The process which leads to increased paradigmaticity is termed by Lehmann paradigmaticization. This process ‘integrates syntactic constructions as periphrastic forms into morphological paradigms and leads to increasingly small, homogeneous paradigms’. Loss of paradigmatic variability occurs through obligatorification: ‘Within the paradigm choice amongst its members becomes constrained by grammatical rules. The whole category represented by the paradigm becomes increasingly obligatory in the sentences of the language.’ The scope of a sign is reduced through condensation. ‘The more a sign is grammaticalized, the less complex become the constituents with which it can combine.’ An increase in bondedness occurs through coalescence which ‘leads from juxtaposition via cliticization, agglutination and fusion to symbolic alternation’. Finally, Lehmann terms the loss of syntagmatic variability fixation; ‘the grammaticalized sign tends to occupy a fixed syntactic, then a morphological position and becomes a slot filler.’ Table 2.2 (taken from Lehmann 1985, 309) outlines the relation between synchronic parameters and diachronic processes, and indicates the effect of such processes. Each parameter or process is separate from the others. As a linguistic sign grammaticalises the parameter values change progressively. While usually each value changes, they may not all change at the same rate; this helps to account for the fact that signs do not fall into clear-cut categories, and that they may ‘pass’ one traditional test for a particular status but still ‘fail’ another. Traditional criteria are still very useful, but as guidelines for placing signs on a scale, not as ‘pass/fail’ mechanisms. Many of Lehmann’s criteria are very similar to traditional criteria, but they allow more readily for signs to be placed relative to each other rather than in discrete categories. Semantic attrition is similar to the criterion of semantic loss which was considered in chapter 1. We saw there that it is not always easy to measure such loss, or even to decide what the principal meaning of a ‘full verb’ is. However, although this criterion does present difficulties, when looking at examples of tener+past participle we must consider what possible semantic value tener has in each case. We have seen that in many cases tener retains some notion of possessing or of keeping something, but that in some cases this notion is lost. Tener undergoes no phonological attrition: there is no phonological distinction between tener as a full verb and tener in any of its posited auxiliary functions.
Table 2.2 Lehmann’s parameters and processes of grammaticalisation Parameter
Weak grammaticalization—Process
Strong grammaticalisation
Integrity
bundle of semantic features; attrition possibly polysyllabic
few semantic features; oligoor monosegmental
Tener +past participle in modern Castilian
73
Paradigmaticity item participates loosely in semantic field
paradigmaticization small, tightly integrated paradigm
Paradigmatic
free choice of items accord
obligatorification
variability
ing to communicative intentions
constrained, use largely obligatory
Scope
item relates to constituent of condensation arbitrary complexity
item modifies word or stem
Bondedness
item is independently juxtaposed
coalescence
item is affix or even phonological feature of carrier
Syntagmatic
item can be shifted around
fixation
item occupies fixed slot
variability
freely
choice systematically
Paradigmaticity is not easy to assess either. It seems that a high degree of paradigmaticity is indicated by a particular construction having a closely constrained grammatical function, rather than exhibiting a range of possible interpretations which are not part of the grammatical ‘system’ of the language. However, how one decides whether something is a member of a ‘small, tightly integrated paradigm’ is not clear. (This point will be raised again later (see p. 90).) Under paradigmatic variability we need to consider whether tener+ past participle can be replaced by any other construction. Is there any other means of saying the same thing? If there is, is there a free choice between the various possibilities, or is that choice constrained in any way? Also, we must consider whether tener can be used with the past participle of all verbs, and—if it cannot—what the limitations on its cooccurrence are. Under scope, we need to consider whether tener relates to a ‘constituent of arbitrary complexity’ or whether it modifies a single word or stem. In order to do this, we need to decide on the need or otherwise for tener+past participle to have a direct object: if the direct object is obligatory, then tener not only relates to the past participle but to the object. We should also place the question of the agreement of the past participle with the direct object under the heading of scope. This might not seem to relate directly to the scope of tener, but it is an important factor. Firstly, we can say that as tener usually requires a direct object, it combines with a complex set of constituents, not merely with the past participle. In addition, we can say that if the past participle must agree with the object, then tener takes a noun phrase with a predicative adjunct (see Lehmann’s comments on Latin habere (1985, 308)). Tener has thus not lost its ability to predicate, which Lehmann sees as an important step in the process of condensation. Usually the relaxing of the agreement requirement goes hand in hand with a relaxing of the requirement for a direct object. We must also take into account whether or not the subject of tener is the same as the agent of the past participle. If the two are identical, then tener is tied more closely to the past participle, and less closely to the direct object, resulting in a decrease in its scope; this is indicative of greater grammaticalisation. Under bondedness and syntagmatic variability we need to consider how closely tener is linked to the past participle—for example, whether other elements such as the direct object can be interposed, and what limits there are on such intercalation—and whether
Tener+Past participle
74
tener+past participle exhibits a fixed word order or whether they can be freely rearranged. Various of Lehmann’s parameters overlap each other. The process of obligatorification leads to a decrease in paradigmatic variability: i.e. a sign increases its range of co-occurrence with other signs. This is brought about by the loss of some of the sign’s selectional restrictions. For example, in the case of tener, it would mean that tener could be used with the past participles of intransitive as well as transitive verbs. But this loss of selectional restrictions is also involved in the process of condensation (i.e. the decrease in the scope of a sign). We have seen that one factor in determining the scope of tener is whether or not there is an obligatory direct object. The necessity for a direct object—and also the nature of that object—is dictated by the selectional restrictions. So, loss of selectional restrictions means that tener widens its range of cooccurrence (i.e. decreases its paradigmatic variability) and condenses its scope (it relates only to the past participle and not to the direct object). 15 Let us now return to the question of paradigmaticization. I suggested above that although Lehmann uses paradigmaticity as one of his criteria, his definition is not very precise. By way of illustration he considers auxiliary verbs, but merely states (1985, 307) that The primary auxiliary verbs of French, avoir ‘have’ and être ‘be’, are completely integrated into the conjugational paradigm, which their Latin predecessors habere and esse/stare were not. The secondary auxiliaries such as aller ‘go’ and venir ‘come’ are slightly less paradigmaticized, but still much more so than the Latin ambulare and venire, from which they derive. Vincent (1987, 242–3) takes a similar view: One widely used criterion for according a verb auxiliary status is that the periphrasis it defines should have become absorbed into the overall conjugation system. Thus…habere in its classical uses meets none of the criteria for auxiliaryhood, while its Romance descendants avoir, avere, haber, etc. are classified as auxiliaries on the grounds precisely that they are at the heart of the so-called compound tenses j’ai mangé, ho mangiato, he comido, etc. But Vincent’s approach suggests that grammaticalisation is to some extent independent of some of the factors put forward by Lehmann, in that a form may exhibit various syntactic rather than morphological properties and still be highly grammaticalised: Synchronically, it seems clear that the morphological and syntactic subcomponents [of a grammar] will have to have the capacity to overlap, assigning a double analysis to certain forms, such as the Romance perfects, whose syntax may not compositionally reflect their morphosemantic function. (Vincent 1987, 250)
Tener +past participle in modern Castilian
75
He suggests that certain constructions may exhibit properties of linear order, surface separability and perhaps even deletability which suggest that they are syntactic, while at the same time entering into contrastive opposition with elements more normally regarded as morphological. (p. 251) Considerable importance, then, is placed on the question of paradigmaticity. However, while people may have an intuitive idea of whether or not a form is part of a highly integrated paradigm, it is not easy to formalise this intuition. If a form can be seen to fill a ‘gap’ in the system, then it may be considered to be part of that system. That presupposes a predetermined system which exists at least partially independently of the members of that system. But as Vincent (1987, 240) points out: It must be borne in mind that at the outset any periphrastic formation has an internal motivation of its own as a function of the syntax and semantics of its constituent parts. This…means that if a periphrasis gets drawn into an inflectional system, it does so, in part at least, on its own terms, sometimes creating new morphosemantic oppositions, sometimes generating new forms on the analogy of its own structure. 16 There appears to be no other morphological means of expressing what tener+past participle expresses. Tener+past participle is like estar+ past participle in that both indicate the present result of a past action; but they differ in that what is expressed by the subject of tener (which is frequently, although not always obligatorily, the agent) is not expressed in the estar+past participle construction, where the subject is the patient. (Compare, for example, El bedel tiene cerrada la puerta and La puerta está cerrada (*por el bedel).) There is no other morphological means of indicating the repetition or duration in some cases expressed by tener+past participle. Haber+past participle can only do so through the addition of adverbial phrases such as muchas veces. We have to decide whether tener+past participle is a ‘morphological’ category, even if it is still a periphrasis, and whether it contracts regular oppositions with other forms in the tense/aspect paradigms of Spanish. The fact that it may be expressing a category which was not previously morphologically distinguishable in the language does not mean that it is not at least partly integrated into the inflectional system. We cannot exclude a form from paradigmatic oppositions because it expresses a tense-aspect combination which previously was neither given formal expression in the paradigm, nor was an obvious ‘gap’ in the paradigm. So, the task of deciding whether or not a new form is integrated into a paradigm is not an easy one. Considerable attention must be paid to other parameters proposed by Lehmann, such as the extent to which a particular form is obligatory (i.e. cannot be replaced by any other form) and—in the case of an auxiliary verb—whether it can cooccur with all verbs in the language. If we take the view that verb paradigms may include both morphological and syntactic forms, then, when deciding to what extent tener is to be thought of as an auxiliary, we might place more emphasis on the parameters of integrity, paradigmaticity, paradigmatic
Tener+Past participle
76
variability, and scope, as compared with those of bondedness and syntagmatic variability. To do so is not to deny the importance of these last two parameters, but merely to suggest that while they represent an important part of the process of grammaticalisation, they do not play such a major role in deciding whether something is or is not an auxiliary: they come to the fore once something is already an auxiliary, and may be developing further along the grammaticalisation scale towards becoming an affix. To sum up the position of modern tener+past participle as measured by Lehmann’s criteria, we can say that in most cases it has undergone little semantic attrition, but in examples such as Ya te tengo dicho que no hagas eso semantic attrition is considerable. There is no phonetic attrition. The question of paradigmaticity has been discussed, and while instinctively tener+past participle appears to be near the weakly grammaticalised end of the scale, it is hard to measure its degree of paradigmatic integration. As for paradigmatic variability, at the least grammaticalised end, use of tener+past participle is largely a matter of free choice, but with the more grammaticalised examples, such as the one cited above, tener+past participle is the only way of indicating iterative or durative perfective action through the selection of a particular verbal construction, though even here its use can hardly be considered to be obligatory. Again, scope varies according to the particular example under consideration: the construction usually requires a direct object (which does not, however, have to be a noun) but in a few cases this is not obligatory; usually the participle has to agree with a nominal direct object but we have seen that in a few cases it does not do so. Under bondedness and syntagmatic variability, we have seen that tener is still an independent item which does not occupy a fixed slot. Tener+past participle has begun to grammaticalise, and some examples have developed quite a way according to certain parameters. However, as measured by other parameters its grammaticalisation is very slight. This accords with the conclusion reached earlier that tener+ past participle satisfies some tests for traditional auxiliary status but does not satisfy others.
2.9 OVERVIEW In chapter 1 we looked at various approaches to the classification of auxiliaries. We saw that many approaches provide a general outline of what constitutes an auxiliary or a main verb, but that the neat, clearly delineated categories do not correspond to the realities of language. Using tener+past participle as an example, we saw how provision must be made for graduated membership of each category; that some auxiliaries are more auxiliary-like than others; and that examples may pass one test and fail another. In this second chapter we have looked in more detail at use of tener+ past participle. Initially this was done without reference to grammatical categories such as ‘auxiliary’. The different uses of the construction were described irrespective of whether they indicated that tener was a main verb or an auxiliary. We found a range of uses which could be ordered according to the degree of acceptance granted them by different speakers. The variation in acceptability broadly corresponded to the extent to which a particular example had moved away from the durative value inherent in tener (a value which we found to be more useful in providing an overall description of tener+past participle than the more usually attributed characteristic of possession). In general, the
Tener +past participle in modern Castilian
77
further the construction moves from expressing present duration—by moving to express either resultant state or past action—the less widely accepted it becomes. This ordering of acceptance provides not only a unified means of classifying tener+past participle, but also helps to explain differences in acceptability of different examples. If one does not have recourse to this kind of explanatory description, it becomes very difficult to describe the various uses of tener+past participle according to one underlying principle. We have seen that while resultant state—the value usually ascribed to the construction—plays a major role, there are examples which do not express resultant state. Neither is it satisfactory to equate tener+past participle with some other construction which is more fully integrated into the language, such as haber+past participle (for a fuller discussion of this point see section 5.3). Tener+past participle has come into the verb system on its own terms, expressing a tense-aspect combination not expressed by other constructions, and this has to be accommodated within any overall description of the language, if necessary by creating new categories within the language. In short, tener +past participle cannot be reduced to a single use; neither can it be dealt with either as a replacement for another construction or in terms of a reorganisation of existing categories. Finally, in this chapter a new classificatory system has been examined, which—while owing much to earlier tests for grammaticalisation—allows for graduated membership of classes. It allows for constructions to grammaticalise faster along one parameter than another, and is not restricted to a single test for grammaticalisation. As was to be expected after the conclusions reached in chapter 1, such a system provides a more satisfactory means of accounting for constructions such as tener+past participle, especially as it enables us to combine synchronic and diachronic considerations in one description. We have seen how different uses of tener+past participle and differences between speakers can be attributed to a greater or lesser move away from the basic durative value of tener. This introduces a diachronic element into our description, and it is to a more specifically historical study that we turn in the next chapter.
3 Historical development of tener + past participle 3.1 AIMS OF THE HISTORICAL SURVEY In the previous chapter an analysis of native-speaker judgements was put forward in which variation between speakers was ordered according to the degree of grammaticalisation which it indicated. Grammaticalisation is a diachronic process, and this variation could correlate with differing degrees of diachronic change: i.e. by looking at variation in modern speech we may be able to form an idea of the historical changes undergone by the construction. Uses which all speakers have in common presumably belong to an early stage of development, and subsequent developments can be chronologically ordered according to their degree of acceptance by the informants. It is now time to look at the historical evidence to see how it ties in with such a hypothesis. This is a two-way process: information gained from native speakers can be used to help piece together our inevitably incomplete historical knowledge; and historical evidence can shed light on synchronic variation, helping to decide whether such variation could be the result of changes having worked through different people’s speech to differing extents. In chapter 2 we also saw how Lehmann’s approach can be adopted for both synchronic and diachronic studies. Chapter 2 concentrated primarily on the synchronic aspect, and applied it to tener+past participle. The same considerations can be used in a historical survey of the construction. From the discussion of Lehmann’s framework, we can see that a number of factors should be taken into account in such a survey. Throughout the textual study the following factors must be borne in mind: the semantics of tener and whether tener has undergone any semantic attrition; whether there is an obligatory direct object and what sort of direct object it is; the range of past participles with which tener occurs; agreement of the past participle with the object; the syntactic ordering of elements within the construction and the possibility of inserting other elements between them. This study will also consider the role of haber+past participle from two main standpoints: firstly, it provides a means of comparing two similar constructions which have grammaticalised to very different degrees; secondly, it enables us to consider the interaction of the two constructions and the way in which they contract oppositions with each other.
Historical development of tener+past participle
79
3.2 TEXTUAL STUDY An examination of the historical development of tener+past participle necessarily has to be based on the study of texts. However, textual surveys have some inherent drawbacks. They provide only partial information about the syntax of the language under examination. Given the wide range of possible syntactic constructions, examples of a particular construction may not be very numerous. Apparent gaps in usage as illustrated by the texts may be coincidental; on the other hand, they may indicate important constraints on usage. There is also the problem of interpretation. Analysis of nativespeaker judgements indicates that modern tener+past participle has various different interpretations, varying from a ‘present durative’ value to a ‘past punctual or iterative’ value. While past developments of the construction may not have produced the same results, it is reasonable to assume that in past stages too there were differing interpretations. At times, context gives a very good indication of the most likely interpretation; at other times it does not. It is also not easy to decide on the basis of textual evidence alone whether a particular factor must necessarily be present in order for the construction to be grammatical, or whether its presence in a specific example is merely compatible with (rather than essential to) the construction. For example, we saw in chapter 2 that the question of possession is often irrelevant in assessing the grammaticality of a particular instance of tener+past participle. The sentence Tengo escritas las cartas gives no indication at all as to whether or not the subject has the letters in his or her possession. So possession is not a necessary part of the interpretation, and is not an essential condition for the acceptability of the sentence. But if we were to come across such a sentence in a text, where context indicated that the subject did still possess the letters, then we would be unable to discover from this example whether or not possession was a necessary condition for grammaticality, or an important factor in interpretation. To view written texts as a poor reflection of the spoken language of the time (although of major importance in historical linguistics) is the usual approach found in linguistics (see, for example, Labov 1972a, 100–2). 1 In this study, I shall view written texts as an indication—albeit an inadequate one—of the spoken language. If we are to make comparisons between modern usage and historical developments we must make sure that the comparison is as accurate as possible: i.e. as the survey of modern Spanish tener+past participle concentrated on what is possible in the spoken language, then we must attempt to discover what was possible in the spoken, not the written language of the past. Labov (1972a, 102) encapsulates this in his Principle of Convergence: ‘the value of new data for confirming and interpreting old data is directly proportional to the differences in the methods used to gather it’. We cannot use the same methods for collecting information on former states of the language as we use for looking at the modern language, but we need to make sure that we are at least examining the same phenomena. One further problem must be borne in mind. We have seen in chapter 2 that even within a narrow sector of the speech community, native competences can vary quite considerably. Just as the native-speaker judgements examined earlier differed from each
Tener+Past participle
80
other, the language of each writer will be different. This variation will be at least as great as that discovered amongst native informants, and may be greater, as there may well be additional dialect differences, and differences in style and register conditioned by the type of work being written. In the same way that some native speakers were more tolerant of the tener+past participle construction than others, some writers will be more forwardlooking in their use of language, and others will be more conservative. We must remember this when using textual evidence, and should be wary of throwing together examples of the construction taken from different authors, even when they are from the same period. For that reason, this survey looks in detail at individual texts, rather than collecting examples from a range of texts from each period. The survey is based on 25,000-word samples from each of a series of texts which cover the period from the mid thirteenth century to the twentieth century. (For details of the texts and editions used, and the abbreviations by which they are referred to, see the list on pages vii–viii.) 2 3.3 STUDY OF TENER +PAST PARTICIPLE 3.3.1 Semantics of tener +past participle According to Seifert (1930, 266–7) the foundations for subsequent use of tener as a main verb are laid in the thirteenth century. It appears more frequently, and in usage it increasingly includes areas covered by haber. Seifert observes that ‘es sorprendente ver cuántas veces alternan ambos verbos [haber and tener] sin distinción perceptible de sentido’. However, at this stage tener is most frequently used with the meanings of ‘to keep or maintain’, ‘to be in charge of’, or ‘to occupy’, and it is essentially durative. In the early texts, when combining with a participle, tener usually retains much of this lexical meaning, or at least the interpretation is fully compatible with lexical tener. Looking in detail at the earliest text we find that in some cases tener has a full lexical meaning of ‘to hold’ or ‘to sustain’: (1) Metióli so los piedes do estava colgado las sus manos preciosas, tovolo alleviado (Mil 150 a, b) ‘Beneath his feet where he was hanging She put her precious hands, she sustained him’ (2) trovaron al bon omne con ábito estranno teniendo el ninnuelo envuelto en un panno (Mil 568 c, d) ‘They found the good man in strange attire holding the child wrapped in a cloth’ (3) Paráronse delante al Ninno coronado el qe tenié la Madre dulzement abrazado (Mil 692 a, b) ‘They stopped in front of the crowned child who the mother held in a gentle embrace’ In other cases tener expresses inalienable possession. (4) Demostrava el brazo qe tenié livorado (Mil 265 a) ‘He showed his arm which was bruised’ (5) abrió luego los ojos qe tenié adormidos (Mil 794 d) ‘then he opened his eyes which had been asleep’
Historical development of tener+past participle
81
Many examples indicate that the subject of tener has the direct object in their power or under their control in some way. For example: (6) Los malos qe vinieron afontar la tu ciella, bien los tovisti presos dentro en tu capiella (Mil 745 a, b) ‘The wicked who came to dishonour your cell, you held them well captive in your chapel’ (7) Quando lo entendió la gent adïablada quitóse de la alma la qe tenié legada (Mil 260 a, b) ‘When the demons heard this they left the soul which they held in their power’ At times the object is under the control of the subject in the sense that the subject has it stored away, or hidden: (8) sopolo la Gloriosa tener bien escondido (Mil 348 c) ‘the Virgin could keep him well hidden’ (9) el cesto en qe vino el aver bien contado so el so lecho misme lo tiene condesado (Mil 694 c, d) ‘the basket in which the money came, carefully counted, he keeps it stored away under his bed’ (10) Doquiere qe la tenga el diablo metida (Mil 849 a) ‘Wherever the devil may have it hidden away’ In the last two examples it is difficult to ascertain the relative importance of possession, having something in one’s control, and duration. In some other cases the predominant function of tener seems to be the indication of duration. It may suggest some active involvement of the subject in keeping the object in a particular state, but this element seems fairly weak: (11) las puertas de mi casa aviertas las tenía (Mil 639 c) ‘I kept the doors of my house open’ (12) Y tenié la imagen de la sancta Reína la que fue pora’l mundo salut e medicina, teniéla afeitada de codrada cortina (Mil 515 a–c) ‘And his image of the holy Queen who was salvation and solace to the world, was adorned with a coloured cloth’ At times tener may be little more than a copula: (13) qe tenié so amor en Dios bien encendida (Mil 730 b) ‘whose love of God burned strongly’ In such cases tener does not indicate possession or having in one’s power, but is still essentially a marker of duration. It can be used in this way when not combining with a participle too. Seifert (1930, 266) says that in the thirteenth century ‘tener es un instrumento a propósito para el mero enlace de los miembros de la oración, haciendo esta construcción resaltar el interés que toma el sujeto por las personas u objetos, de los cuales se refiere lo que representa el punto capital de toda la enunciación’. In all the above examples, it could be argued that tener has a lexical value: in each case, the past participle is compatible with the lexical value of tener. However, as has already been indicated, the role played by this literal meaning in the interpretation of the construction is not immediately obvious. In some cases this lexical value is considerably weakened, and the participle carries much more weight. All examples have a current durative value, and in no case is the emphasis on the previous action rather than on the current state. But where more lexical information is conveyed by the participle than by
Tener+Past participle
82
tener, and where the subject of tener is also the agent of the participle, there is the possibility of increased emphasis on past verbal action. At this stage it remains little more than a latent possibility. However, there are cases where these conditions are fulfilled. If one removes the examples where tener+past participle indicates some kind of possession, or has a durative copular function, four examples still remain which may exhibit a greater degree of grammaticalisation. In the following example the expression of the agent moves the emphasis from resultant state or inalienable possession to the previous action: (14) tenié con sus oncejas las massiellas rompidas (Mil 364 b) literally ‘she had her cheeks lacerated with her nails’ (see below) Yllera (1980, 286) suggests that in this example, it is the fact that the subject of tener is the same as the agent of the participle which brings about the grammaticalisation of the construction. This is not strictly accurate. The two are also identical in various other examples where there is little suggestion of grammaticalisation. What encourages a more auxiliary interpretation of tener in (14) is the presence of con sus oncejas. This agentive phrase refers to the action of tearing the cheeks rather than to the resultant state. The interpretation seems to be both that ‘She had lacerated her cheeks with her nails’ and ‘Her cheeks were lacerated’. Without the agentive phrase the construction reverts to being solely an expression of inalienable possession, and the fact that the subject of tener and the agent of rompidas are the same does not in itself lead to a more grammaticalised interpretation. Another possible candidate for auxiliary (or semi-auxiliary) status is: (15) La duenna pïadosa qe fue ante irada fue perdiendo la ira e fue más amansada; perdonólis la sanna qe lis tenié alzada (Mil 395 a–c) ‘The holy lady who had been angry before gradually grew less angry and became more calm; she pardoned them the anger which she had felt towards them’ (see below) If we interpret tenié alzada as having the same time reference as perdonólis, then tener+past participle is once again to be interpreted as expressing a current resultant state, with no indication of the previous action: ‘She pardoned them the anger which she felt towards them.’ However, alzar could also be regarded as punctual, referring to the moment in which the anger was aroused. In this case the interpretation is closer to ‘She pardoned them the anger which she had felt towards them’, and tener+past participle moves towards expressing previous action. Yllera suggests that in Berceo there are examples where tener+past participle does not indicate a resultant state but a repeated action which has led to the formation of a habit in the subject, and that tener underlines the continued existence of that habit: (16) el enclín e la Ave teniéla bien usada (Mil 80 c) ‘he habitually genuflected and said the “Ave”’ (17) El ‘Salve Sancta Parens’ sólo tenié usado (Mil 221 c) ‘He only knew the “Salve Sancta Parens”’ While this type of construction may be the forerunner of the iterative value later to be found in tener+past participle, it cannot be claimed that in these examples the tener+past participle construction is of itself iterative. The iterative value does not lie in the
Historical development of tener+past participle
83
construction, but in the lexical value of the past participle. According to Corominas and Pascual (1980), usar derives from Vulgar Latin *USARE, a frequentative form of UTI, meaning ‘to accustom’, ‘to be accustomed’, or ‘to frequent’. Tener usado indicates a state resulting from an iterative action in the same way that tener escondido indicates a state resulting from a punctual action. The difference stems from the different aspects of the two past participles. However, such examples at least demonstrate that tener was compatible with an iterative interpretation. This pattern of a largely lexical tener, with some cases where it weakens semantically to an indication of duration, is repeated in other texts. In PCG, tener+past participle is closely linked to the semantic value of tener as an indication of having something in one’s power or control. The breakdown of examples of the construction is as follows: besieging or surrounding an enemy; holding someone captive
6
commanding forces; having at one’s disposal; watching over
5
inalienable possession and attributes
2
mental preparation
1
Virtually all the examples of tener+past participle to be found in the Alfonsine text are covered by the most frequent uses of lexical tener, as highlighted by Seifert, in that the past participle is compatible with one or more of these meanings. Again, the important question is: ‘What is the balance between this basic meaning and any verbal force in the past participle?’ Much the same pattern is repeated in LBA. Examples indicate possession, inalienable possession, having someone in one’s power, etc., as in earlier texts. Also as in earlier texts, there are examples where the durative notion of keeping something in a particular state (present in other examples) comes more to the fore, even if there is still the possibility of some lexical value. (18) aquel buen omne vos ternia defendida (LBA 743 d) ‘that good man would keep you well protected’ (19) ¿por esto teniades a mi la puerta cerrada? (LBA 877 b) ‘so that is why you kept the door shut on me?’ (20) dios e mi buena ventura mela touieron guardada (LBA 877 d) ‘God and my good fortune kept her safe for me’ (Context indicates that (18) is to be understood as ‘that good man would keep you well protected’ rather than ‘would have protected you’.) In this text we also find examples indicating mental preparation or thought, as is found in the modern language: (21) byen cantava la rrana con fermosa rraçon mas al tiene pensado en el su coraçon (LBA 411 a, b) ‘the frog sang well, with fine words, but in his heart he had something else planned’ (22) Vnas palabras tenia pensadas por le desir (LBA 655 a) ‘I had thought out a few words to say to her’
Tener+Past participle
84
Yllera (1980,289) suggests that the use of tener amado in (23) ‘pone de relieve que la acción se viene realizando desde hace mucho tiempo y se supone que se sigue realizando en el presente’. (23) sy mucho la amades mas vos tyene amado (LBA 798 d) ‘if you love her dearly she loves you more’ Although there is no unequivocal indication that this is so, it does seem plausible. And if it is so, then it suggests a further development of tener +past participle. As Yllera points out, it is not truly a perfect here, as it does not indicate the state resulting from an action but rather the continuation of an action begun in the past. (Yllera links this to the aspectual nature of the participle, pointing out that perfective participles yield examples which express causative action and resultant state, whereas imperfective participles tend to produce a present tense rather than resultant state interpretation; cf. the comments made on examples (4)–(9) in chapter 2.) If Yllera’s interpretation is correct, then (23) indicates another step towards grammaticalisation: tener+past participle no longer refers only to a present imperfective state or action, but shifts towards past time reference by incorporating both past state and its continuation in the present. This usage can be seen as an extension of the tener usado type of example found in Berceo. Both (23) and the tener usado examples indicate the continuation of a past imperfective action. But whereas tener usado can still be regarded as ‘posesión espiritual’, in (23) there is no possible ‘literal’ interpretation for tener (not even that of ‘posesión espiritual’). Also, although the aspect of the participle does of course influence the overall aspect of the construction, it does not play such a prominent role as in tener usado. The interpretation ascribed to tener usado is based on the iterative nature of the participle (see p. 100). But there is nothing inherent in the aspectual value of amado which leads to the ‘continuation of past action’ interpretation rather than to a purely present tense interpretation, for example. (These points and their relevance to the historical evolution of tener+past participle will be taken up again later.) Cor provides no new developments. The few examples of tener+past participle it contains can all be classified under the headings of possession or of having under one’s control. In Cel we find that although many of the examples come into these categories, there are other less lexical examples, and cases where the interpretation can clearly no longer be that of current durative state or action. (24) es mayor tal galardón que el servicio, sacrificio, devoción y obras pías que por este lugar alcanzar yo tengo a Dios ofrecido (Cel 46) ‘such a reward is greater than the service, sacrifice, devotion and good works which, in order to reach this position, I have offered up to God’ (25) Vender un poco de hilado, con que tengo cazadas más de treinta de su estado (Cel 109) ‘To sell a little thread, with which I have hunted down more than thirty [women] of her sort’ These two examples are extensions of the type illustrated in (16), (17), and (23). Examples (24) and (25) are more clearly ‘past iterative’ than the previous examples, but can still be viewed as expressing the continuation of past action. However, in (24) and (25) there is more emphasis on this past action and much less on its continuation in the
Historical development of tener+past participle
85
present. At this point tener+past participle has moved away from purely representing some form of resultant or continuing state to representing past iterative or durative action. There is even one example where tener+past participle refers to a past action, with no indication of repetition or duration. (Context gives no indication of repeated action.) (26) ¿Mías señora? Antes ajenas, como tengo dicho (Cel 93) ‘Mine, my lady? Rather other people’s, as I have said’ The investigation of contemporary use of tener+past participle suggested that tener dicho was one of the most widely accepted uses of the construction—accepted even by speakers whose use of the construction was largely restricted to current imperfective states. It is interesting that historically it seems to be one of the first ‘past tense’ examples to appear. In Ter there is a range of different values for the construction. In some cases it has an essentially lexical interpretation, expressing for example, inalienable possession (as in (27)), or having in one’s power or stored away (as in (28)). (27) tiene ánimo aparejado para serlo (Ter 39.4) ‘you are in the right frame of mind to be so’ [literally ‘you have a mind prepared to be so’] (28) Suplico a vuestra merced me la [=la imagen] tenga muy guardada hasta que yo la pida (Ter 17.3) ‘I beg you to keep it [=the image] well guarded for me until I ask for it’ There are also cases where tener has a copular value, with tener+past participle indicating an imperfective state (as in (18)–(20) above): (29) Ansí que a tiempo que tenía aborrecidos dineros y negocios, quiere el Señor que no trate en otra cosa (Ter 24.19) ‘So at a time when I hated money and business, the Lord wants me to deal with nothing else’ Most of this type involve verbs of mental activity: (30) Bien creo tiene vuestra majestad entendido el ordinario cuidado que tengo de encomendar a vuestra majestad a nuestro Señor (Ter 50.1) ‘I fully believe that your majesty understands the habitual care I take to commend your majesty to our Lord’ There are also a couple of cases of resultant state tener+past participle, where there is some emphasis on the action involved: (31)cuanto toca a hacer y comprar la casa hágolo yo, que con el favor de Dios hanme dado dos dotes antes que sea y téngola comprada …(Ter 2.5) ‘I undertake whatever the house needs to do or to buy, for by the grace of God they have given me two dowries earlier than expected, and I have bought it…’ (32) Cuando me dieron la carta de vuestra señoría, ya tenía escrita ésta (Ter 32.1) ‘When they gave me your letter, I already had this one written’ Both examples are compatible with a current state possessive interpretation of tener, but in both there is emphasis on the change of state. In this way they can be viewed as pivot examples, linking lexical tener+past participle with examples which are non-lexical and which have some element of past time reference. It may be that the early examples of
Tener+Past participle
86
a shift from current state to resultant state/past tense involve cases which are still compatible with some lexical interpretation of tener. But in this text there is an increased proportion of non-lexical examples. Approximately one third of the examples come into this category. In some cases, there is an iterative interpretation of the type discussed earlier (see examples (16), (17), (23), (25)) which may well continue up to the present: (33) el Señor se lo pagará con las demás mercedes y buenas obras que vuestra señoría me tiene hechas (Ter 14.4) ‘God will repay you for it, along with the other favours and good works which you have done for me’ But other examples are neither durative nor iterative, and do not refer to events which continue into the present: (34) A no hallar el regalo que vuestra señoría tenía mandado en está casa, fuera peor (Ter 8.3) ‘If I had not found the present which you had sent to this house, it would have been worse’ (35) hágame vuestra merced saber…si fue adelante lo que vuestra merced tenía concertado hacer (Ter 29.4) ‘let me know…if what you had planned to do went ahead’ There are also examples which are compatible with a lexical interpretation of tener, but where context suggests that they are similar to examples (34) and (35) in that they indicate past action rather than current state: (36) Y para lo que vuestra merced tiene recibido, no me parece cumplía con menos que lo que hace (Ter 40.3) ‘And considering what you have received, what you are doing seems the least you can do to repay it’ Cases such as those exemplified in (34)–(36) are comparable with examples of haber+past participle, and at times there appears to be little difference in usage between the two constructions. However, this comparability applies only to specific examples of the construction. Overall there is still a marked difference between the largely durative present or resultant state tener+past participle and the usually past tense value of haber+past participle. (For further comparison of tener +past participle and haber+past participle see section 3.5.) Qui presents a similar picture in terms of the range of the construction. Once again there is a decrease in the importance of lexical tener. A few examples are present where tener can be interpreted as indicating inalienable possession or having something in one’s power or possession. But in many cases which are compatible with some lexical interpretation of tener, the link with such an interpretation is much less clear than in earlier texts. Frequently the semantics of tener are of much less importance than those of the participle. All that several cases which are compatible with lexical tener retain semantically of tener is the idea of duration, which combines with the participle to indicate either current or resultant state. (See the suggestion made earlier (p. 104) that
Historical development of tener+past participle
87
some of the earliest cases to be affected by semantic loss and grammaticalisation involve examples which are compatible with lexical tener, thus providing a link between ‘old’ and ‘new’ uses of tener+past participle.) (37) en fin se ha de acabar con el mesmo mundo, que tiene su fin señalado (Qui 96) ‘eventually it will destroy the world, whose fate is sealed’ (literally ‘which has its end indicated’) (38) ya yo tengo reducida a mi mujer a que me deje ir con vuestra merced (Qui 86) 3 ‘I’ve now persuaded my wife to let me go with you’ There are various examples where tener+past participle has a primarily durative or resultant-state value. (39) tan creído tiene aquello de la ínsula (Qui 54) ‘he believes that stuff about the island so completely’ (40) tengo determinado de volver a servir a mi amo don Quijote (Qui 73) ‘I have made up my mind to go back and serve my master don Quixote’ (41) …que vuesa merced me diese la ínsula que me tiene prometida (Qui 88) ‘…that you should give me the island which you have promised me’ (42) si…me tiene aparejada el diablo alguna zancadilla donde tropiece (Qui 71)
‘if…the devil has got any trap laid for me where I might slip up’ In some cases the action or change of state is more to the fore. (43) los que tenían méritamente granjeada y alcanzada gran fama por sus escritos (Qui 65) ‘those who deservedly have gained and achieved great fame through their writings’ (44)—La respuesta está en la mano—respondió don Quijote—más es resucitar a un muerto. —Cogido le tengo—dijo Sancho (Qui 98) ‘“The answer is obvious,” replied don Quixote, “it is the greater achievement to resuscitate a dead man”. “I’ve got you”, said Sancho’ While (43) is compatible with lexical tener, the emphasis is on the achievement, the action of becoming famous, as well as on the resultant state. In (44) tener cogido refers to the moment of solving the problem rather than to the state of understanding the argument. There are cases where tener+past participle indicates past action. (45) según dice el bachiller Sansón Carrasco, que así se llama el que dicho tengo (Qui 57) ‘according to the student Sansón Carrasco, for that is the name of the man I have mentioned’ Here, tener dicho seems to refer to a single, punctual, past event (compare the comments made earlier with regard to tener dicho (p. 103)). There is another possible example of a ‘past action’ tener+past participle:
Tener+Past participle
88
(46) Cuando Sancho oyó la firme resolución de su amo se le anubló el cielo y se le cayeron las alas del corazón, porque tenía creído que su señor no se iría sin él (Qui 89) ‘When Sancho heard the resolute determination of his master the heavens clouded over and his spirits fell, for he had believed that his master would not go without him’ Use of tener creído is not in itself at all surprising. We have seen various examples of this sort of construction. But tener creído would usually be viewed as expressing a current imperfective state (as in (39)). However, context implies that here tenía creído is acting as a pluperfect, i.e. it is closer to había creído than to creía. Sancho had believed that his master would not go without him, but once his master spoke, the situation no longer held good. Tenía creído still refers to a durative situation, not to a punctual event; but its time reference has changed from being the same as that of tenía alone to a time further back in the past (in the same way that había creído is ‘one tense further back’ from había). This suggests a further evolution in the construction. It also backs up the view expressed earlier that some of the first examples to be used in a more ‘advanced’ way tend to be ones which have some link with lexical tener or already have a well-established, less grammaticalised use: some of the first ‘nonpossession’ participles to appear with tener are those covered by what Yllera terms ‘posesión espiritual’. From this point onwards, we find no evidence of further developments in the semantics of the aspectual and temporal value of tener+past participle. In spite of the fact that Feijóo and Jovellanos come from north-western Spain, where, as will be seen in chapter 5, tener+past participle has undergone a slightly different evolution from that followed in standard Spanish, results from texts by these writers show no significant variation from the texts which immediately precede them. In the last two texts there is, if anything, a move away from tener+ past participle as an expression of past action. There are no clear examples where tener+past participle refers primarily to a past action, whether repeated or not, although there are still resultantstate examples. There are also a number of set phrases and metaphorical expressions. (47) Mi mujer lo tiene sentado en la boca del estómago (STP 65) ‘My wife is sick to death of it’ (48) Pero él ya se lo tenía bien tragado (CHM 19) ‘But he had already prepared himself for it’ It seems that in the nineteenth and twentieth centuries the use of tener+ past participle as a past tense has decreased (although as we saw in chapter 2, it has certainly not disappeared) and that it is now primarily a marker of duration (with, in some cases, reference to causative action). These points will be taken up later. 3.3.2 Identity of subject and agent Unsurprisingly, there is a greater tendency for the two to be different in cases where tener has a more clearly lexical value and the participle has more autonomy. In examples (49)– (51) the two are clearly different.
Historical development of tener+past participle
89
(49) tenie escripto en la palma: fasta aqui llego Hercules (PCG 8, 45–6R) ‘it had written on its palm: Hercules reached here’ (50) Esta dueña me ferio de saeta enarbolada atraviesa me el coraçon, en el la tengo fyncada (LBA 597 a, b) ‘This lady wounded me with an arrow. it pierces my heart, I have it lodged there’ (51) las [manos] tyene adormidas del grand frío que es el pecado en que enbuelto anda (Cor 116) ‘his hands are numb from the bitter cold which is the sin in which he is entangled’ In other cases, the subject and agent must be identical: (52) el enclín e la Ave teniéla bien usada (Mil 80 c) ‘he habitually genuflected and said the “Ave”’ (53) el conseio que ella tenie asmado de fazer (PCG 38, 23–4L) ‘the advice which she had decided to carry out’ (54) Mas como ya tengo escrito a vuestra merced bien largo (Ter 2.3) ‘But as I have already written to you at great length’ In other cases it is not possible to decide whether or not they are the same: (55) sacó su cuchillejo qe tenié amolado (Mil 193 b) ‘he took out his knife which he had sharpened/was sharpened’ Sometimes it seems that the identity of the agent of the participle is irrelevant. It does not matter whether or not it is the same as the subject of tener. As was pointed out in section 2.5.3, for there to be a shift from current or resultant state to previous action, subject and agent must be identical. It is likely that this is preceded by a period in which the two are usually assumed to be the same. Here, we are back to the problem of interpretation. Frequently they are the same, but it is impossible to tell whether such identity is grammatically necessary or not. In cases where they may or may not be identical we cannot tell how likely it is that they would have been assumed to be so. Such factors make it very difficult to chart the development of tener+past participle in this area. There is an apparent increase in obligatory identity of the two over the period in question. However, it may be that there is just a lower proportion of cases where such identity is automatically ruled out (rather than an increase in cases where the two actually are the same). Or it may be that the increase is real, and is due to the fact that there is a higher proportion of resultant state and past tense examples. As stated above, identity of subject and agent is necessary for this change to take place, but if we can test this identity only when the change has occurred then it cannot be used as an independent measure of grammaticalisation. To say that we know that the two must be identical where there is an increased emphasis on the previous action, and that therefore where such a change in emphasis is present subject and agent must be identical is of little practical help. In view of these difficulties it would seem safer to avoid trying to evaluate the role of subject-agent identity in the texts, while accepting its theoretical importance to grammaticalisation.
Tener+Past participle
90
3.3.3 Agreement of past participle and direct object The overwhelming tendency is for the participle to agree with the direct object. In the texts studied there are only four cases of non-agreement. Two cases display what may be thought of as disturbing factors: (56) a los de esta vida días há que los tiene vuestra señoría dado carta de pago (Ter 35.6) ‘Days ago you gained earthly peace and rest’ (57) téngame aderezado de almorzar alguna cosa caliente (Qui 86) ‘Have prepared something hot for me to eat’ In (56) dar carta de pago is a set phrase, and while this does not automatically mean that agreement will not occur, it may affect it. In (57) the direct object is the feminine alguna cosa caliente. However, the invariant participle could be caused by the close link between aderezado and de almorzar, under the influence of instances where a verb’s direct object is formed by de+infinitive (compare, for example, sentence (65) below). The example from Cel is perhaps more interesting: (58) es mayor tal galardón que el servicio, sacrificio, devoción y obras pías que por este lugar alcanzar yo tengo a Dios ofrecido (Cel 46) ‘such a reward is greater than the service, sacrifice, devotion and good works which, in order to reach this position, I have offered up to God’ Here, the past participle may agree with the first noun in the list, although that might be more likely if the participle were immediately adjacent to the noun (e.g. ‘Tengo ofrecido servicio, sacrificio, devoción …’); or lack of agreement may be encouraged by the fact that the participle is separated by a considerable distance from the direct object. (On this point see Macpherson (1967).) There is also the possibility that the service, sacrifice, etc. are being treated as a single collective entity ‘all that I have offered to God’. A further point to note is that this is one of the least lexical examples, with more emphasis on past action (albeit one that may still be continuing) than on current state. As semantically and aspectually it is one of the most evolved examples it is one of the ones most likely to exhibit non-agreement. 4 There is one more example of lack of agreement, in Jov: (59) …una carta de varias en que hago una descripción de Asturias, y de las que tengo ya escrito cinco (Jov 183.301) ‘…one letter out of several in which I give a description of Asturias, and of which I have now written five’ It is tempting (and possibly appropriate) to ascribe this to the fact that Jovellanos was from Asturias. But as mentioned earlier, there is little or no discernible north-western influence on his writings. However, whether or not this provides an explanation for example (59), we can still conclude that there is no real movement towards use of an invariant past participle.
Historical development of tener+past participle
91
3.3.4 Nature of direct object In the great majority of cases the direct object is a noun or pronoun. In the earlier texts, every example contains an overt nominal direct object. Cel is the first text to show any change in this situation, with one example: (60) ¿Mías, señora? Antes ajenas, como tengo dicho (Cel 93) ‘Mine, my lady? Rather other people’s, as I have said’ Here, although there is no overt direct object, one is implied. Similarly with two examples from Ter. (61) Mas como ya tengo escrito a vuestra merced bien largo…(Ter 2.3) ‘But as I have already written to you at great length…’ (62) …como a vuestra merced más largo tengo escrito (Ter 2.3) ‘…as I have written to you at greater length’ (On tener escrito see the comments in chapter 2: informants agreed that it implied tengo escrito algo.) But in Ter we find the first clear examples of a direct object other than a nominal one: (63) lo que vuestra merced tenía concertado hacer (Ter 29.4) ‘what you had decided to do’ (64) la limosna que tiene determinado a hacer (Ter 40.3) ‘the alms which you have decided to give’ This trend is continued in Qui, where four of the eighteen examples are followed by a preposition+infinitive or a que complement: (65) tengo determinado de volver a servir a mi amo don Quijote (Qui 73) ‘I have made up my mind to go back and serve my master don Quixote’ (66) tenía creído que su señor no se iría sin él (Qui 89) ‘he had believed that his master would not go without him’ In Fei and Jov there are a couple of examples with no overt direct object: (67) En orden al ejercicio, ya tengo dicho (Fei 70.7) ‘With regard to practice, I have already spoken [of this matter]’ (68) no pudo asistir el señor Herreros, a quien yo tenía hablado (Jov 139.246) ‘señor Herreros, to whom I had spoken, could not be present’ STP and CHM show no new developments. In this area grammaticalisation has not proceeded very far. But such movement as there is away from the obligatory presence of a nominal direct object occurs, as one might expect, in the later texts, beginning with Cel, which is in line with an overall process of grammaticalisation.
Tener+Past participle
92
3.3.5 Word order There is a clear preference for the order ‘tener+past participle’, as the figures indicate. The only possible deviation from this pattern occurs in Cor, and there are so few examples here that it is impossible to draw any firm conclusions. In her work on haber+past participle Concepción Company (1983, 249) suggests that as the order ‘haber+past participle’ becomes standard, the reverse order is limited to very specific contexts: exclamations and short phrases (usually at the beginning of a phonic group), and phrases which sum up everything which has just been said. This would account for the examples found in Qui, the one example in CHM, and some, but not all, of the earlier examples. In general, the order is ‘tener+past participle’, and there may be some movement towards the fixation of this order, although as overall there are so few examples it is hard to draw firm conclusions. However, even if ‘past participle+tener’ does not necessarily die out, it apparently becomes limited to specific contexts, which may in itself be considered a reduction in its possible usage.
Table 3.1 Word order in tener+past participle construction Mil PCG LBA Cor Cel Ter Qui Fei Jov STP CHM Tener+past participle
23
14
17
3
14
22
15
9
15
8
2
Past participle+tener
1
0
1
2
1
1
3
0
0
0
1
3.3.6 Interpolation The information regarding the interpolation of elements between tener and the participle is set out in Table 3.2. (Examples of ‘past participle+ tener’ are not included.) 5 From these figures we can see that there is a movement towards greater cohesion of tener and the participle until the Golden Age and eighteenth-century texts, but then this move seems to die away. Throughout the period under examination it is possible for them to be separated. If we look at the sort of element which can be placed between tener and the participle we find that a variety of elements can be interposed: subject and object nouns or pronouns, and adverbs. In most cases the interposed element is fairly short—a single pronoun, or a short adverb—although at times a slightly longer element or combination of elements is found: (69) …qe tenié so amor en Dios bien encendida (Mil 730 b) ‘whose love of God burned strongly’ (literally ‘who had his love of God well kindled’) (70) según tiene mi lengua y sentido ocupados y consumidos (Cel 75) ‘as it holds my speech and my reason captive and oppressed’ It would be interesting to see if the distance separating tener from the participle decreases over the period covered by the texts. However, given the small number of cases of separation, and the fact that in all the texts the interposed element is usually short, it is not possible to say whether there is a tendency towards less lengthy separation of tener
Historical development of tener+past participle
93
and the participle. What can be stated is that although examples which are semantically and aspectually amongst the furthest evolved may not exhibit lengthy separation of tener and the participle, by no means is interpolation totally absent from such examples.
Table 3.2 Interpolation of elements between tener and the participle Mil PCG LBA Cor Cel Ter Qui Fei Jov STP CHM No interpolation
9
6
10
3
11
16
11
6
12
3
1
At least one element between tener and the participle
14
8
7
0
3
6
4
3
3
5
1
A final point is that if we include in our study the second participle in the ‘twoparticiple’ constructions (see note 5) we obtain a different picture. Tener does not obligatorily appear with each participle, and the second participle may be well separated from tener. (71) ¿Piensas que no tengo sentidas tus pisadas y entendido tu dañado mensaje? (Cel 96) ‘Do you think that I haven’t been aware of what you are doing and understood your evil message?’ (72) los que tenían méritamente granjeada y alcanzada gran fama (Qui 65) ‘those who had deservedly gained and acquired great fame’ Unlike interpolation, two-participle constructions do not become less frequent during the Golden Age, nor are they restricted to the most lexical examples of tener+past participle. However, there are no examples of two-participle constructions amongst those which— according to factors such as semantics, tense/aspect, type of direct object—are most grammaticalised. And if one looks at the texts from the eighteenth century onwards, one finds a distinct drop in the number of two-participle constructions. (In the last four texts there is only one example.) 3.4 STUDY OF HABER +PAST PARTICIPLE 3.4.1 Outline development From the earliest texts haber is well ahead of tener in its development as an auxiliary verb. The most obvious indication of this is frequency. As the following table shows, haber+past participle is consistently much more common than tener+past participle. The way in which the constructions are used also points to the greater grammaticalisation of haber. We have seen that, almost exclusively in the early texts and to a large extent in the later ones, tener retains its durative character; the temporal reference of the construction is the same as that which would be indicated by tener as a main verb; the participle is semantically compatible with the semantics of lexical tener. We have also seen that tener
Tener+Past participle
94
almost always appears with the participle of a transitive verb, which agrees with the nominal direct object.
Table 3.3 Relative frequency of tener+past participle and haber+past participle Mil
PCG LBA Cor Cel Ter Qui Fei Jov STP CHM
tener+past participle
24
14
18
5
15
23
18
haber+past participle
112
112
41
24
132 347
177
9
15
157 298
8
3
264
167
In contrast, haber+past participle usually indicates past action rather than resultant state. This is so from the earliest texts. For example: (73) si don Adam oviesse de tal fructo comido de tan mala manera non serié decibido (Mil 15 b, c) ‘if Adam had eaten of such fruit he would not have been deceived in the same way’ (74) Yo nunca te tollí valía de un grano e tú hasme tollido a mí un capellano (Mil 229 c, d) ‘I never took the slightest thing from you and you have taken a priest from me’ (75) el grand danno que auien fecho los d’Affrica en Espanna (PCG 16, 44–5L) ‘the great damage which the men from Africa had wrought in Spain’ Haber also appears with a wider range of participles. In many examples haber has no possible lexical interpretation, owing to the nature of either the participle or the object. (76) el mal qe he passado contar no lo podría (Mil 296 b) ‘I could not recount the suffering which I have undergone’ (77) mas non se te oblide lo qe te he mandado (Mil 485 d) ‘but do not forget what I have commanded you’ However, in the early texts haber+past participle is occasionally used lexically, for example to indicate inalienable possession. (78) avié con el desarro las piernas embargadas (Mil 438 b) ‘with the fright her legs were paralysed’ (literally ‘she had paralysed legs’) Compare also: (79) e tenie escripto en la palma: fasta aqui llego Hercules (PCG 8, 45–6R) ‘it had written on its palm: Hercules reached here’ (80) e avie escripto en la palma: estos son los moiones de Hercules (PCG 8, 12–13R) ‘and it had written on its palm: these are the boundary markers of Hercules’ These lexical examples of haber+past participle are found only in the first two texts. After this point only tener+past participle is found used in this way. Again in the earlier texts, haber+past participle sometimes has a durative value similar to that of tener+past participle. In such examples the subject of haber and the agent of the
Historical development of tener+past participle
95
participle are not always the same. This use is found more frequently in PCG than in the earlier Mil. It seems that although syntactically haber+past participle is more grammaticalised in PCG than in Mil—see below for comments on participle agreement and word order—semantically and aspectually the reverse is the case. While there are plenty of examples of haber+past participle as a perfective form of past tense in PCG, there are also durative ‘tener type’ examples. (81) E fizieron por toda la tierra cantares de llantos dEspanna que dizien que Dios la auie ayrada (PCG 14, 11–13L) ‘And throughout the land they performed laments for Spain, for they said that God was angry with her’ In some cases it is not easy to ascertain whether haber+past participle expresses a durative state, or a completed action, or both. For example, in (82) it could be argued that haber+past participle is used in a similar way to the pluperfect -ra form, or that there is a contrast between them and that haber+past participle indicates resultant state or achievement rather than action. (82) e aquel fiziera la puente, e auie tod el canno fecho pora traer ell agua (PCG 12, 26– 7L) ‘and he had built the bridge, and had built all the conduit/had got all the conduit built to bring the water’ However, among the texts here examined, this durative or resultant state use of haber+past participle seems to be a particular characteristic of PCG and is not found at all in any of the later texts. 3.4.2 Había +past participle and the -ra form We have seen that one possible way of testing for grammaticalisation is to see whether or not there is any other form which has the same function. In the earlier texts había+past participle is in competition with the -ra pluperfect. As the following example suggests, the two forms could in many cases be used interchangeably as far as the meaning is concerned: (83) ca no.1 tollieran nada ni.1 avién ren robado (Mil 195 b) ‘for they had not taken anything from him, nor had they robbed him of anything’ Gradually había+past participle takes over from the -ra form to become the only pluperfect, while -ra evolves as an alternative pluperfect subjunctive to the -se form. Figures from the texts are shown in Table 3.4: 6
Table 3.4 Comparative frequency of haber+past participle and -ra form (pluperfect and subjunctive uses) había+past participle
-ra form pluperfect
conditional/ subjunctive
Tener+Past participle
96
Mil
44
25
6
PCG
39
211
6
LBA
12
6
17
Cor
6
4
24
Cel
3
0
59
In all texts apart from PCG, había+past participle is the more common of the two pluperfects. The fact that the -ra form is so much more frequent than había+past participle in PCG may be linked to the tendency for había+past participle to retain some of its durative value in this text (see above). From LBA onwards the -ra form is used increasingly in conditional constructions and less and less as a pluperfect. 3.4.3 Haber +past participle and ser +past participle Another partial rival to haber+past participle is to be found in ser+ past participle. The rivalry is partial in that ser is used as an auxiliary only for intransitive and reflexive verbs. As Table 3.5 shows, haber+ past participle is increasingly used in these contexts at the expense of ser +past participle, until eventually haber+past participle becomes the only possible perfect construction for all verbs. Use with intransitive and reflexive verbs:
7
Table 3.5 Relative frequency of haber and ser with intransitive and reflexive participles Mil
PCG
LBA
Cor
Cel
Ter
Qui
haber+past participle
9
3
3
5
21
93
55
ser+past participle
42
19
17
5
6
2
0
Both these results and those dealing with the -ra form indicate an increasing integration of haber+past participle into the verb paradigm as it ousts alternative forms and as haber widens its range of cooccurrence to include all verbs. 3.4.4 Agreement of past participle and direct object Here again haber+past participle is quickly ahead of tener+past participle. From the earliest texts there are examples of haber with an invariant participle (although the three examples of non-agreement in Mil all occur at the end of a line and could thus be influenced by questions of assonance). From Cor there are no examples in the texts of the participle agreeing with the object. 8
Historical development of tener+past participle
97
Table 3.6 Agreement in haber+past participle construction Mil
PCG
LBA
Cor
Agreement
43
32
7
0
No agreement
3
13
3
4
3.4.5 Word order Although there is a certain amount of variation in the earlier texts, there is an increasing use of ‘haber+past participle’ as ‘past participle+ haber’ becomes less and less common (see Table 3.7).
Table 3.7 Word order in haber+past participle construction Mil PCG LBA Cor Cel Ter Qui Fei Jov STP CHM Haber+past participle
107
111
36
19
Past participle+haber
5
1
5
5
123 340 9
7
177 0
157 298 0
0
259
167
5
0
Concepción Company’s classification (see 1983, 167) accounts for some, but not all, of the ‘past participle+haber’ examples in the texts up to Ter. In STP all five examples of ‘past participle+haber’ are of the type ‘past participle que hubo’. For example: (84) dejado que hubo el mozo el Seminario por el Cuartel (STP 28.10) ‘as soon as he had left the seminary for the barracks’ (85) Sentado que se hubo, Frasquita echo una pierna sobre la otra (STP 50.26) ‘As soon as she had sat down, Frasquita crossed one leg over the other’ This would seem to be a particular stylistic device used by Alarcón. If this set of examples is discounted, then there are no examples of ‘past participle+haber’ after Ter. The trend towards a more fixed word order is in line with greater grammaticalisation, although in this area grammaticalisation is slower than in other areas such as the ousting of the -ra pluperfect and loss of participle agreement. 3.4.6 Interpolation Not only is the rate of interpolation lower in the later texts than in the earlier ones; also, the length of what appears between haber and the participle decreases. In the early texts, although there are many very short interpolations, there are also cases such as (86) and (87). (86) si oviesse su lengua un poco retenido (Mil 70 b) ‘if he had held his tongue a little’
Tener+Past participle
98
(87) auie con su marido fecha poca morada (LBA 478 b) ‘she had lived with her husband for only a short time’
Table 3.8 Interpolation of elements between haber and the participle Mil PCG LBA Cor Cel Ter Qui Fei Jov STP CHM No interpolation
64
92
27
17
At least one element between haber and the participle
43
19
9
2
118 314 167 149 285 256 5
26
10
8
13
8
162 5
(Note 5 also applies here)
From Cel onwards virtually all the interpolated elements are very short—usually pronouns and short adverbs (e.g. ya, bien, hoy). Fei provides the only exception to that. For example: (88) …de quien hago grata memoria por haberle, siendo yo jóven, conocido obispo de mi diócesis de Orense; y conocido asimismo su amabilísimo genio (Fei 135.15–17) ‘…who I remember with pleasure, as I have, when I was young, known him as bishop of my diocese of Orense, and likewise known his very kindly disposition’ So here there is a further contrast—albeit not a very clear one—between haber and tener. While in both cases interpolation becomes less frequent, with haber+past participle the interpolated elements also tend to become shorter; there is no indication of a similar tendency with tener+past participle. 3.5 SUMMARY OF DEVELOPMENT OF TENER/HABER +PAST PARTICIPLE 3.5.1 Summary of development of tener +past participle So far we have traced the development of tener+past participle as indicated by various different criteria: semantic loss; changes in the temporal and aspectual value of the construction; syntactic criteria such as word order and agreement of past participle and direct object. According to some of these criteria, tener+past participle changes very little throughout the period, whereas other criteria indicate important developments. Syntactically, little changes: at the end of the period under examination tener can still be separated from the participle in much the same way as it could in the earliest texts; the orders ‘tener+past participle’ and ‘past participle+tener’ are both found throughout the period; the participle continues to agree with the direct object, with very few exceptions. But if we look at the semantics of tener+past participle and its temporal and aspectual values we see that its use has expanded. In the earliest texts, examples were compatible with lexical uses of tener. Seifert observes that the thirteenth century sees a marked expansion in uses of tener, as more and more it invades the area covered by haber. These new lexical uses of tener can also be found amongst the examples of tener+ past
Historical development of tener+past participle
99
participle. The initial expansion of tener+past participle can be ascribed to the expansion of lexical tener. if lexical tener increases its range of application, then it is unsurprising that tener+past participle does so too. This expansion of tener+past participle is thus not the result of grammaticalisation, but of lexical expansion. But gradually the expansion of tener+past participle takes on a different character, and later developments of tener+past participle arise not from lexical expansion, affecting all uses of tener, but from loss of lexical value of tener as used in the tener+past participle construction. This second part of the expansion of tener is limited to the tener+past participle construction, and can be viewed as grammaticalisation. In this second stage, in some instances of the tener+past participle construction tener is essentially a marker of duration, and in certain cases even this element is lost. Tener takes over from haber as a lexical verb of possession, and also begins to develop as a perfect auxiliary. But while it eventually ousts haber from its lexical role, it has little impact as a rival to haber’s auxiliary role. The two developments—semantic expansion and an incipient auxiliary function—are presumably not entirely separate from each other. Increased use of tener as a main verb brings with it increased use of tener+past participle (even though one might still regard tener as a main verb in many of these cases) and from this gradually develop the more auxiliary type uses of tener. During the analysis of tener+past participle I suggested that tener gradually loses its specific semantic content of possession or of actively keeping something in a particular state, and can in some cases be essentially an aspectual marker of duration. This is very similar to its copular function (see Seifert, 1930, pp. 256, 266). 9 At this stage it is a present tense marker of a current state. Gradually it acquires a possible resultant state interpretation, in which there is some indication of previous action, even if the emphasis is still primarily on the state. As already mentioned, Yllera suggests a further development, which either precedes or overlaps with the early stages of resultant state tener+past participle, in which tener+past participle expresses a durative or iterative action which has existed in the past and is still going on at the present moment (see the discussion of examples (16), (17), and (23) on pp. 100 and 101). While the texts examined here provide some possible examples of such a development, they neither strongly support nor refute such a theory. There is a further point to be made about the way in which tener+ past participle changes in use. As I suggested in the analysis of the history of the construction, it seems that when tener+past participle begins to acquire a new function, the first examples to be used in this way are ones which are compatible with an ‘older’ interpretation, i.e. there is no immediate introduction of completely new examples: rather, examples which are already possible appear with a slightly altered application. However, this is a very delicate area in which to work, as once again we are faced with problems of interpretation: both how something is to be interpreted, and whether something is a necessary part of the interpretation or merely incidental. One use seems to shade almost imperceptibly into another, and ‘lexical’ and ‘grammaticalised’ examples overlap, producing a vast number of possible degrees of grammaticalisation. 10 Other factors, such as word order or participial agreement, are much easier to measure. However, in the early stages of grammaticalisation at least, they provide less information than semantic considerations do. For example, if we look at participial agreement in the tener+past participle construction we find that there is virtually no change throughout the
Tener+Past participle
100
period under study here. Similarly with interpolation. This is not to say that such factors are not important but, as was suggested earlier (section 2.8), they seem to come into play at a later stage. They do not contradict any posited grammaticalisation, but neither do they provide positive support for it. Let us now attempt to sum up the development of tener+past participle. In the first texts all the examples contain either a literal value for tener or, at their least lexical, indicate duration; this durative value is similar to the copular function which could be exercised by tener when not combining with a participle. In the first four texts there is little or no movement away from this position. But from Cel onwards, lexical tener becomes less important. This is not to say that examples of lexical tener +past participle no longer occur, but that there is an increased number of examples where a lexical interpretation of tener is either irrelevant or impossible. The decreased importance of lexical tener continues up to the contemporary language. As was pointed out earlier, deciding where the lexical value is significant and where it is merely incidental is not an easy task. However, Table 3.9 provides a rough guide to the development. More or less contemporary with the onset of this development is the appearance of examples which focus on past action rather than current state. An approximate indication of the development also appears in Table 3.9.
Table 3.9 Change in use of tener+past participle Mil PCG LBA Cor Cel Ter Qui Fei Jov STP CHM Total
24
14
18
5
15
23
18
9
15
8
3
No possible/likely lexical interpretation
9
1
8
0
10
19
15
7
13
5
3
Concentration on past action
0
0
0
0
3
6
5
2
2
0
0
These figures suggest that, unlike the decline of lexical tener+past participle, the frequency of tener+past participle as a form of past tense is not maintained. In these texts it first appears in Cel, continues during the Golden Age and then dies down. This conclusion would seem to tie in with Seifert’s (1930, 382) opinion that ‘En el siglo XVI tiende tener a establecerse como auxiliar verdadero. Continúa esa tendencia en la época de Cervantes…para luego caer en desuso conservando la lengua haber.’ But what exactly does this falling away consist of? If we consider the frequency of the construction, then it does indeed appear to die away, although it clearly does not disappear. Nor is its use restricted entirely to cases of lexical tener+past participle. There do indeed seem to be fewer examples of ‘past tense’ tener+past participle. Tener+past participle has become, above all, an expression of duration and of resultant state (to that extent the ‘past action’ element has not been lost). Once again, we see the essentially durative nature of tener+past participle coming to the fore. We saw in chapter two that tener+past participle can still have a past-action interpretation. But while the survey discussed there showed the possible uses of the construction, it gave little indication of their relative frequency. An approximate idea of relative frequency can be formed by examining the examples of tener+past participle to be found in transcriptions of conversations recorded for a study of the contemporary
Historical development of tener+past participle
101
speech of Madrid (Esgueva and Cantarero 1981). Of these forty examples, twenty-one apparently express resultant state (with, of course, varying degrees of emphasis on state and action). Of the remaining nineteen examples, fifteen indicate some form of present tense, including expressions of inalienable possession, of keeping something in a particular state, and use of tener as a copula. That leaves four possible candidates for ‘past action’ tener+past participle. Two points should be made about these four examples. Firstly, that they all involve the same participle: oído. So in that sense the construction appears to be quite restricted. Secondly, although there is no resultant state in these examples, the degree of current relevance which they display is greater than in many earlier examples of ‘past action’ tener+past participle. Tener oído seems to be nearly equivalent to tener entendido. Compare, for example: (89) Y encima, llegarás casi con la cara negra, porque, tengo oído que como en Madrid hay tantísima porquería…(Esgueva and Cantarero, 1981, p. 346) ‘And on top of that, when you get there your face is almost black, because I’ve heard that because Madrid is so dirty…’ (90) Pero yo tengo entendido que, que en Persia y luego en los países árabes que en fin, que se casa la gente muy joven, pero no sabía que…(p. 379) ‘Because I’ve heard [literally ‘understood’] that, that in Persia and in Arab countries, well, that people get married very young, but I didn’t know that…’ So it would seem that in modern Spanish ‘past action’ tener+past participle is to be found, but it is less frequent, more restricted, and perhaps more closely linked to the present than it was in Golden Age Spanish. While the construction retains the various possible uses which it had at the height of its development, some of the more grammaticalised of these uses have become less frequent. Tener+past participle appears to be used principally as a copular construction or as an expression of resultant state. 3.5.2 Summary of development of haber +past participle From the start of the period under consideration here there are already clear indications of haber’s role as an auxiliary verb. 11 During this period this role is strengthened, while lexical use of haber ceases altogether. Loss of haber as a lexical verb means that the more lexical examples of haber+past participle (as in examples (78) and (80)) also come to an end. Semantic loss is not the only indication of grammaticalisation. Morphosyntactic developments also point to increased grammaticalisation. Although in the early texts there are a few examples of invariant participles, as a general rule the participle agrees with the direct object. However, from Cor onwards the participle is always invariant. There is also a trend towards a fixed word order of ‘haber+past participle (+ direct object)’. However, it is still possible for elements to be placed between the auxiliary and the participle, although this becomes less common. We have seen that it is more helpful to view some linguistic phenomena in terms of continua rather than of discrete categories, and we have considered the framework provided by Lehmann. But as well as needing to place constructions relative to each other on a scale—constructions which admit of no change in word order being more grammaticalised than those where items can be moved
Tener+Past participle
102
around; affixes being more grammaticalised than independent items, and so on—we also need to take the frequency of such factors into account. For example, the fact that something can be placed between haber or tener and the participle indicates a relatively low level of grammaticalisation. But in other constructions where fusion has taken place (e.g. the synthetic future arising from the fusion of infinitive+haber) such a development cannot just suddenly happen; it presumably occurs as more and more often the two elements are placed immediately next to each other, even though throughout this stage it is still possible for them to be separated. So while separation of the two elements may be possible at various points in the history of a construction, it is more likely at some points than at others. We cannot place a construction at a certain point on a continuum purely according to whether or not a particular factor is present or absent; we also need to consider the extent to which that factor is present. Although separation of haber from the participle is still possible, the fact that it has become less common is an indication of greater grammaticalisation. Haber+past participle also displaces the -ra pluperfect and the ser+ past participle perfect. This points to a fuller integration of the paradigm into the verbal system. All the factors which have been examined indicate the increased grammaticalisation of haber+past participle. 3.5.3 Comparison of tener and haber One of the main difficulties associated with tracing and measuring the grammaticalisation of tener+past participle is that this construction does not follow one single path, nor do all its developments lead towards greater grammaticalisation. Tener expands its uses both as a main verb and as a possible auxiliary. Some of the tener+past participle uses relate to the expansion of tener as a main verb; others relate to the grammaticalisation of tener. The period studied here sees an increased variety of uses of tener+past participle which cannot all be grouped together. It is easier to trace an overall pattern of grammaticalisation for haber. Haber is used less and less as a main verb indicating possession, and is increasingly used as an auxiliary. This development can be measured in semantic and syntactic terms. It might be tempting to see the history of haber and tener in terms of a ‘sharing out of tasks’: after a period of rivalry tener takes over entirely from haber as a verb of possession, while haber repels any attack from tener on its new position as an auxiliary. However, while it is fairly clear that tener becomes the only one of the two verbs capable of expressing possession, the situation regarding auxiliary usage is less clear-cut. In terms of grammaticalisation, haber is ahead of tener in every area we have looked at, and in many of these areas it moves even further ahead of tener as the development progresses. But while haber is clearly more grammaticalised than tener, we cannot say that haber appropriates all the auxiliary uses. And while tener has some non-lexical uses, it is not clear that these are uses which it has taken over from haber. Pinkster (1987, 215) notes that Latin TENERE+past participle is much less common than HABERE+past participle, and that while HABERE is already on the way to becoming an auxiliary at this stage, the TENERE construction nearly always means ‘to keep someone/something in a certain physical/psychic condition’. Pinkster’s observations support the view that haber is always more grammaticalised than tener and also suggests that the idea of keeping
Historical development of tener+past participle
103
something/someone in a particular state is characteristic of tener but not of haber. (Pinkster notes that there is only a partial semantic overlap of the two verbs.) From the earliest texts examined in this study tener+past participle is used to indicate duration. Gradually, the semantic content of tener in such cases is eroded until it is just a marker of duration. This indicates a movement towards greater grammaticalisation—as measured by some, though not all, of the criteria used. At no stage is such a use so grammaticalised as the auxiliary use of haber according to any of the criteria. (And we saw in chapters 1 and 2 that in the modern language tener+past participle passes fewer of the tests for auxiliary status than haber+past participle does.) But this use is one which is particularly characteristic of tener. We saw in the texts that it is only in PCG that haber regularly has this durative usage. So tener lays virtually undisputed claim to one area of at least semigrammaticalised usage. The reason for this may well lie in the aspectual value of haber and tener as main verbs. Seifert points out that although both verbs are essentially durative, there is a tendency for tener to be used where duration is particularly emphasised, and for haber to express inchoative aspect. Seifert stresses that this is no more than a general rule, and that tener and haber do overlap considerably (and of course, tener eventually takes over haber’s use as a main verb). However, this tendency seems to be accentuated in past participial constructions, with tener quickly establishing and then sustaining a use as a marker of duration. Although haber is clearly the more grammaticalised of the two verbs and has the more obvious auxiliary role, tener is not completely absent from this field. We have seen that tener develops other partially grammaticalised auxiliary functions, not all of which are fully sustained: it is used to express resultant state; iterative or durative situations beginning in the past and continuing in the present; and occasionally it refers to a single event in the past. In this respect it appears to follow the same path as haber, without proceeding so far along it. But we have seen that there is at least one important difference between the two verbs, in that tener has this particular durative use. Earlier I suggested that the other uses of tener+past participle develop from this use. We also saw in chapter 2 how contemporary uses of the construction can be seen to stem from the notion of duration. If this is the case then tener+past participle does not just follow the path of haber+past participle, although—unsurprisingly, given the considerable overlap between the two verbs—the two developments have much in common.
3.6 COMPARISON OF THE HISTORICAL DEVELOPMENT WITH MODERN VARIATION How closely does the historical evolution of tener+past participle correspond to the synchronic variation between contemporary speakers? We have seen how historically the first examples of tener+ past participle are compatible with lexical tener in that they indicate that the subject possesses something or is maintaining it in a particular state; this notion then weakens to an expression of duration; later there is a shift in emphasis towards the causative action, and past durative or iterative examples appear. This development corresponds closely to the order put forward in chapter 2 for the relative degree of acceptance of the different semantic and aspectual interpretations of tener+past participle and also for the hierarchy of likely interpretations. All speakers accepted
Tener+Past participle
104
lexical and durative examples. The further the examples moved towards a past tense interpretation, the fewer speakers there were who accepted them. Also, where there was a current state interpretation, this took precedence over a past action one. Syntactically, too, historical development and synchronic variation coincide. Tener+past participle first appears with a nominal direct object, then with a sentential or infinitival complement, or with an implicit direct object. This corresponds to the hierarchy of acceptance by contemporary speakers. (In the survey of modern usage there are also examples of intransitive verbs with certain adverbial quantifiers. No similar examples were found in the historical survey. However, in the contemporary survey they were found only in more grammaticalised usage.) Participle agreement provides us with little evidence, as in both the historical and the contemporary investigation very few examples of non-agreement arose. But in both cases, when non-agreement did appear it was in examples which other criteria indicated were among the more grammaticalised ones. Questions of word order and interpolation are largely matters of frequency rather than of the presence or absence of a particular word order. As the contemporary survey dealt only with what is possible and did not attempt to measure frequency, no comparison can be made here between the two surveys. However, the fact that varying word orders and insertion of elements between tener and the participle are possible throughout the history of the construction supports the view expressed in chapter 2 that Lehmann’s parameters of bondedness and syntagmatic variability come into play later than the other parameters. In so far as the two surveys can be compared, such a comparison supports the view that variation in the modern language can be hierarchically ordered in a way which corresponds to the historical development of tener+past participle. There are also some more general comparisons to be drawn between synchronic and diachronic studies of tener+past participle. In this chapter we have seen that while the historical development of tener+ past participle follows a broadly similar path to that of haber+past participle there are important differences between the two. In particular, the essentially durative nature of tener strongly influences the development of tener+past participle. Similarly, in the modern language, tener+past participle has some characteristics in common with other constructions (notably haber+past participle and estar+ past participle), but once again, it cannot be accounted for simply in terms of these other constructions (see sections 2.9 and 5.3). Both the synchronic and the diachronic surveys lead us to the view that tener+ past participle enters the verb system on its own terms, and maintains its own internal motivation, not adjusting to fit into pre-existing slots in the verb system. Unsurprisingly, it has characteristics in common with other constructions, both in terms of its general pattern of grammaticalisation (for example, in the way it conforms both synchronically and diachronically to Lehmann’s parameters) and more specifically in the similarities it exhibits to other verbs which move from expressing possession to some form of perfect auxiliary (for example, haber+past participle). But in spite of these similarities, both synchronically and diachronically tener+past participle must be dealt with on its own terms, not as a poor copy or a replacement for another existing construction. In the synchronic and the diachronic studies we come across similar problems of interpretation: not only of deciding what a particular example means, but also of discovering which elements of the context are essential to the interpretation and
Historical development of tener+past participle
105
acceptability of that example and which are merely incidental. With the contemporary language it is possible to question native speakers and to elicit all the necessary information. This of course cannot be done with studies of previous language states. But where synchronic variation and diachronic developments pattern in the same way we can cross-reference between the two (see Labov 1972a). Before this can be done we need enough independent information on variation and development to be sure that the two do pattern in the same way. But once this has been ascertained, then information from contemporary studies can be used in historical surveys. We should be wary of taking this approach too far and of projecting information about the modern language onto historical developments where there is no independent confirmation of such a development, but our more complete knowledge of the contemporary language can at least highlight possible lines of historical research and help to make sense of incomplete historical data. This is an additional reason for looking for ordered variation in contemporary language, rather than idealising to a uniform state by ignoring differences between speakers or by ‘ironing out’ these differences to produce a description which represents the lowest common denominator of all speakers.
4 Historical development in Portuguese 4.1 HISTORICAL SURVEY In chapters 2 and 3 we saw how a comparison can be made between the current use of tener+past participle in standard Spanish and the historical development of the construction, and how each can help to provide a clearer understanding of the other. We now turn to a comparison of the evolution of Spanish tener+past participle and Portuguese ter+past participle. The evolution of Portuguese ter+past participle is rather better documented and discussed than that of its Spanish cognate—probably because, in contrast with developments in Spanish, ter+past participle was eventually victorious over haver+past participle, and is now the usual perfect auxiliary construction. Before embarking on a textual analysis, I shall outline the most generally accepted views on this evolution. In medieval Portuguese both ter+past participle and haver+past participle are found. Initially, both constructions appear only with transitive verbs, and the participle agrees with the direct object. Mattos da Silva (1981) suggests that the two constructions are in free variation until the fifteenth century, when ter begins to displace haver. Naro and Lemle (1977), on the other hand, believe that haver functioned as an auxiliary ‘since the earliest attested stages of the language’, and that later on, ter catches up and eventually ousts haver. The two articles agree that ter gradually displaces haver which, in the modern language, is restricted to highly literary usage. In both constructions the participle ceases to agree in every case with the direct object, and intransitive verbs begin to appear in the construction. This development expands until ter (and, within the already mentioned restrictions, haver) comes to be used with the participle of all verbs. This development is very similar to the one outlined in chapter 3 for Spanish haber+past participle. However, two major differences can be pointed out: firstly, ter appears both with an invariant participle and with an agreeing participle, whereas the general rule in Spanish is for haber to appear with an invariant participle and tener to appear with an agreeing participle; secondly, while ter+ agreeing participle and tener+agreeing participle may be thought of as broadly equivalent constructions, ter+invariant participle is by no means the same as haber+past participle. The present tense forms of ter +past participle have a durative or iterative value not shared by haber +past participle. Examination of texts may provide information on the development of this value. It may also indicate how the separation between the invariant and the agreeing constructions occurred. The survey of Portuguese texts which follows was carried out in the same way as the examination of Spanish texts described in chapter 3. (Details of the texts used are to be found on page viii.)
Historical development in Portuguese
107
4.2 STUDY OF TER +PAST PARTICIPLE 4.2.1 Outline development This construction provides arguably the most interesting information of all the constructions considered in this study, in either Spanish or Portuguese, as it is possible to trace its development from the very early stages to its use as a virtually undisputed auxiliary. On the one hand we can chart its course from a much earlier stage than is possible with haber +past participle; and on the other hand, as it grammaticalises much further than tener+past participle, it provides more information on later stages than does its Spanish cognate. In the earliest texts ter has a literal interpretation. The only example in Dem is (1) Eu achei ora mortos dous cavaleiros e ũa donzela, que tiinha a cabeça cortada (Dem 116) ‘Just now I found dead two gentlemen and a lady, who had her head cut off’ The four examples from ODE can also be interpreted literally, although they do not express inalienable possession: (2) no cessarey porem de chamar o nome de Jhesu, porque o tenho scripto emno meu coraçom (ODE 10, 18) ‘but I will not cease to call the name of Jesus, because I have it written in my heart’ (3) tomou-lhe a cinta, que tiinha cingida aredor das rẽẽs (ODE 24, 38) ‘he took her sash, which she wore girded round her waist’ (4) uyo hũa muy grande pedra que tiinha per sy muytos furacos retornados (ODE 49, 36– 7) ‘he saw a very big stone which had several holes bored in it’ (5) mostrou-lhe rey Poro todos seus thesouros que tiinha escondidos (ODE 62, 19–20) ‘king Poro showed him all his treasures which he kept hidden’ These are the only examples to occur in the sample. However, Naro and Lemle (1977, 265) indicate that in the text as a whole ‘in a very small number of examples the meaning of a compound tense, rather than the complement construction, can be felt’. If this is so, then there is some indication of a move away from purely lexical ter+past participle. By RDJ this movement has developed considerably. While it is not always easy to pinpoint the exact interpretation, at a conservative estimate approximately half of the thirty-six examples have a past action interpretation rather than any literal, durative, or copular value. For example: (6 El-Rey respomdeo que lho gardeçia muyto, dizemdo que aquella era sua vontade, ajnda que lho ataa entom nom teuesse dito (RDJ 13) ‘The King replied that he took careful note of it, saying that that was his wish, although until then he had not said so’ (7) …temdo ja dias auia mandado com recado a el-Rey de Castella Gomçallo Marinho (RDJ 25)
Tener+Past participle
108
‘…having days earlier sent Gomçallo Marinho with a message to the King of Castile’ In this text we find the first use of ter+past participle without a direct object. However, seven of the eight cases are in clauses following como or segundo, of the type: (8) por fallar ao Mestre…sobre as mercadorias que lhe forom tomadas acerca do porto da dita cidade nas duas naaos, como teendes ouujdo (RDJ 11) ‘to speak to the Master…about the goods which had been taken from him near the port of the aforementioned city from the two ships, as you have heard’ (9) Partio elRey de Cojmbra como tinha hordenado (RDJ 17) ‘The King left Coimbra as he had arranged’ (10) E mandaua…que se ueessem pera elle pera emtrar em Portugall pella parte de Badalhouçe, segundo tinha hordenado (RDJ 37) ‘And he ordered…that they should come to him in order to enter Portugal via the Badajoz region, as he had planned’ These can be discounted if one argues that they contain an implicit direct object. The final example is: (11) os portugueses tijnham mandado a Ingraterra por huma grossa ajuda de gentes (RDJ 55) ‘the Portuguese had sent to England for a large contingent of men’ It can be argued that in this example too there is an implicit direct object. The first convincing examples of ter with the participle of intransitive verbs occur in Vie, where nine of the sixty-one examples involve intransitive verbs and three involve reflexives. For example: (12) Tinham êles saído na Ilha de Taparica (Vie 29) ‘They had departed from the Island of Taparica’ (13) …nos lugares onde havia pouco tinham estado (Vie 47) ‘…in the places where they had been a short while ago’ (14) …não se tendo achado o dia de antes no assalto (Vie 59) ‘…not having been [literally ‘found himself’] the day before in the attack’ (15) tendo-se uma vez apartado da graça divina (Vie 67) ‘having once strayed from divine grace’ This development continues in later texts, as the auxiliary use of ter spreads through the language, until it can be used with all verbs. The ‘past action’ development also continues. I shall leave until later a discussion of what type of past action ter+past participle denotes, simply making a broad distinction here between a present ter+past participle construction (encompassing literal and resultant state interpretations) and a past ter+past participle construction. In VME sixtyeight of the seventy-five examples involve the past construction. For CPA the figure is forty-two out of forty-five, and for Vin, twenty-nine out of thirty-four.
Historical development in Portuguese
109
4.2.2 Agreement of past participle and direct object Unsurprisingly, in the earlier texts, the overwhelming tendency is for the participle to agree with the direct object. As with the introduction of intransitive verbs, it is not until Vie that any significant change occurs (see Table 4.1 and chapter 3, note 8). From Vie onwards we see a split between the past tense and the current state/lexical constructions. It was suggested above that in terms of tense, aspect, and semantics the split begins much earlier. But it is not until Vie that there is any formal indication of such a split, with the past tense construction increasingly having an invariant participle, and the less grammaticalised construction continuing to take an agreeing participle. Of the nine cases of agreement in Vie, only three are past constructions: the other six involve lexical or current state values of ter+past participle. From VME onwards all examples of agreement involve this lexical or current state value, and the past construction always has an invariant participle.
Table 4.1 Agreement in ter+past participle construction Dem Agreement No agreement
ODE
RDJ
PJ
Vie
VME
CPA
1
3
18
8
9
4
1
1
0
1
0
25
20
13
15
0
1
Vin
Here we see clearly the link between agreement of the participle and grammaticalisation. There is a move from a situation in which the participle always agrees with the direct object and the construction is primarily lexical, through a stage where the participle begins not to agree with the object when there is a past interpretation, leading to a division in which the literal/resultant state construction has an agreeing participle and the past construction has an invariant one. 4.2.3 Word order There is a distinct preference for the order ter+past participle throughout the texts, as indicated by the figures in Table 4.2.
Table 4.2 Word order in ter+past participle construction Dem
ODE
RDJ
PJ
Vie
VME
CPA
Vin
Ter+past participle
1
4
26
12
60
45
75
34
Past participle+ter
0
0
10
1
1
0
0
0
The high proportion of examples of the past participle+ter order in RDJ seems to be due to a particular stylistic device of the author. All the examples exhibit the same pattern, with the construction appearing at the end of a relative clause. For example: (16) os priuillegios e liberdades que lhe dados tinha (RDJ 9)
Tener+Past participle
110
‘the privileges and liberties which he had given him’ (17) …que lhe quitasse a menagem que por o logar feita tinha (RDJ 26) ‘…that he should call off the siege that he had laid on the place’ 4.2.4 Interpolation
Table 4.3 Interpolation of elements between ter and the past participle Dem
ODE
RDJ
PJ Vie
No interpolation
0
3
19
7
At least one element between ter and the participle
1
1
7
5
VME
CPA
Vin
40
39
49
26
20
6
26
8
(Examples of past participle+ter are not included. As in chapter 3, only the first participle in multiparticiple constructions is considered: see chapter 3, note 5.)
It may seem that the figures for this phenomenon given in Table 4.3 are far from conclusive, in that there is no progressive decrease in the number of cases where ter is separated from the participle. However, some conclusions can be drawn from the information on word order and interpolation. In modern Portuguese the standard word order for the invariant past tense construction is ‘ter+past participle+direct object’. The agreeing resultant state construction may be ordered as ‘ter +past participle+direct object’ or as ‘ter+direct object+past participle’; some writers suggest that the latter is the more usual order. If we look at instances of ‘ter+nominal direct object+past participle’ in these texts we find that they account for only twenty-two out of a total of seventy-four cases where ter is separated from the participle. But all twenty-two instances involve lexical or resultant state examples: i.e. there are no examples of the past construction with the order ‘ter+ nominal direct object+past participle’. It can of course be argued that in the earliest texts all examples of’ ter+direct object+past participle’ would have to be lexical or resultant state examples, as there are no past action examples in these texts. But the trend continues in RDJ and later texts where such examples do occur. This suggests that the distinction in word order is drawn fairly early on in the development of the split between the lexical/resultant state and the past constructions, and that while even in more grammaticalised examples ter can readily be separated from the participle, in such examples separation by a nominal direct object does not generally occur. The figures also suggest that for lexical/resultant state examples the order ‘ter+nominal direct object+past participle’ is more usual than ‘ter+past participle+nominal direct object’, with twenty-two examples and nine examples respectively being found in the texts (as compared with fifty-one examples of ‘ter+past participle+nominal direct object’ for the past construction). No similar conclusions can be drawn for word orders involving pronouns, as weak object pronouns are usually attached to the end of the auxiliary (although in certain circumstances they precede the auxiliary). The position of weak object pronouns in this construction is decided according to syntactic constraints rather than interpretation. Normally, they follow the verb, but they usually precede it in negative and subordinate
Historical development in Portuguese
111
clauses, and in clauses introduced by certain short adverbs. And so we find examples such as: (18) Tinha-a tomado por piedade (CPA 28) ‘She had taken her in out of pity’ (19) Talvez o dr. Bruno o tenha conhecido (Vin 68) ‘Perhaps Dr Bruno may have known him’ (20) No dia seguinte tinha-os ella embrulhados n’um papel (CPA 85) ‘The next day she had them, wrapped in paper’ (21) Enquanto ao sobrinho, devia de morrer na briga, porque o não tinha preso (Vie 60) ‘As for the nephew, he must have died in the fighting, because he hadn’t taken him captive/was not holding him captive’ While these rules for weak pronoun position apply to all auxiliary and periphrastic constructions, and are not specific to haver/ter+past participle, the fact that pronouns frequently separate the auxiliary from the participle is an indication of a low level of grammaticalisation as measured by Lehmann’s criterion of bondedness. (This point will be taken up again during the comparison between Spanish and Portuguese in section 4.8.) Particularly in the later texts only very short elements separate ter from the participle in the past construction. In VME there are four cases where ter is separated from the participle in a past construction. In two examples a short subject pronoun comes between ter and the participle; in the other two cases they are separated by a short adverb. In CPA the twenty examples of separation are made up of fourteen cases of pronouns and six cases of short adverbs. In Vin the three cases of separation involve se or short adverbs. In these three texts, of the fifteen examples of the lexical/resultant state construction, all thirteen examples where ter is separated from the participle have the order ‘ter+ direct object+participle’. The two ter+past participle constructions are thus differentiated not only by interpretation and participial agreement but also by word order. 4.3 STUDY OF HAVER +PAST PARTICIPLE 4.3.1 Outline development In the early texts haver+past participle seems to be ahead of ter+past participle, in terms of frequency at least, although later on it falls back and eventually all but disappears. However, Mattos da Silva’s (1981, 98) figures, given below, produce rather a different picture.
Table 4.4 Relative frequency of ter+past participle and haver+past participle Dem
ODE
RDJ
PJ
Vie
VME
CPA
Vin
Ter+past participle
1
4
36
13
61
45
75
34
Haver+past participle
33
11
15
4
12
0
0
3
Tener+Past participle
112
Table 4.5 Mattos da Silva’s assessment of the relative frequency of ter and haver DSG 1
DSG 2
LRR
CDP
Ter+past participle
34
34
1
24
Haver+past participle
5
5
32
18
Mattos da Silva uses the fact that two of the texts (or rather, two versions of the same text) show a preference for ter, one for haver, and one no clear preference for either, to argue that the two constructions are in free variation. While the figures may indicate that we need to be wary of drawing firm conclusions as to relative frequency, it does not follow that the two constructions are in free variation. The way in which they are used in the texts in this study provides a strong suggestion that while there is some overlap between the two, they are by no means equivalent. In Dem, although there are a few examples of ‘literal’ haver+past participle (as in (22) and (23) below) the majority of examples indicate past action (see (24)–(26)). (22) o corpo havia bem talhado (Dem 42) ‘he had a fine figure’ (23) o braço talhado havia (Dem 99) ‘his arm was cut’ (24) E por êsto lhe semelhava que o havia perdudo (Dem 39) ‘And because of this it seemed to him that he had lost him’ (25) Pois que Lançalot houve feito quanto a cavaleiro conviinha, disse…(Dem 42) ‘When Lancelot had done all that became a knight, he said…’ (26) era a seeda daquel cavaleiro que fôra morto, assi como o conto há já devisado (Dem 51) ‘it was the seat of that knight who had died, just as the tale has already described’ Similarly in ODE, although there are some possible resultant state examples, the majority indicate past action. In both these texts there are very few examples of ter+past participle, and it has only a literal interpretation. But if we look at RDJ, in which examples of ter+past participle outnumber those of haver+past participle, we still find that in several cases ter has a literal or a durative interpretation. For example: (27) E os da uylla tijnham as portas abertas (RDJ 52) ‘And the people of the town kept the doors open’ (28) el tinha o mar por seu e Lixboa aficada de fame (RDJ 57) ‘he held the sea as his own and Lisbon beset by hunger’ (29) juraua a Deus que, se el-Rey de Castella lhe mais deçepasse, que el tinha oitemta homeens do arreall pressos, e que todos lhos mandaria deçepados (RDJ 54) ‘he swore to God that if the King of Castile mutilated more of his men, he held eighty men from the camp captive, and he would order them all to be mutilated’ On the other hand haver+past participle always indicates past action.
Historical development in Portuguese
113
Here is a list of the participles to be found with ter and haver in RDJ. (Spelling has been modernised.) Participles with ter: ordenado (occurs six times); feito; presso (both occur four times); ouvido (three times); começado, prometido, mandado, tomado, aberto (twice); dado, dito, devisado, entrado (transitive), obrigado, posto, nomeado, aficado (once). Participles with haver: feito (five times); dado, mandado, entrado (intransitive), sofrido, enviado, partido, buscado, dito, perdido, cobrado (once). While a list such as this does not give a complete indication of how the constructions are used, it does enable some comparisons to be made. There are certain participles which occur with both ter and haver (feito, dado, dito, mandado). 2 They are used in much the same way with both verbs, and this would indicate some overlap between ter+past participle and haver+past participle. However, in other examples, ter +past participle is used lexically in a way that haver is not. As with its Spanish cognate, ter+past participle is used to indicate duration while haver+past participle is not. In this way ter+past participle has some specific uses linked to its lexical value, not open to haver+past participle. Conversely, haver+past participle has some uses not found with ter+past participle. Haver is already used with the participle of
intransitive verbs of movement (the last group of verbs to move from use with ser to Portuguese haver/ter or Spanish haber). It is not until Vie that ter appears with the participle of these verbs. This information indicates that haver+past participle and ter+past participle are not completely interchangeable, and that haver is more grammaticalised than ter, although by RDJ ter has caught up a certain amount. From then on, ter+past participle expands, as we saw above, gradually pushing haver+past participle aside. In the modern language haver+past participle is restricted to literary usage, and is not found in the spoken language. Although there are some examples of it in Vin, they all occur in descriptive passages, not in direct speech. 4.3.2 Agreement of past participle and direct object There are few examples where agreement can be tested, but haver+past participle follows a similar path to ter+past participle. With haver+ past participle there is no split between an invariant and an agreeing construction. There is a single development towards an invariant, past tense construction (see Table 4.6 and chapter 3, note 8).
Table 4.6 Agreement in haver+past participle construction Dem
ODE
RDJ
PJ
Vie
Vin 3
Agreement
10
4
9
cannot be
0
0
No agreement
0
1
1
tested
3
1
4.3.3 Word order Once again, the order ‘auxiliary+past participle’ is much more common than ‘past participle+auxiliary’ (see Table 4.7).
Tener+Past participle
114
Table 4.7 Word order in haver+past participle construction Dem
ODE
RDJ
PJ
Vie
Vin
Haver+past participle
29
11
14
4
12
3
Past participle+ haver
4
0
1
0
0
0
4.3.4 Interpolation Here the results are inconclusive (see Table 4.8).
Table 4.8 Interpolation of elements between haver and the past participle Dem ODE RDJ PJ Vie Vin No interpolation
19
10
13
4
8
3
At least one element between haver and the participle
10
1
1
0
4
0
4.4 HAVIA/TINHA +PAST PARTICIPLE AND THE -RA FORM As we saw in chapters 2 and 3, one way of measuring the grammaticalisation of a construction is to see whether there is any other form which has the same function. We saw that in Spanish había+past participle was initially an alternative to the synthetic pluperfect inherited from Latin; eventually había+past participle became the only pluperfect in Spanish, as the synthetic -ra form evolved into an imperfect subjunctive. The early Portuguese texts provide a similar picture to the early Spanish texts: havia+past participle, the -ra form, and slightly later on tinha+past participle are all found. However, as Table 4.9
Table 4.9 Comparative frequency of havia+past participle and -ra form as pluperfect Dem
ODE
RDJ
PJ
Vie
VME
CPA
Vin
havia+past participle (pluperfect)
4
7
13
3
2
0
0
3
tinha+past participle (pluperfect)
0
0
29
5
28
3
44
17
124
37
98
56
87
9
107
199
-ra*
Note: *Two problems arise when counting instances of the -ra form: firstly, as the third person plural form is for most verbs homonymous with the third person plural of the preterite, it is not always easy to decide whether a form is to be interpreted as pluperfect or preterite; secondly, in the texts up to and including Vie, some of the -ra forms appear to be dictated by sequence of tense rules
Historical development in Portuguese
115
in conditional clauses, and in these cases the -ra form is not interchangeable with havia/tinha+past participle. For example:
e fezera-se de grado afora, se podesse, mais nom podia polos outros (Dem 69) ‘and he would willingly have gone outside, but he could not because of the others’ E se nom fora que hia bem armado da cabeça, fora morto (RDJ 24) ‘And if he had not been well protected on the head, he would have been killed’ indicates, the -ra form continues as a pluperfect. It can be argued that in modern Portuguese the -ra form is not a true alternative to tinha+past participle, as it is largely restricted to the literary language. In this style it is, however, a very common alternative to tinha+past participle, as the figures in Table 4.9 indicate. In later texts there are apparently no examples of third person plural forms. Willis (1965, 200) points out that this form is rarely used now, owing to its formal identity with the third person plural of the preterite (see note to Table 4.9). In CPA and Vin which both contain direct speech and descriptive passages, tinha+past participle is found in both direct speech and description, whereas the -ra form is not in general found in speech. (There are two examples in CPA occurring in speech, where the -ra form is used more as a conditional.) This is in line with the view that -ra is restricted to the literary language. 4.5 HAVER/TER +PAST PARTICIPLE AND SER +PAST PARTICIPLE As in Spanish, the perfect of some verbs was at one time formed with ser. The ser+past participle perfect was gradually replaced by haver+past participle and later ter+past participle (see Table 4.10).
Table 4.10 Use of haver, ter, and ser, with intransitive and reflexive verbs Dem
ODE
RDJ
PJ
Vie
VME
haver+past participle
0
3
2
0
9
0
ter+past participle
0
0
0
0
12
8
ser+past participle
4
1
2
5
5
0
From VME onwards the ser+past participle perfect disappears completely. This is in line with greater grammaticalisation as haver, and then ter, replace ser in this use and widen their range of co-occurrence until they are used with the participles of all verbs.
Tener+Past participle
116
4.6 ASPECT It has already been noted that certain forms of ter+past participle are specifically iterative or durative. Irmen (1966) provides a description and a possible explanation for the development of this iterative and durative use. He suggests that in its early stages ter+past participle indicates resultant state. This gradually evolves into an indication of the past action—either durative or iterative—which produces the state. Irmen argues that the link between these two uses is provided by the fact that both uses establish a relation between past action and current state: No primeiro caso…o que faz esta ligação é o resultado duma acção inteiramante terminada e acabada e, como tal, separada do momento presente, ao passo que, no segundo, se encara a própria acção, começada num passado mais ou menos longínquo, no seu desenrolar até ao momento presente, sem que este, participando ainda dela, lhe ponha termo definitivo. (p. 234) It should be made clear here that there is nothing preordained about the direction of this development. It would be just as ‘logical’ for a current-state use to develop into a past punctual one, as the focus of attention switched from resultant state to previous causative action. I am not suggesting that Irmen claims any special status for the development from current state to durative/iterative past action; I merely want to underline that other developments are perfectly possible. Indeed, Irmen later goes on to suggest that there is a tendency for a resultant state use to move towards a past punctual one. Irmen suggests that the modern durative/iterative use develops directly from the resultant state use. While accepting that some of the resultant state examples may focus to some extent on the previous action (which is punctual, not durative or iterative) he specifically rejects any notion that at any stage in their history haver or ter+past participle may have referred to past punctual actions. However, Irmen has to stretch the notion of resultant state to the limit in order to accommodate all his examples within it. For example, he classifies the following as indicating resultant state (1966, 226–7): (30) Nosso Senhor me há encomendado êste homem ‘Our Lord has entrusted this man to me’ (31) Amigos, eu nom sei mais que diga, do que vos já tenho dito ‘Friends, I do not know what more to say, other than what I have already said’ (32) E desta guisa como teemdes ouvido ‘And in this way as you have heard’ We have seen that in the earliest stages of its development ter+past participle is essentially stative. (The early stages of haver+past participle take place before the time of the first texts examined here; I am assuming that they are in principle the same as those of ter+past participle, although, given the semantics of the two verbs and the differences found between them in the first texts, ter+past participle may well have a more specific
Historical development in Portuguese
117
element of keeping something in a particular state.) As Irmen suggests, this stative construction may develop something very close to a past action use. But to say that one use develops from the other is not to say that the two uses are essentially the same. Compare the following examples (already cited but repeated here for convenience). (33) nó cessarey porem de chamar o nome de Jhesu, porque o tenho scripto emno meu coraçom (ODE 10, 18) ‘but I will not cease to call the name of Jesus, because I have it written in my heart’ (34) mostrou-lhe rey Poro todos seus thesouros que tiinha escondidos (ODE 62, 19–20) ‘king Poro showed him all his treasures which he kept hidden’ (35) por fallar ao Mestre…sobre as mercadorias que lhe forom tomadas acerca do porto da dita cidade nas duas naaos, como teendes ouujdo (RDJ 11) ‘to speak to the Master…about the goods which had been taken from him near the port of the aforementioned city from the two ships, as you have heard’ (36) era a seeda daquel cavaleiro que fôra morto, assi como o conto há já devisado (Dem 51) ‘it was the seat of that knight who had died, just as the tale has already described’ While (35) and (36) may have considerable current relevance, they are different from (33) and (34), which involve something actually being in a particular state. Irmen implicitly recognises this, but seems loath to acknowledge it. There are two possible contributory factors to his determination to group everything under the notion of resultant state which are worth mentioning here. Firstly, part of Irmen’s argument seems to be based on the premise that throughout the history of the ter +past participle construction, a clear distinction should be drawn between a construction where the participle agrees with the direct object (which he classifies as a ‘giro sintáctico’) and a construction in which the participle is invariant (which constitutes a ‘forma verbal’). According to Irmen, only examples of the invariant construction should be eligible for this possible past action interpretation. As he finds some examples of the agreeing construction with a similar possible interpretation, he uses this as an argument for rejecting a ‘past action’ interpretation. While in principle it is desirable to find some means of separating the two constructions, and while they do develop differently, it seems reasonable to allow for a transition period in which some possible past action examples would have an invariant participle while others would still have an agreeing participle. This is what has been argued above. The acceptance of such a transition period would remove one of Irmen’s objections to a past punctual interpretation. But it says nothing about whether such an interpretation does or does not exist. With the transition period it would still be possible for the development to be from an agreeing construction which expresses resultant state to an invariant construction which indicates past durative or iterative action. An indication of the second possible factor in Irmen’s decision is given in the following quotation: Surgira a forma composta como giro sintáctico com particípio variável. Neste, o carácter de ‘perfectum praesens’ com sentido do resultado presente é muito mais forte que na forma simples do verbo, graças ao
Tener+Past participle
118
facto de concordar o particípio com o complement directo e manter o verbo finito (haver ou ter) a função de verbo independente com pleno sentido de posse presente. A forma composta, pelo contrário, pelo facto de o particípio não concordar mais com um complement directo, desliga deste o verbo finito, transformando-o em mero auxiliar privado de funções próprias, e assim desloca a atenção, de certo modo, do resultado para a própria acção verbal. Ao nascer a forma composta como forma verbal, existe portanto, desde o princípio, o perigo de confundí-la com a forma simples do pretérito perfeto. Por outro lado, continua em vigor a construção com particípio variável, de sentido resultativo present inconfundível. Neste dilema, a língua portuguesa acaba por sacrificar o significado perfectivo-resultativo da forma composta e substituí-lo pelo imperfectivo, afigurando-se-nos entre todas as línguas românicas a única que soube dar solução limpa e inconfundível ao problema e criar, em oposição tanto a construção com particípio variável quanto a forma simples, um meio de expressão suplementar que nenhuma outra conhece. (1966, 235) Irmen’s argument seems to be that the resultant state construction tends towards a past punctual interpretation; this then triggers the durative/ iterative development, which keeps the various constructions separate: if the resultant state construction did become a past punctual construction it would merge with the preterite, and probably preclude the reinterpretation of ter/haver+past participle as a durative/iterative past tense. But this is not the case. Portuguese may be unique in the particular arrangement it has developed, but this does not mean that this is the only means of keeping the three constructions separate, nor that if ter/haver +invariant past participle became a past tense it would merge with the preterite. Modern Spanish has tener+agreeing participle, haber+ invariant participle, and a simple past form. The three constructions function in different ways, without haber+past participle being durative or iterative. Haber+past participle is a type of past tense, but is differentiated from the simple preterite by the restrictions which operate on their respective uses. There is no theoretical reason for rejecting the possibility that haver/ ter+past participle had some type of past punctual use. There are not a great number of examples of present tense of haver/ter+past participle (and other tense forms do not have this durative/iterative aspect) but there is little evidence to suggest that in the early texts the ‘past action’ examples of either haver+past participle or ter+past participle are predominantly durative or iterative, as the following examples illustrate. 4 (37) Todo êste mal, que te eu digo, eu hei feito em uũ soo dia (Dem 75) ‘All this evil, which I recount to you, I have committed in a single day’ (38) E porém vos digo que nhuũ nom será tam ardido que o a seu colo deite, senam a quem o Nosso Senhor há outorgado (Dem 99–100) ‘And so I say to you that no-one will be so bold as to put it round his neck, except the man to whom Our Lord has granted it’ (39) outro tal edifício não havemos visto (PJ 84) ‘we have not seen another such building’
Historical development in Portuguese
119
(40) Estamdo em Bragaa Vaasco Lourenço…depois que perdo Neyua como teendes ouuido (RDJ 28) ‘When Vaasco Lourenço was in Bragaa…after he lost Neyua as you have heard’ (41) Do Príncipe, que Deus guarde, tenho já dado as novas a V. Exa (Vie 87) ‘Of the Prince—may God preserve him—I have already given you news’ It seems then that alongside Irmen’s view of a resultant state construction developing into a durative or iterative construction we have to allow for some form of past punctual use, albeit a severely restricted one. As we have seen, Irmen effectively acknowledges such a use, but rejects it apparently on theoretical grounds. To posit the existence of such a use at some stage in the development of the construction may make for a ‘messier’ description, but such a use does seem to be a part of the development and should be recognised as such. As for the relative chronology of the various stages, two possibilities suggest themselves. The first is an evolution from resultant state to a limited past punctual use, and then to a durative/iterative use. The second involves a move from resultant state to past action (albeit closely linked to the present) which could be punctual, durative, or iterative; from this stage the durative/iterative interpretation would eventually become the norm. Given the difficulties of interpretation—of deciding when resultant state shades into causative action, or whether an example is punctual or durative/iterative—it would be difficult to judge which is the more likely scenario. The most that can be said is that a past punctual stage seems likely, although it does not survive in modern Portuguese. The hypothesis that ter+past participle could at one time express past punctual action but can no longer do so is supported by judgements from a native speaker. She rejected sentences (40)–(41) above, which indicate events which are past punctual, saying that ter+past participle cannot be used in this way in the modern language. So far I have largely been in agreement with Irmen’s account of what happened, but have disagreed with his explanation of events. There is one further quibble with Irmen’s account. Irmen implies that the emergence of a tendency not to make the participle agree with the direct object leads to a reinterpretation of the construction as some form of past tense. It would seem more sensible to argue that it is the development towards a past tense interpretation which encourages the invariant participle construction. This is the view taken by Naro and Lemle (1977, 265): In the new compound tense construction the restriction against intransitive use of ter, the unmarked order with the object between ter and the participle, and especially the agreement in gender and number of the participle with the object are no longer natural. None the less, at least in the beginning stages of the new construction all of these formal peculiarities of the ancestor construction were maintained, disappearing only slowly over the next few centuries. This would accord with the fact that while there are some ‘past action’ examples with an agreeing participle, there are virtually no invariant participle examples with a resultant state interpretation. For example, we saw above that in RDJ approximately half the examples of ter+past participle have a past action interpretation, but at this stage the
Tener+Past participle
120
participle still agrees with the direct object. It would also be in line with the development of the haber+past participle construction in Spanish, and of similar constructions in French and Italian where even in the modern languages constructions which clearly indicate some form of past action may, in certain circumstances, show agreement of the participle.
4.7 TER +PAST PARTICIPLE IN MODERN PORTUGUESE Some indication of the use of ter+past participle in the modern language has already been given. However, before going on to a comparison of Spanish and Portuguese, modern usage must be summed up. We have seen that Portuguese develops both an invariant and an agreeing ter+past participle construction. The invariant construction indicates repeated or durative events which, while beginning in the past, continue up to the present moment. Castilho (1967, 88) describes the Portuguese construction as follows: ‘Se indica acção cursiva…ou iterativa…; em ambos os casos, a noção é a de passado que se estende ao presente, quer num movimento contínuo, quer num movimento interrompido e retornado diversas vêzes.’ Similar views are expressed by Paiva Boleo (1937, 4–7), Irmen (1966, 235–6), 5 Mattoso Câmara (1972, 145–6). For Sten (1973) ter+invariant past participle is indefinite, whereas the preterite is more precise: ‘Ce qui est essentiel, c’est que le perfeito exprime le “parfait” de façon vague, indéterminée, sans les contours nets du prétérit…. Dans les cas concrets, ce “sens fondamental” peut se manifester par une nuance durative ou itérative’ (p. 240). Later he gives a number of examples in which ‘l’alternance entre le perfeito et le pretérito est déterminée assez clairement par le contraste entre une action aux contours peu précis, durative ou répétée, et une action precise, très souvent unique (parfois répétée pour exprimer, dans beaucoup de cas, quelque chose de concret)’ (p. 254). In this way he aims to explain the usually iterative or durative nature of ter+past participle by reference to a more general underlying principle. Not all writers mention ter+agreeing participle. Castilho (1967, 87) mentions it as part of the historical development of ter+past participle but does not treat it as part of modern Portuguese. Mattoso Câmara (1972, 141) mentions it only to discount it as a periphrastic construction, on the grounds that the two verbal constituents do not form a semantic unity. Sten (1973, 234) suggests that where it combines with an agreeing participle, ter is a full verb rather than an auxiliary. Irmen (1966, 224–5) describes the use of ter+agreeing past participle in examples such as Temos o problema resolvido, Dizem que já têm o trabalho feito, Tenho uma carta começada há que tempos, etc.: O que nestes exemplos se tem em vista não é a acção passada, mas sim o resultado desta acção no momento presente, não é o acto de resolver ou de começar algo, e sim o facto de que, em consequência de um acto passado e completamente terminado, alguma coisa está resolvida ou começada e que o falante a tem com tal. O giro sintáctico em questão tem sentido perfectivo, com nítido carácter resultativo, e tem valor de presente muito mais que preterital.
Historical development in Portuguese
121
Dias da Costa (1976, 210–12) relates this construction to the semantic and aspectual content of ter and its copular value. He points out that ter +agreeing past participle is linked to the idea of ‘to maintain’, and that the construction either indicates the present result of a past action, or has the meaning of ‘conserving’ or ‘maintaining’ something in a particular state. Vázquez Cuesta and Mendes da Luz (1961,498) mention both the agreeing and invariant constructions, but do not describe how they differ in usage: ‘Con los verbos ter…y haver…usados como auxiliares de los tiempos compuestos, el participio permanece, en general, en portugués invariable… Pero cuando el complemento directo precede al verbo, el participio puede concertar con éste.’ 6 However, they implicitly recognise the difference between the two constructions, as they use Spanish haber +past participle to translate the invariant construction and tener+ past participle to translate the agreeing construction. One further point about the agreeing participle construction should be made: while, like its Spanish cognate, it indicates resultant state, unlike its Spanish counterpart it is not used to indicate repeated or durative action in the past. This difference in usage can be related to differences in the present perfect constructions in Spanish and Portuguese. As indicated above, the Portuguese present perfect—ter+ invariant participle—is specifically marked for repetition or duration: it refers to events which have begun in the past and which continue up to the present moment. There is thus little room for ter+agreeing participle to move from expressing resultant state to indicating past iterative action, as this slot is already filled by the invariant construction. In Spanish the haber+past participle present perfect places events within the extended present or indicates that they have particular current relevance, but is not marked for repetition or duration. And so there is more room for tener+past participle to come to indicate repeated or durative past action, as this is not specifically indicated by haber+past participle. To sum up, we can say that Portuguese has two ter+past participle constructions: one where the participle is invariant and precedes any direct object; and one where the participle agrees with the direct object, which it may precede or follow. The invariant construction indicates repeated or durative action which continues up to the present moment. The agreeing construction expresses resultant state.
4.8 COMPARISON OF SPANISH AND PORTUGUESE The major differences in developments in the two languages—a different choice of auxiliary and differences in use of their respective present perfect constructions—have already been outlined. Now that we have looked at developments in each of the languages individually we can proceed to a more detailed comparison. Early developments are much the same in the two languages. In both cases derivatives of HABERE have already begun to acquire auxiliary uses by the time of the first texts examined here, while derivatives of TENERE are much more restricted in this respect, with tener/ter+past participle having a primarily lexical interpretation. It may be that Spanish haber proceeds more quickly along the road to grammaticalisation than Portuguese haver. If we look at participial agreement we find that in Spanish there are sporadic examples of invariant participles right from the earliest text examined (early/mid thirteenth century), and that there are no cases of participial
Tener+Past participle
122
agreement from Cor (1438) onwards. In Portuguese there are no invariant participles in Dem (late thirteenth/early fourteenth century), and in ODE (late fourteenth/early fifteenth century) and RDJ (1440) there is still a clear tendency for the participle to agree with the direct object. We can thus assume that participial agreement disappeared earlier in Spanish than in Portuguese. The information on word order and separation of haber/haver from the participle does not provide much evidence either way, but more support—albeit rather tentative—comes from an examination of the speed with which haber/haver ousts potential rivals. We have seen that in Spanish the -ra form has disappeared as a pluperfect by Cel (1499). In Portuguese this form survives as a pluperfect in contemporary Portuguese, even if only in the written language; not even the development of ter+past participle—later but more vigorous than that of haver+past participle—displaces it completely. As for ser +past participle, in both languages ser disappears as a possible perfect auxiliary. The data presented here suggest that its disappearance is earlier in Spanish than in Portuguese. Haber is already found with the participle of intransitive and reflexive verbs in Mil, and has taken over completely from ser by Qui. In Portuguese haver does not appear with intransitives and reflexives in Dem, and ser is still found as a perfect auxiliary in Vie. Guiter (1980, 743) comes to a different conclusion: he argues that there is never any ser+past participle perfect construction in Portuguese, and that ser is used as an auxiliary only in the passive construction. While it is sometimes difficult to decide whether a particular example represents an active perfect or a passive (especially where the distinction between ‘action passive’ ser+past participle and ‘resultant state passive’ estar+past participle has yet to be clearly drawn) there are examples in the texts examined here which are fairly clearly active perfects. For instance: (42) Sabee de çerto que el-Rey he partido de Guimaraaes (RDJ 31) ‘Know for sure that the King is gone from Guimaraaes’ (43) Não me valia dizer que eram idos (PJ 79) ‘It was no use for me to say that they were gone’ (44) As coisas grandes não se acabam de repente: hão mister de tempo e todas têm seu tempo. O desta parece que é chegado (Vie 93) ‘Great things are not completed quickly: there is a need for time, and everything has its own time. It seems that the time for this matter is arrived’ Guiter’s remarks are thus not borne out by the texts examined here. On the contrary, it seems that ser+past participle persists longer as a perfect auxiliary in Portuguese than in Spanish. The overall conclusion is then that while Spanish and Portuguese begin in much the same way, Spanish moves slightly faster towards a grammaticalised perfect auxiliary than does Portuguese. But, of course, the eventual perfect auxiliary in Portuguese is not haver but ter. However, here again it seems that in the early stages Spanish is ahead of Portuguese. Spanish tener+past participle grammaticalises very little, but while we may not be able to show that it was ever ahead of Portuguese ter+past participle in its formal development, figures from Guiter suggest that at an early stage the Spanish construction was more common than its Portuguese counterpart. In the texts studied by Guiter, ter+past participle accounts for only 4.5 per cent of occurrences of compound past tenses in the thirteenth-century Portuguese texts, as compared with 7.1 per cent for tener+past
Historical development in Portuguese
123
participle in thirteenth-century Spanish. Added to the fact that the compound forms (derivatives of HABERE, TENERE, and ESSE with a past participle) are less common in Portuguese than in Spanish (see below) this would indicate that ter+past participle is slower to come into use than tener+past participle. However, the situation changes in the fourteenth century, where ter is comparatively more common than tener, and in the fifteenth century ter accounts for over 50 per cent of instances of compound perfects in the Portuguese texts. 7 In Spanish, on the other hand, tener+past participle drops to only 3 per cent of examples in the fifteenth century. The study I have carried out indicates that ter takes off as an auxiliary from RDJ, a conclusion which is compatible with Guiter’s views. But although ter has a slower start than tener, one of the ter+past participle constructions does eventually grammaticalise much further than tener. By the time of RDJ ter has moved further from its original lexical interpretation and closer to a past interpretation than tener+ past participle. And by Vie it has clearly overtaken tener+past participle in its degree of formal grammaticalisation. But what happens when we compare Portuguese ter with Spanish haber? Unsurprisingly, haber is more grammaticalised than ter in the early stages of the language, as ter begins its development much later than haber. Let us turn our attention to the contemporary language. I have already mentioned the presence of the -ra pluperfect in Portuguese as a rival to tinha+past participle. Also, ter has not completely replaced haver as the perfect auxiliary in the literary language, although haver is severely restricted. Looking at word order and separation of auxiliary and participle, we find that both haber and ter can be separated from the participle. However, this separation is more common in Portuguese than in Spanish. A major reason for this is that Portuguese rules on weak pronoun position dictate that a pronoun is often suffixed to the auxiliary: i.e. it is placed between the auxiliary and the participle (see p. 135). We have already noted that this is a syntactic point not applicable solely to ter+past participle. But it may have repercussions for the grammaticalisation of auxiliaries. In Spanish there are very few circumstances in which something has syntactically to be placed between the auxiliary and the participle. (It does occur, for example, where the auxiliary is an infinitive or a gerund: después de haberlo visto; habiéndolo visto.) The fact that other factors in the language dictate that the participle will be separated from the auxiliary more frequently in Portuguese than in Spanish could affect relative grammaticalisation of the two constructions. This is not to say that if a higher degree of fusion between auxiliary and participle did occur, rules governing pronoun position would not gradually change so that pronouns would no longer ever split up the construction. What is being claimed is that while in both Spanish and Portuguese the auxiliary and the participle are still separate entities, the fact that they tend to be immediately next to each other more frequently in Spanish than in Portuguese means that such fusion (and thus greater grammaticalisation) is more likely to happen in Spanish, although complete fusion is as yet far from occurring. It seems that not only historically but also in the modern language ter is less grammaticalised than haber. Another interesting point, which may be related to grammaticalisation, is frequency. In Guiter’s study of Spanish texts from the twelfth to the sixteenth centuries an average of 10.3 per cent of indicative past tense forms are compound forms (haber/ tener/ser+past participle). His figures for Portuguese texts over the same period produce an average of 2.2 per cent for compound forms. Two matters
Tener+Past participle
124
might be considered here. The first is the relation between frequency and grammaticalisation. It is quite plausible that forms which are less common should tend to grammaticalise more slowly than forms which are used more frequently, although this still begs the question as to whether lower frequency entails lower grammaticalisation, or less grammaticalised forms tend to be less common, or whether both factors apply; nor does it explain why the compound forms are initially less common in Portuguese. The second point involves interpretation, and could help to explain why ter+invariant past participle is less common than haber+past participle in the modern languages. In addition to referring to events which continue up to the present moment the present tense of ter+past participle (i.e. the present perfect) is specifically marked for duration or repetition: it is not used for single punctual events. There are no such limitations on the Spanish present perfect: the overriding factor here is whether or not the event is viewed as having current relevance or taking place within the extended present; aspect (duration, repetition, etc.) is not important. It is unsurprising then that Spanish haber+past participle should be more common than Portuguese ter+past participle, as it has a wider range of application. Again, of course, this does not explain why ter+past participle should be less common than haber+past participle earlier in the language, but it does provide a link between interpretation and frequency. It may be possible to link grammaticalisation and frequency in another way. First we must look more closely at the interpretation of the construction and the historical development of this interpretation. In his analysis of the history of the present perfect throughout Romance, Harris (1982, 49–50) puts forward the following outline development, using HABEO FACTUM as an example: 8 1 derivatives of HABEO FACTUM are used to describe ‘present states resulting from past actions’; 2 derivatives indicate current relevance and are also marked for duration or repetition; 3 they express ‘past action with present relevance’ (without being aspectually marked for duration, repetition, etc.); 4 they take over the functions of the preterite, and the opposition between present perfect and preterite is neutralised. 9 Portuguese ter+past participle is at stage 2 of this development, while Spanish haber+past participle has moved a stage further and is at stage 3. Can the fact that ter+past participle has not proceeded so far through these semantic changes be linked to its lower level of grammaticalisation? Such a link would not necessarily have to be causal: it is not that less semantic evolution means lower grammaticalisation or vice versa—a durative/iterative perfect is not in itself less grammaticalised than an aspectually unmarked one. But these factors—semantic and grammatical—could have their roots in a common cause. Degrees of both semantic evolution and grammaticalisation suggest that ter+past participle has changed less than haber+past participle. We have seen that ter begins to develop as a perfect auxiliary considerably later than haber, and that compound perfect forms are less common in Portuguese than in Spanish. These factors—a shorter history and lower use—could help to explain why ter +past participle has advanced less along the paths of both grammaticalisation and semantic evolution. Turning now to resultant state constructions, we see that although Spanish and Portuguese now have different perfect auxiliaries, in both languages the resultant state constructions have their roots in literal tener/ter, and in particular in its copular or durative value. They undergo similar developments and have similar uses in the contemporary languages. HABERE and its derivatives may have gone through a
Historical development in Portuguese
125
predominantly resultant state stage before the time of the texts studied here, but at an early stage TENERE appropriated this function almost exclusively. Once again, it seems that Spanish was quicker off the mark than Portuguese. Similarly, it appears that the regular distinction between resultant state and present perfect is drawn sooner in Spanish. This can be attributed to two related factors: firstly, haber establishes itself as a perfect auxiliary more quickly than haver. Secondly, tener does not rival haber in the way that ter rivals and then overtakes haver. Spanish haber is more firmly entrenched than Portuguese haver, and even though tener is initially more advanced than ter, it never seriously threatens haber: in Spanish the two verbs fairly soon settle into the pattern which exists today. (There are slight variations, and in the sixteenth century tener may come a little nearer to being a perfect auxiliary than it does today, but the broad outline is the same.) In Portuguese haver initially begins to evolve as the perfect auxiliary, but is later overtaken by ter. The sort of distinction made in Spanish between haber+past participle and tener+past participle eventually occurs in Portuguese, between ter+invariant participle and ter+agreeing participle. But although from a fairly early stage Portuguese has a resultant state construction and an incipient perfect construction, the clear formal distinction between them comes later in Portuguese than it does in Spanish. When haver is the principal perfect auxiliary there is a formal distinction between haver+past participle, which indicates some form of perfect, and ter+past participle, which indicates resultant state. But when ter+past participle begins to evolve a perfect function, this formal distinction is for a time blurred, until the perfect construction reaches the formal stage where the participle never agrees with the direct object, and is in this way differentiated from the resultant state construction, where the participle continues to agree with the direct object. Turning now to interpretation, we saw in section 4.7 that in the modern languages Portuguese ter+agreeing participle is more restricted in use than standard Spanish tener+agreeing past participle. However, both the agreeing constructions follow a similar pattern of development to the one from Harris outlined above, although they have not proceeded so far with this development as ter+invariant participle or haber+past participle. Tener+past participle is straddling stages 1 and 2 of the development, and also retains its copular uses. In general it is still at stage 1, as it principally indicates resultant state, but in some cases it has reached stage 2 (as in the tener dicho examples, or examples with specifically iterative adverbial phrases—see chapter 2, examples (10) and (53)–(56)). In Portuguese ter+agreeing participle has only a stage 1 interpretation: as was pointed out in section 4.7 it cannot be used to express durative or iterative aspect. It was also argued there that differences between the Spanish and Portuguese agreeing constructions can be linked to use of ter/haber+invariant participle. Spanish haber+past participle has moved as far as stage 3, leaving room for tener+agreeing participle to begin to develop a stage 2 use. Portuguese ter+invariant participle has remained at stage 2, rendering any similar use of the agreeing construction redundant. We have seen then that at all stages Spanish is ahead of Portuguese in its grammaticalisation of a present perfect construction. Progression through the various stages outlined by Harris is not to be equated with grammaticalisation, and—particularly in the case of a move from stage 2 to stage 3—is more a question of changing temporal and aspectual values. But here too Spanish is ahead of Portuguese. The Spanish construction is also more common than the Portuguese one. I have given an indication of
Tener+Past participle
126
how these factors—grammaticalisation, temporal/ aspectual evolution, and frequency of occurrence—can be linked, without the need to posit a direct causal relation between them.
5 Comparison with other Romance languages 5.1 OTHER REFLEXES OF TENERE+PAST PARTICIPLE This study has concentrated on the development of reflexes of TENERE+ past participle in Spanish and Portuguese. But similar constructions exist in some other Romance languages—notably the dialects of north-west Spain, Catalan, and some dialects of Italian. While it is beyond the scope of this study to look at these in detail, an outline will be given, before going on to an overview of the development of TENERE+past participle in Romance. 5.1.1 Galician Use of ter+past participle in Galician is very similar to use in Portuguese. In both languages ter+invariant participle forms a durative/iterative perfect, and ter+agreeing participle indicates resultant state. Paiva Boleo (1936, 13–15) claims that Galician ter+ past participle is identical in usage to Portuguese ter+past participle, apart from an occasional difference caused by the influence of Castilian on Galician. For Rojo (1974, 130–2) the construction expresses repeated action, but he does not mention any possible durative value. Porto Dapena (1972, 14) states that ‘Teño dito…, non ten somente un puro sentido terminativo ou perfeutivo como o castelán he dicho, senón que engade a acción un matiz reiterativo: a traducción en castelán sería He dicho (ou dije) muchas veces’ Rojo (1974, 129) remarks on ter+agreeing past participle in Galician, drawing a parallel between the Galician construction and Castilian tener+agreeing past participle: Cuando se quiere marcar específicamente el carácter perfectivo de una acción, se utiliza corrientemente ter+participio. Pero entonces no equivale a la castellana haber+participio, sino a la latina habere +participio…. Esta posibilidad existe también en castellano por medio de la construcción paralela. Santamarina (1974) distinguishes two constructions in Galician—one with an agreeing past participle and one with an invariant one. The interpretations he ascribes to them agree with those put forward by Rojo. In principle, Galician ter+past participle behaves as Portuguese ter +past participle does, both as an iterative or durative perfect, and as an expression of resultant state. There is, however, a slight difference in use of the agreeing construction, which will be discussed below (section 5.2).
Tener+Past participle
128
5.1.2 Asturian Very little work has been done on perfect auxiliaries in Asturian or in the Spanish of this area. Several writers claim that Asturian does not possess compound forms, and some state that one of the characteristics of the Spanish spoken in this region is the use of simple forms where standard Spanish would use compound forms. Menéndez Pidal (1906, 310–11) talks of the ‘carencia total del Perfecto perifrástico en algunas regiones como Asturias, donde nunca se usa más que el Perfecto simple, aun en el habla castellana de las personas cultas’. Similar comments are made by Rodríguez-Castellano (1952, 181; 1954, 241) and García Alvarez (1960, 422–3). However, it may well be that when writers say that the perfect does not exist in Asturian they really mean that haber+past participle does not exist. For example Grossi (1962, 452) says of the central Asturian dialect he is studying: Podemos hacer notar la ausencia total de formas compuestas. Un hablante de esta zona, por grande que sea su castellanización, será incapaz de distinguir la diferencia entre las formas castellanas ‘he dicho’ y ‘dije’, debida esta indiferenciación a la falta dialectal de la primera forma. Martínez Alvarez (1973, 300) states that ‘la peculiaridad más saliente de ese conservadurismo en los dialectos occidentales [of the Iberian peninsula] es la originaria inexistencia de las llamadas “formas compuestas” del verbo, es decir de los significantes constituidos por la combinación del “auxiliar” haber y el participio’. These quotations suggest that when linguists claim that there are no compound forms in Asturian they may well be referring purely to haber+past participle. Similarly, Canellada (1944, 33) states: ‘Se emplea solamente el pretérito simple. (Rasgo común con el portugués.)’ Portuguese certainly does possess compound forms, formed by ter+past participle, but their usage does not correspond to that of Castilian haber+past participle, and they are used less frequently than the Castilian compound perfect. Paiva Boleo (1936, 12–19) claims that a similar difference to the one found in Portuguese between the simple preterite and the compound perfect is to be found in Galician and to some extent in Asturian and Leonese. But in his discussion of Asturian and Leonese he does not mention tener+past participle, citing only a few examples of haber+ past participle. However, some writers do mention tener+past participle in the dialects of the region. Millán Urdiales (1966, 174–5) comments that in Leonese, as in the Castilian of Galicia, the preterite covers the functions carried out by the perfect in standard Spanish, but he also mentions that in the Leonese dialect he is studying tener+past participle is sometimes used. Although Millán does not analyse the way in which the construction is used, the examples he gives all point to an iterative interpretation. Yllera (1980, 285) suggests that tener+past participle is more grammaticalised in Asturian than in standard Spanish. And according to Seifert (1930, 383) ‘el asturiano moderno, igual que el portugués, emplea casi exclusivamente tener como auxiliar’. My work with informants in Oviedo indicates that use of tener+past participle in the Spanish of this region is closer to use in Portuguese or Galician than in standard Spanish, but that there are important differences. As there has been no clear description of tener+past participle in the Spanish of Asturias, and as it displays some interesting
Comparison with other Romance languages
129
differences from neighbouring dialects, I shall describe it in more detail than some of the other dialects. 1 In outline, a preposed participle may or may not agree with the direct object; a postposed participle must agree with the object. This difference between agreeing and invariant participles corresponds to a difference in interpretation. At its most basic, an invariant construction has a past tense interpretation and an agreeing participle construction has a present tense interpretation. The most common use of the ‘past’ construction is to indicate an iterative or durative event. All speakers agreed that this was the most likely interpretation for the following sentences: (1) Tengo ido a Madrid (2) Tengo comido en ese restaurante (3) Tengo perdido la cartera (4) Mis padres me tienen prohibido muchas cosas In other cases, where an informant was given an example of this construction he or she would provide a context which indicated iterative action. For example, when given Tengo olvidado los libros que me prestaste and Los niños se tienen lavado they produced sentences such as (5) and (6). (5) Tengo olvidado los libros que me prestaste muchas veces, pero hoy te los traigo (6) Los niños se tienen lavado en la fuente del patio cantidad de veces This is perhaps the interpretation one might expect. We saw with the informants from Valladolid and Zaragoza that where tener+past participle has more of a past interpretation it tended to imply repetition or duration. The second major use of tener+invariant participle is to place an event in close relation to the present moment, or to something which takes place in the present moment. One informant explained that, while out of context sentence (2) above would be interpreted as expressing repeated action, it could also be the reply to the question ¿Por qué no vamos a ese restaurante nuevo? In this case (2) would express an event of particular relevance to the present moment. It is also frequently used to place an event within the extended present, as in (7) and (8). (7) ¿Le tienes prestado el coche alguna vez? (8) Todo el mundo tiene perdido la cabeza en alguna ocasión de su vida For tener+past participle to be used in this way it must refer to an event which takes place at a time which is not specifically defined, but which is placed within the extended present. For example, in (7) or (8) there is no indication as to when the event actually happened; it merely has happened at some point within a past which stretches right up to the present moment. 2 One of the informants pointed out that this construction is frequently used in boasting of past deeds, in cases such as (9). (9) Ayer pesqué una trucha que media… Pues eso no es nada. Yo tengo pescado una que media…
Tener+Past participle
130
Again, in this type of example, the action took place at some unspecified moment within the extended present. These two categories—repeated or durative events and the extended present or close relation to the present—cover most examples of tener+ invariant participle. In both cases the action is not tied to a specific moment and is related to the present moment. It has either been going on over a period of time up until the present or takes place at an unspecified moment in the extended present. However, there are cases where tener+invariant participle refers to events which are not linked to the present in this way. The contexts in examples (10)–(13) were all produced by the informants themselves, and were not given to them as ready-made sentences for them to judge. (10) Tengo comido muchas patatas durante la guerra (11) Tengo castigado a la niña cuando tenía diez años y se portaba muy mal (12) Tengo aprendido los nombres de toda esta gente aunque ahora ya me los olvidé (13) Tengo pensado mucho en ese problema y ya llegué a la solución These examples are either iterative or durative, or can be thought of as such. But in all of them any close relation to the present moment is clearly ruled out. There are also cases in which tener+past participle does not indicate repeated or durative action and does not place an event within the extended present. Informants produced sentences such as: (14) Tengo roto la pierna en esos días que llovía y estaban las calles resbaladizas (15) Se tienen puesto de acuerdo que había que exterminar las ratas cuando la peste Informants varied as to how ready they were to ascribe a punctual interpretation or to provide a context which ruled out any close connection to the present, but all produced some examples of both of these types of sentences. The more readily a speaker accepted the construction, the more likely he or she was to ascribe or to permit a punctual interpretation. In general, tener+invariant participle links an event closely to the present—either ascribing an iterative or durative aspect or placing a punctual event in the extended present. However, we have seen that there are cases where there is no such link with the present, although this use is not so common. The relation between these different uses and the possible chronology for their development will be discussed in section 5.2. It seems that there is a division of functions between tener+invariant past participle and the simple preterite. Tener+past participle is used to refer to past events which are usually iterative or durative, or are placed at some indeterminate point within the extended present. The simple preterite can refer to events at any point in the past, even within what would be classified as the extended present in Castilian. This use could explain why tener+past participle rarely refers to ‘definite’ punctual events, as this use is covered by the preterite. (The preterite is punctual unless specifically marked for repetition or duration.) Haber+past participle, when used at all, seems to refer very specifically to past events which have caused a still-existing current state. Haber+past participle is apparently little used, and the functions it has in standard Spanish are carried out either by tener+invariant past participle or the simple preterite. The judgements which the Oviedo informants gave on tener+ agreeing past participle were very similar to those given by speakers of standard Spanish. However, there is one
Comparison with other Romance languages
131
area in which they differ. We saw in chapter 2 that in standard Spanish, in addition to its resultant state use, tener+agreeing participle can refer to previous repeated or durative events. The Oviedo informants were less willing to accept this use of the agreeing construction, frequently substituting an invariant participle for an agreeing one in such cases. Thus, for example, they changed Tengo perdida la cartera varias veces and No me creo las cosas que me tienen dichas to Tengo perdido la cartera varias veces and No me creo las cosas que me tienen dicho respectively. However, tener does occur with an agreeing participle where context indicates that an action which has been occurring in the past will continue into the future. So there are contrasting pairs of examples such as: (16a)
Tiene perdidos cinco kilos y espera perder cinco más
(16b)
Tiene perdido cinco kilos pero después engordó diez
(17a)
Tengo vistas muchas películas de Orson Welles y voy a ver más
(17b)
Tengo visto muchas películas suyas, pero no me gustaron
In section 4.8 we saw how the restriction of Portuguese ter+agreeing participle to an indication of resultant state (as compared with standard Spanish tener+agreeing participle which can also indicate durative or repeated action) can be related to the presence of a durative/iterative invariant construction. A similar situation appears to obtain in the Spanish of Oviedo. However, although the agreeing construction in Oviedo Spanish has a narrower range of application than its standard Spanish counterpart, it is less restricted than in Portuguese, where examples similar to (16a) and (17a) are not possible. The question of the agreeing construction in the Spanish of Asturias will be taken up again in section 5.2, where further comparison will be made with standard Spanish and Portuguese. 5.1.3 Catalan Tenir+past participle is found in modern Catalan, and both its meaning and the syntactic restrictions on use are similar to those of Spanish tener+past participle, although it seems that the construction is less common in Catalan. Badia Margarit (1962a, 396–7) states that: La perífrasis tenir+participio añade, al tono general perfectivo o terminativo y de anterioridad, un aspecto resolutivo de que carece la perífrasis con haver (comp. aquest llibre el tinc demanat y l’he demanat ‘este libro lo tengo pedido’ y ‘lo he pedido’); aunque sin poseer la riqueza ni los matices que caracterizan la perífrasis cast. ‘tener’+ participio, son evidentemente correctos, en catalán, ejemplos como el citado y otros (tenim decidit de fer-ho ‘tenemos decidido hacerlo’ etc.); en este caso el participio concierta como adjetivo—igual que sucede en castellano—con el sustantivo sujeto o complemento verbal. Yates (1975, 89) ascribes a similar interpretation and also underlines the link with possession: ‘Tenir is used in a construction equivalent to a compound tense when
Tener+Past participle
132
emphasis is on final state rather than action, often when the idea of possession is uppermost.’ However, some other writers (e.g. Moll 1952; Gili 1943) make no mention of the tenir+past participle construction, which would appear to back up Badia’s view that it is more restricted in Catalan than in Spanish. If we look at the historical development of the Catalan construction, we find that it is very similar to that of the Spanish construction. The most detailed historical survey is probably that provided by Seifert (1957–8). She divides instances of tenir+past participle into two major categories: cases where tenir has its original meaning of ‘to hold’; and cases where tenir has a more grammaticalised role in the sentence. The various subgroups listed under the first of these categories correspond closely to the various types of example of tener+past participle found in Old Spanish: they indicate that something or someone is being kept hidden, shut up, imprisoned, or bound, or is held ready, or is in someone’s power or under their protection. Here are some of Seifert’s examples: (18) en la qual [cova] un pastor tenia amagada .I.a fembra ‘in which [cave] a shepherd kept a woman hidden’ (19) i a vós rés no tendria cellat ‘I would not keep anything hidden from you’ (20) que no esguart ab qui tinga l’ull clus ‘he whose eye is closed does not see’ (21) fon-se venir .X. francès, d’aquels que tenia preses ‘they make ten of the Frenchmen come whom they were holding prisoner’ (22) manà a l’amirall…que tota hora tengués cinquanta galees adobades e aparellades ‘he ordered the admiral…at all times to have fifty galleys prepared and fitted out’ (23) E així que tendría la casa de Castella pujada a sa voluntat ‘And so he would have the house of Castile at his disposal’ As can be seen from these examples, usage is very similar to Spanish. In the second category, where tenir has largely lost its semantic content of ‘to hold’, Seifert suggests that tenir+past participle is more emphatic than, for example, estar+past participle would be. She also emphasises the durative nature of the construction. She suggests that through use the emphasis and the duration may fade, and tenir moves closer to becoming a perfect auxiliary. Once again the examples indicate that use is very similar to Spanish: (24) En l’any…733 aprés de la traició del Comta Julià, com molts tenen escrit que los moros…ocupàren tota la Hespanha ‘In the year…733 after the treachery of the count Julià, as many people have written, when the Moors…occupied the whole of Spain’ (25) Gran es estat…lo digres, que en los vltims capitols del precedent llibre tinch fet ‘Great has been…the discourse which in the last chapters of the previous book I have made’ (26) La cara tinch embotornada ab les arpes del gat ‘My face is swollen from the cat’s claws’
Comparison with other Romance languages
133
(27) Curial viu que tots los camins que solia tenir uberts li eran tancats ‘Curial saw that all the roads which were usually open to him were closing on him’ (28) I no em parleu de perills…; ho tinc pensat i apamat i…estic ben decidida ‘And do not speak to me of dangers…; I have thought about it and weighed it up and…my mind is completely made up’ (29) la creença de que les havia de fer per respondre dignament al concepte d’artista genial que de mi mateixa tenia format ‘the belief that I had to make them in order to do justice to the concept of the artist of genius which I had formed of myself’ The conclusions reached by Roca Pons (1957–8) and Vallcorba i Rocosa (1978) support Seifert. Roca Pons (pp. 165–6) states that in Old Catalan tenir ‘pot entendre’s com a designació expresa d’una idea de possessió o bé, només, en el sentit d’exigir-se que l’objecte estigui, respecte al subjecte, en una relació possessiva o derivada, la qual justifica l’us de l’auxiliar’. He goes on to say that ‘En el catalá antic tenir… arriba, bé que rarament sense una veritable significació concreta, a expresar l’interès actual d’una acció passada i acabada.’ These two descriptions fit in with Seifert’s two major categories: in one case tenir indicates some form of possession, or of keeping or holding something; in the other it indicates durative resultant state. Vallcorba plays down the possessive element of tenir+past participle, asserting that in medieval Catalan tenir does not indicate possession, for which haver is used. He stresses the primary durative value of tenir, (including cases where it is coupled with a past participle). He also underlines the relation between tenir and estar (a point already briefly made by Roca Pons). Seifert does not give such a detailed chronology for Catalan as she does for Spanish, but she does say that tenir+past participle is found in the work of R.Lull (thirteenthfourteenth century), appears frequently in fifteenth-century poetry, and that occasional examples are found from that time up until the present day. Such an outline suggests that tenir+past participle declined earlier than Spanish tener+past participle, which reaches its peak in the sixteenth century and is still found reasonably frequently in the seventeenth and eighteenth centuries. This is compatible with the view that in the modern language tenir+ past participle is more restricted than tener+past participle. To sum up, we can say that while tenir+past participle follows a similar development and has similar uses in the modern language to tener+past participle, it does not apparently gain such a foothold in Catalan as tener+past participle does in Spanish; although it is present in Old Catalan (whether with the same frequency as tener+past participle in Old Spanish we cannot tell) it dies away earlier, and does not have the same degree of acceptance in the present-day language. 5.1.4 Italian The last area of Romance which we have to consider is Italian, and in particular the dialects of southern Italy. Rohlfs (1969, sec. 733) points out that ‘l’italiano in generale conosce questa costruzione soltanto nel significato originario di “mantenere”, per esempio, tieni chiusa la porta, mi tiene ocupato’ Such a use is of course also found in Spanish. But standard Italian does not share the iterative or durative perfect function
Tener+Past participle
134
found in some Spanish examples of the construction. This perfect function is found, however, in some southern Italian dialects. According to Seifert (1935, 87) it occurs in all of southern Italy except Calabria and Sicily. (The examples given by both Rohlfs and Seifert are taken from a variety of dialects and periods of time. 3 ) Rohlfs cites various examples, among them: (30) il marito teneva detto che allumasse il fuoco ‘her husband had told her (several times) to light the fire’ (31) teneva na funtana frabbecata, rose e viole nce teneva chiantate ‘he had built a fountain, I had planted roses and violets there’ (32) maritimo i tè condati ‘my husband has counted them’ However, he concludes that ‘in nessun luogo si è ancor giunti ad una totale sostituzione di avere con tenere in nesso col participio passato. Si ha nel Meridione la prima fase d’uno sviluppo la cui fase finale è presentata dal portoghese dove “tenere” ha veramente soppiantato il precedente “avere”.’ 4 A similar picture is painted in the more detailed historical survey by Seifert (1935). According to Seifert the construction is much less common in Italian—even in the southern dialects—than in Spanish and Portuguese. Many of her examples could be classed as ‘lexical’ tenere+ past participle, and have parallels in both Spanish and Portuguese. For example, where tenere is used meaning ‘to keep’, as in (33) la sua tardanza mi aveva lungamente tenuto sospeso nel timore ‘his lateness had for a long time kept me waiting in fear’ Also, where tenere means ‘to wear’ or ‘to have on or in one’s body’: (34) [il nome] lo tengo rinserrato nel cuor mio ‘I keep it [the name] hidden in my heart’ As in Spanish and Portuguese tenere can be used with certain past participles to mean ‘to keep protected’: (35) le ténghe stepate (chiuse nell’armadio) 5 ‘I have them stored (closed in the cupboard)’ (36) ‘na trizza di li to’ tegno sarvatu ‘I keep saved a lock of your hair’ [literally ‘a lock of yours’] or ‘to hold ready’: (37) na bona mazziata—da no piezzo le tengo preparata ‘I have had a good cudgelling ready for her for some time’ Seifert stresses the peculiarly durative nature of tenere+past participle present in more grammaticalised examples as well as in the lexical ones. She suggests that it can combine in one construction the aspects of avere+past participle and stare or andare+past participle.
Comparison with other Romance languages
135
She also points out that, depending on the semantics of the participle, the construction may emphasise either an action or a resultant state, but that it is always essentially durative—a situation very similar to the one we have already come across in Spanish. Seifert cites a couple of examples where she claims that tenere+past participle comes close to a perfect function (1935, 85): (38) nu curtiello m’aggi ‘astipato—‘o tengo a trentasè menate ‘I have armed myself with a knife—I have wielded it against thirty-six [people]’ (39) vna Donna bellissima, che notriua cinque fantolini soi figlioli, li qual tenea partiti, trè dalla parte dritta, & li doi altri figlio [sic] li tenea dalla sua parte mancha charamente ‘A very beautiful woman, who was nursing five babies who were her children, whom she had separated [or ‘kept separate’], three on the right side, and the two other children she kept affectionately on her left-hand side’ (As Seifert believes that these are close to being examples of the perfect, she presumably interprets the use shown in (39) as ‘she had separated’ rather than ‘she kept separate’. However, lack of context makes it difficult to evaluate her reasons for so doing; it is hard to decide whether tenere+past participle in this example is really to be classified as a ‘near-perfect’, or whether it is primarily durative.) There is one example which Seifert classifies as a ‘pure’ perfect (1935, 86): (40) ma lo Gesellatore sujo lo tengo schiaffato e rebattuto ‘but its Creator I have slapped and spurned’ However, Seifert concludes that the construction retains a very strong element of duration and of keeping something in a particular state, and so tenere is rarely found with a participle which stresses action rather than state. From the accounts given by Rohlfs and Seifert we can see that tenere +past participle in southern Italian has much in common with its cognates in Spanish and Portuguese. But Seifert states directly and Rohlfs implies that the Italian construction is used less frequently and has a more restricted range of application than its Spanish or Portuguese counterparts.
5.2 COMPARISON OF DEVELOPMENT Having looked at the reflexes of TENERE+past participle in the various Romance languages which possess them, let us now try to summarise and to compare their development. Seifert (1930, 371–3) suggests that the beginnings of a periphrastic construction formed by TENERE+past participle can be found in Latin texts, but that there are very few examples in early Romance texts. She concludes that ‘no es admisible, pues, hablar de una continuación de la tradición latina en la lengua vulgar románica’. While examples of TENERE+past participle are found in Latin, development of this construction as a periphrasis appears to be a Romance phenomenon, which has followed a similar path in the different languages, progressing along that path to a greater or lesser degree.
Tener+Past participle
136
We saw in section 4.8 the four stages in the development of the present perfect in Romance, as described by Harris (1982). First the construction indicates present states resulting from past action; then current relevance plus duration or repetition; then past action with present relevance; and then past action without being marked for present relevance. If we apply this description to reflexes of TENERE+ past participle we see that in standard Italian and Catalan they have proceeded no further than stage 1. In Italian in particular tenere+past participle is very limited in scope, and tends to incorporate the notion of keeping something in a state. In standard Spanish, and seemingly in some southern Italian dialects, it usually has a stage 1 meaning but in some examples is at stage 2 (as in tener dicho). In Portuguese, Galician, and the non-standard variety of Spanish in Oviedo there is a split between the agreeing resultant state construction and the invariant construction. Portuguese ter+invariant participle is at stage 2 (but see the comments on its historical development in section 4.6). The Galician invariant construction apparently functions in the same way. Oviedo tener+invariant participle is usually a stage 2 construction, in that it is usually iterative or durative. However it does also have a stage 3 use: it can apply to punctual events which have a particular present relevance, as in (41); it can also refer to punctual events which take place within the extended present, as in (42). (41) ¿Porque no vamos a ese restaurante nuevo? Ya tengo comido en ese restaurante (42) Todo el mundo tiene perdido la cabeza en alguna ocasión de su vida There is, however, an indication that it is moving to stage 4: all informants produced examples where the event described was not placed in a period of time extending up to the present—that possibility being excluded by context. Sentences (10), (11), (14), and (15) above exemplify this. Such examples are less common, and a stage 4 interpretation is less readily applied to examples than a stage 2 or stage 3 interpretation, but nevertheless it is clear that Oviedo tener+invariant participle can be used in this way. Turning now to the agreeing constructions, we have seen that there are differences here between the various languages. The agreeing construction in Portuguese is more restricted than in the speech of the Oviedo informants, where it is, in turn, more restricted than in standard Spanish. These differences in usage can be linked to use of the invariant construction. The Oviedo speakers rejected several of the standard Spanish examples where tener+agreeing participle indicated repeated action (e.g. Tengo perdida la cartera varias veces). This is unsurprising, as the invariant construction can be used in cases such as these. In the Oviedo dialect, where tener+agreeing participle indicates repeated action it usually implies future continuation of the action (see p. 159), which is not a necessary part of the interpretation in standard Spanish. In standard Spanish there is no opposition between invariant and agreeing constructions, and the agreeing construction can refer to past iterative events. Similarly, differences between Portuguese and the Oviedo dialect in use of the agreeing construction can apparently be related to differences in the invariant construction. For the Oviedo informants tener+invariant participle can indicate past action that is no longer taking place (see section 5.1.2). And so a contrast can exist between tener+agreeing participle—indicating action which continues into the present and implies future continuation—and tener+invariant participle which indicates an action which is no longer occurring. In Portuguese ter+invariant participle indicates action which continues up to the present, thus making the agreeing construction redundant in
Comparison with other Romance languages
137
such examples. Portuguese ter+agreeing participle seems to be restricted to cases where it clearly indicates a current resultant state. In this respect, Galician is apparently closer to Asturian Spanish than to Portuguese. Santamarina (1974, 161) cites the following examples of ter+agreeing participle from Galician: (43) En tódalas ocasióis que me cuadra deso—dixo—entérome ben, e teño feitos muitos casos deses ‘Whenever it suits me—he said—I find out all about it, and I’ve done that several times’ (44) Teño lidos muitos libros ‘I’ve read a lot of books’ However, according to Santamarina, in these examples feitos and lidos could be replaced by feito and lido respectively. Unfortunately, Santamarina does not give enough detail for us to discover how closely Galician resembles Asturian in this respect, and no other writer comments on this use of the agreeing construction. Use of the invariant and agreeing constructions is interdependent, with one helping to delimit the other. In standard Spanish, where there is no invariant iterative/durative perfect construction, the agreeing construction can indicate repeated past action. 6 In Portuguese, which does have an iterative perfect, ter+agreeing participle is limited to resultant state constructions. In Asturian Spanish, where the invariant construction can express punctual events or events which have no clear link with the present, there is room for the agreeing construction to express repeated or durative action which continues into the present. The situation in Galician is less clear-cut. It seems to be part-way between Portuguese and Asturian Spanish: the invariant construction indicates only repeated or durative events which continue into the present, as in Portuguese; but the agreeing construction can apparently be used to indicate repeated events continuing into the present, as in Asturian Spanish. We have seen that in all the Romance dialects considered here, TENERE+past participle follows a similar pattern of development, but varies in the extent of its evolution. This development is largely the same as that indicated by Harris (1982) for the present perfect in Romance, where he deals primarily with reflexes of HABERE+past participle. It is not only in the temporal/aspectual development that the Romance dialects behave similarly: the semantic development is also apparently similar. Where we have looked at the historical development—in detail in Spanish and Portuguese, and more superficially in Catalan—tener, ter, and tenir appear with the participles of the same types of verbs, and the range of co-occurrence expands in the same way. This development seems to coincide with synchronic preferences too: the participles which historically appeared first with reflexes of TENERE are also those which most readily cooccur in the modern languages. Although developments in the various Romance dialects have presumably been separate (apart, perhaps, from some Castilian influence on Catalan, and possibly crossinfluence in north-western Spain and Portugal) they have all followed the same pattern, developing to greater or lesser extents, and present similar patterns of preference in the modern language. 7
Tener+Past participle
138
5.3 STARE +PAST PARTICIPLE Various people have pointed to parallels between reflexes of STARE+ past participle and TENERE+past participle (see especially, for Spanish, Pottier (1961) and Roca Pons (1968), and for Catalan, Roca Pons (1957–8) and Vallcorba i Rocosa (1978)). 5.3.1 Spanish Considering Spanish first, we have seen that examples of tener+past participle range from cases where it indicates purely current state, through examples with varying relative emphasis on resultant state and causative action, to cases which focus almost entirely on the action. In this last group the action tends to be iterative. Although all speakers accepted some examples of the last group, in general the higher the degree of grammaticalisation in the speaker’s usage the greater the acceptance of this type of interpretation. To suggest that tener+past participle can focus on a past action is not to suggest that that action is not relevant to the present moment, or that there is no resultant state. With constructions such as haber+past participle and tener+past participle, the action is situated at some point within the ‘extended present’. 8 In such cases it is highly likely that the resultant state still exists, that nothing has taken place to change the situation, and that the action can be deemed to have current relevance. 9 Pottier contrasts the resultant state tener+past participle with haber +past participle, which indicates the past action (though still within the extended present) which leads to this state. He also links haber and tener with ser and estar. Ser+past participle forms the passive of the action expressed by the verb in participial form; estar+past participle indicates the state in which the subject currently is as a result of an action which has been performed upon it, although this action is not directly referred to. Ser+past participle is the ‘action passive’; estar+past participle is the ‘resultant state passive’. According to Pottier ‘con ser y haber se forman modalidades fundamentales que afectan al proceso mismo: la voz y la terminación’ whereas with estar and tener ‘se expresa una captación externa del proceso, de tipo resultativo’ (1961, 327). This produces the following arrangement:
Pottier’s arrangement clearly shows the parallel between tener+past participle and estar+past participle. Others have also emphasised this: for example Roca Pons (1958, 121–2) states that: Por lo que afecta al estado, si decimos, p.e., tengo la casa arreglada… lo importante no es la afirmación de nuestra posesión—que se presupone— sino el estado en que se halla el objecto poseído. Afirmamos, pues,
Comparison with other Romance languages
139
entonces, no qué tenemos, sino, en cierto modo, cómo lo tenemos, o, lo que es lo mismo, cómo está. From these parallels, one would expect pairs of related sentences like those in (45) and (46). (45a)
He pintado la casa
(45b)
La casa ha sido pintada
(46a)
Tengo pintada la casa
(46b)
La casa está pintada
As Alonso and Henríquez (1945, 122) point out ‘Con ser la acción del participio ocurre en el tiempo del auxiliar…. Con estar, la acción del participio es anterior al tiempo del auxiliar.’ The use of tenses points to a difference between the two constructions in (45) and the two in (46): he pintado corresponds to the perfect form ha sido pintado, indicating reference to a past event; tengo pintada corresponds to the present tense está pintado, indicating that although there has been a previous action the emphasis is on the present state. But how well do these parallels work? Consider the following sentences: (47a)
¿Has arreglado el coche? Sí, ya está arreglado Sí, ya ha sido arreglado
(48a)
¿Has organizado la excursion? Sí, está organizada para la semana que viene Sí, ha sido organizada para la semana que viene
(47b)
¿Tienes arreglado el coche? Sí, ya está arreglado Sí, ya ha sido arreglado
(48b)
¿Tienes organizada la excursion? Sí, está organizada para la semana que viene Sí, ha sido organizada para la semana que viene
Although ser+past participle may seem a little clumsy at times, in both sets of examples both ser+past participle and estar+past participle are possible. (As pointed out above, there is a difference in the tense used: estar appears in the present tense, while ser appears in the perfect.) But this does not mean that the idea of linking ser/estar and haber/ tener with the question of action and resultant state has to be abandoned. There are cases where one member of a pair is appropriate and the other is not.
Tener+Past participle
(49a)
140
He vigilado esa casa *Esa casa está vigilada Esa casa ha sido vigilada
(49b)
Tengo vigilada esa casa Esa casa está vigilada *Esa casa ha sido vigilada
(50a)
He vigilado esa casa muchas veces *Esa casa está vigilada (muchas veces) Esa casa ha sido vigilada (muchas veces)
(50b)
Tengo vigilada esa casa muchas veces *Esa casa está vigilada (muchas veces) 10 Esa casa ha sido vigilada (muchas veces)
In some cases tener+past participle relates to estar+past participle, in some cases to ser+past participle, and in some it relates to both constructions. Sometimes haber+past participle relates to ser+past participle and sometimes to both ser+past participle and estar+past participle. However, while the parallels between haber+past participle and ser +past participle, and between tener+past participle and estar+past participle are not absolute or exclusive, they do hold good in principle. Examples (47a) and (48a) indicate that in many cases both haber+past participle and tener+past participle correspond to both ser+past participle and estar+past participle. This is because—as was suggested above—even where the emphasis is on the past action the temporal restrictions on the construction (i.e. that it places an event within the extended present) mean that there is usually a resultant state. But tener +past participle can indicate a durative current state interpretation, which haber+past participle cannot. (Compare (49a) and (49b) above: (49a) indicates past action—albeit within the extended present; (49b) indicates a durative current state.) In such cases tener+past participle relates to estar+past participle, and cannot relate to ser+past participle. And where both haber and tener can be used, if there is both a ‘resultant state’ and a ‘previous action’ value tener+past participle has greater emphasis on the state and haber+past participle on the action. Concentrating now particularly on tener and estar with a past participle, we have seen that in general for each example of tener+past participle there is a related estar+past participle example (regardless of whether or not ser+past participle is also possible). But in (50b) above and (51) and (52) below there is no such related construction with estar. (51a)
Me tiene contadas muchas historias
(51b)
*Muchas historias (me) están contadas
(52a)
Tengo recorridos muchos kilómetros
(52b)
*Muchos kilómetros están recorridos
Comparison with other Romance languages
141
Unsurprisingly, it is in cases where tener+past participle moves from indicating resultant state to focusing on the previous action that estar+ past participle cannot be used. We have seen that historically this use of tener+past participle, where emphasis is on the (durative or iterative) action rather than on the resultant state, is a later development. And in the modern language, although all speakers accept some examples of this use, they are generally less tolerant of it than of resultant state usage. There are also cases where estar+past participle is possible and tener +past participle is not. Similarly, these tend to be cases where estar+ past participle has moved on from its ‘traditional’ role as an expression of resultant state and focuses more on the action, as in for example: (53) Estas fotografías están tomadas esta mañana a las cinco (54) La ausencia de acuerdo…ha estado provocada fundamentalmente por las discrepancias sobre el papel decisivo de los interlocutores del Gobierno…y la insistencia del Ejecutivo en negociar exclusivamente con la banda la reinserción de los activistas 11 The presence of a temporal adverbial phrase referring to the time of the action in (53), and the agentive phrase in (54) serve to shift the emphasis on to the action. I am not arguing that examples of this sort are very common, nor that in such examples estar+past participle is purely an ‘action passive’ in the way that ser+past participle is; I merely want to point out that it is in cases where estar+past participle does not have its more usual resultant state interpretation that the parallel with tener+ past participle cannot be applied. 12 To sum up, the tener+past participle/estar+past participle parallel works best for the ‘traditional’ uses of both constructions, where they express, above all, resultant state. The neatness of the arrangement is disturbed by some of the less common, later developments. This point will be taken up again later. At the moment it is necessary only to point out that, while the traditional analysis of tener/estar+past participle has a good deal of truth in it, when all the different uses of the constructions are taken into account, as so often happens, the situation becomes more ‘messy’, and the different uses cannot easily be reduced to a single underlying principle. 5.3.2 Portuguese In Portuguese the situation is similar to Spanish, although the comparisons between ter and estar with an agreeing participle are neater than in Spanish. We have seen (section 4.7) that ter+agreeing participle acts only as a resultant state construction, and does not have an iterative/durative past action interpretation. In this respect it seems that ter+agreeing participle corresponds to traditional descriptions more closely than tener+agreeing participle does. There are also some differences between Spanish estar+past participle and its Portuguese counterpart. In both languages estar+past participle is used to indicate resultant state. But Willis (1965, 363) points out that in Portuguese ‘if the result of a previous action can be considered as lasting or permanent in nature, then it is frequent for estar to be replaced by ser, the participle then functioning like an adjective’. Willis indicates that in such cases there is usually a choice between ser and estar. Among his examples are:
Tener+Past participle
142
(55a) O telhado está/é coberto com telhas ‘The roof is covered with tiles’ (=‘It’s a tile-covered roof’) (55b) O telhado está coberto de neve ‘The roof is covered with snow (only temporarily)’ (56a) A cabeça está/é coberta com cabelo ‘The head is covered by hair (by natural law)’ (56b) A sua cabeça está coberta de chagas ‘His head is covered with sores (presumably curable)’
Pountain (1982, 141) comments that: It is interesting that Portuguese can use ser in precisely those cases which have sometimes been considered to constitute ‘illogical’ uses of estar in Castilian, i.e. where the adjective or adjectival past participle, though representing a ‘resultant state’, also represents an inherent property. In Castilian it is the ‘resultant state’ idea which seems to take precedence; in Portuguese, it is the idea of inherentness. It may be that Spanish estar+past participle is more clearly a resultant state construction than Portuguese estar+past participle, but the difference between them in this respect would appear to be slight. I suggested above that occasionally Spanish estar+past participle has moved on from its resultant state use and focuses more on the action expressed by the verb (as in (53) and (54)). It seems that Portuguese is more reluctant to accept this kind of use, and in this way it again keeps closer to the arrangement in which ter+agreeing participle and estar+ agreeing participle parallel each other as expressions of resultant state. 5.3.3 Catalan For Catalan, Vallcorba i Rocosa (1978) argues that in the medieval language there is a clear correspondence between tenir and estar, but the two verbs are not used in the same way as their Spanish cognates: medieval Catalan estar+past participle does not primarily indicate resultant state, but rather it stresses that the state lasts for a fixed period of time (and not indefinitely). Vallcorba (pp. 54–5) also points out that with inanimate subjects el verb de I’oració és estar quan s’hi expressa que aquella cosa se situa en tal lloc o es manté en tal estat, per la voluntat (la intenció, el propòsit, la finalitat) d’un agent, sempre que aquesta voluntat aparegui d’alguna manera explícita al context…. Quan no apareix l’agent, ni la seva finalitat, l’oració és confegida amb ésser. Vallcorba stresses the primary durative value of tenir and of its compounds (contenir, retenir, detenir, sostenir, etc.). He concludes (pp. 105–6) that: El verb estar expressa en passiva allò que el català medieval diu amb el verb tenir essent el subjecte de l’oració un agent. Subratllem-ho: tot subjecte agent que té alguna cosa (en els sentits del verb medieval tenir)
Comparison with other Romance languages
143
pot ser dit mitjançant una altra oració equivalent a la qual aquella cosa (complement directe de l’oració amb tenir) és el subjecte de l’oració amb estar. The correspondence between tenir and estar is based primarily on the notion of duration rather than that of resultant state, but the parallel still holds. In modern Catalan there are indications that estar+past participle is coming to be used more as an indication of resultant state, as in Spanish. Badia Margarit (1962b, 145ff.) stresses the temporal criteria for use of Catalan estar, and contrasts its use with the ‘resultant state’ use in Spanish: Se expresan con ésser las cualidades permanentes del sujeto, aquellas que tienen valor imperfectivo…, se expresan con estar las cualidades transitorias o accidentales, que tienen valor perfectivo…. Pero no vale para el catalán la distinción [in Spanish]…entre las cualidades enunciadas en sí mismas y sin consideración de su origen y procedencia (con ‘ser’), y las cualidades como resultado de una acción, transformación o cambio que sabemos o imaginamos que ha tenido, tiene o tendrá lugar (con ‘estar’). (p. 153) However, Badia goes on to say that Catalan estar is increasingly used, especially in the spoken language, with past participles; the examples he gives (e.g. vacillation between Aquest gerro és/està trencat ‘This jug is broken’) suggest that use is moving towards that of expressing resultant state (leaving ésser+past participle as the ‘action passive’). Yates (1975) expresses a similar point of view. He makes the following basic distinction between esser and estar: Ésser expresses a condition or quality considered as inherent, continuous, unchanging or permanent. The verb itself presents no prominent idea of temporal limitation, change, beginning or end. Because of this ésser is said to be imperfective. Estar denotes qualities in which we are interested within a temporal context, realized at a particular moment in time. In the background are such ideas as stage, state or stay, and these cognate terms can provide a convenient guide to the perfective meaning and use of estar. (pp. 94–5) But he goes on to say (p. 102) that ‘the distinction is less easily made when the adjective concerned is a past participle. In modern usage estar tends to be used with such adjectives, even when the context is imperfective.’ Although Catalan estar+past participle does not correspond precisely to Spanish estar+past participle, there is still a parallel between estar+past participle and tenir+past participle. Vallcorba has shown this in some detail for medieval Catalan, when the opposition between them was not the same as between Spanish tener/estar+past participle. It seems that there is still a parallel between them in the modern language, but that the situation is moving closer to the situation in Spanish, with resultant state playing a more significant role. 13
Tener+Past participle
144
5.3.4 Italian Standard Italian shows a much more limited use of stare+past participle (see Pountain 1982; Peral Ribeiro 1958, 161–2). While stare has a wider range of use than its Latin etymon, it has not developed to the extent that it has in the Iberian peninsula. In particular, while stare may appear with some adjectives, to indicate a temporary state, Italian has not developed a resultant state passive with stare. We saw above that a very similar situation obtains with tenere: this verb too has broadened its range much less than in the Iberian peninsula. Tenere does appear with some past participles, but it nearly always retains something of its lexical meaning. In Italian, neither tenere nor stare has developed to the extent that their cognates in the Iberian peninsula have, and Italian has not developed the sort of resultant state past participial constructions which we have seen in Spanish, Portuguese, and Catalan. But we saw earlier that in some southern Italian dialects tenere has undergone greater expansion than it has in the standard language. This is also true of stare. Peral Ribeiro (1958, 163–5) provides various examples in which standard Italian uses essere and the Southern dialects use stare. He concludes that ‘observa-se a preferência do napolitano por stare para, ligado a adjectivos, indicar situação passageira…: para significar o “encontrar-se” em determinado lugar…. ou, com uma “nuance” mais leve, simplesmente o “lugar onde”.’ He notes that in these dialects stare does not regularly appear with past participles, but he provides some examples where there is hesitation between essere and stare: (57a)
O tratturo è [essere] ’mfusso ‘The path is wet’
(57b)
Le vie stanno [stare] tutte ’mfosse ‘The roads are all wet’
(58)
O piatto è/sta [essere/stare] sengato ‘The plate is cracked’
(59a)
A botteglia è [essere] chiena ‘The bottle is full’
(59b)
A botteglia sta [stare] chiena ’e vino ‘The bottle is full of wine’
(60a)
A camisa è [essere] lorcia ‘The shirt is dirty’
(60b)
A casa sta [stare] tutta lorcia ‘The house is all dirty’
(61)
A porta è/sta [essere/stare] aperta ‘The door is open’
As Peral Ribeiro points out, such hesitation is found in Old Spanish and Old Portuguese texts. The southern Italian dialects seem to have moved further than central and northern Italian in their development of stare, towards the situation which prevails in the Iberian peninsula, although the opposition between essere and stare is not so well developed as in Spanish, Portuguese, or Catalan. While there is no clear resultant state category in these southern dialects, we find that both tenere and stare have developed further than in other Italian dialects.
Comparison with other Romance languages
145
5.3.5 TENERE AND STARE We have seen that participial constructions with reflexes of both TENERE and STARE have developed to differing extents in different Romance languages. We have also seen that there are parallels between the two constructions. In each language the development of one construction seems to be matched by the development of the other construction. The parallel between the two is not exactly the same for each language, neither is there a perfect match between the two constructions in any one language. This is what we would expect. Much of the information put forward throughout this study has pointed to the conclusion that linguistic description is seldom as neat as some linguists would have us believe; and that accurate descriptions of usage can rarely be reduced to a few broad categories. However, the notion of a parallel between tener and estar is an attractive one, as is that of oppositions between haber and tener and between ser and estar. Is there a link between tener and estar, or is this merely an invention of the linguist, keen to impose on the language a structure and order which it does not necessarily possess? The reflexes of both TENERE and STARE with a past participle indicate resultant state—or in the case of medieval Catalan, duration. With TENERE+past participle, the agent is expressed as the subject of TENERE. With STARE+past participle, the agent is not usually expressed, and a type of passive is formed. It is quite plausible that when a language begins to develop a category of resultant state it should then exploit this category more fully. Initially, the Romance languages develop a resultant state category through the rise of HABERE+past participle (see Bassols de Climent (1951); Harris (1982)). In most languages HABERE+past participle moves on from this, and in many cases its resultant state function is to a greater or lesser extent replaced by TENERE + past participle. Once these languages have the notion of resultant state as a separately expressed category, then it is unsurprising that both active and passive forms should develop. While there need be no direct causal link between the presence of TENERE+past participle and that of STARE+past participle as expressions of resultant state, it could well be that the presence of one strengthens the presence of the other by reinforcing the category of resultant state. As either one develops and strengthens its position in the language, this will encourage the development of the other, and there is mutual consolidation of the two constructions. The two forms may also be reinforced by the contrast they provide with HABERE+past participle and *ESSERE+past participle. Taking Spanish as an example, we see that while tener+past participle and estar+past participle may both develop independently of haber+past participle and ser+past participle, once tener/estar+past participle begin to appear in the same contexts as haber/ser+past participle an opposition arises between tener+past participle and haber+past participle, and between estar+past participle and ser+past participle. Once an opposition exists in each case between a ‘resultant state construction’ and an ‘action construction’, the opposition reinforces the existence of each of its members. In chapter 3 we saw how historically tener appears with the participles of certain verbs—principally those which involve the idea of having something in one’s power. When it begins to appear with the participles of these verbs it contrasts with haber, which already occurs with a range of past participles. The extension
Tener+Past participle
146
of tener+past participle from this context of having something in one’s power to other contexts could occur independently of other constructions—we have seen its progression through different semantic and syntactic classes of verbs. But it could be encouraged by the contrast it forms with haber+past participle. Once haber+past participle and tener+past participle contrast in certain contexts, why should this contrast not be extended to other contexts? The existence of a contrast between haber+past participle and tener+past participle in certain contexts may encourage the extension of the tener construction to other contexts where haber is already found. Once again, this process need not be viewed as one of direct cause and effect, but of encouraging and strengthening a process which has already begun. Pountain (1982) highlights a similar situation with *ESSERE and STARE: The extension of estar to past participle complements in Castilian, Portuguese and Catalan makes possible the development of a systematic opposition between *ESSERE and STARE in the form of the action passive and the resultant state passive Once the systematic opposition between *ESSERE and STARE is reached, it is natural that these languages should capitalise on it. (pp. 157–8) Pountain also points out that the history of estar is one of encroachment on its partner in the opposition, ser. Once again, there are similarities with tener/haber. And in both cases it seems that this encroachment is still in progress. Most ‘textbook’ descriptions of use of ser and estar make no mention of uses of the type exemplified in (53) and (54) above. 14 While allowing for the fact that descriptions tend to lag behind reality, and for the prescriptive nature of some grammars, it seems that this sort of usage is a relatively recent phenomenon, or at least is currently on the increase. These developments occur after the situation described in most grammars has been reached. Tener+past participle has received much less attention than estar+past participle, and so it is harder to compare current usage with usage as described by grammars. But the historical survey of the constructions which I have undertaken, and the relative tolerance of different uses of the construction in the modern language point to uses where tener+past participle does not have its traditional resultant state interpretation being a later development. We have seen the similarities in resultant state use of tener+past participle and estar+past participle. In their later development are the two constructions evolving in the same direction? The answer to this question appears to be both yes and no. Both have primarily a resultant state interpretation and are moving to encompass cases where the emphasis is on the action. Estar+past participle is moving towards the ‘action passive’ usually denoted by ser+past participle; tener+past participle is moving towards the iterative perfect, apparently following the pattern of development earlier undergone by haber+past participle. But while both constructions can emphasise past action, they differ in that in such cases tener+past participle indicates repetition, whereas estar+past participle does not. However it is always difficult to evaluate changes which are still in progress, and it remains to be seen whether tener+past participle and estar+past participle will continue to develop along similar lines, as ‘active’ and ‘passive’ (approximate) counterparts, or whether their paths will diverge.
Comparison with other Romance languages
147
5.4 SUMMARY In this chapter we have looked at the development of reflexes of TENERE +past participle in various Romance languages. While there is less information available on the development of tenir/tenere+past participle in Catalan and Italian than on that of their Spanish and Portuguese cognates, what information there is suggests that all the languages follow the same basic development. The languages of the Iberian peninsula have taken this development further than other Romance languages; but examples from Italian indicate a similar, albeit more restricted development there, especially in southern Italian dialects. As well as looking at TENERE+past participle, I have suggested that there is a parallel with reflexes of STARE. This parallel may not be as exact or as tidy as the system proposed by Pottier, but I would suggest that it is more interesting. Reflexes of TENERE+past participle and of STARE+ past participle do not form exact active/passive counterparts, but they both develop towards expressions of resultant state (reaching this situation in Spanish, Portuguese, and probably Catalan, but not developing to this extent in Italian). While there is no need—or evidence—to suggest a direct causal link, in each case the position of one resultant state construction would be reinforced by the existence of another such construction. This reinforcement of the category of resultant state strengthens the position of each of its exponents. While this is the principal comparison I have drawn between tener and estar, there are other interesting similarities. Both verbs exemplify the encroachment of one form on another (haber and ser respectively), to the extent that oppositions are formed when both forms appear in the same context; this in turn encourages the extension of the opposition to other contexts. Both tener+past participle and estar+past participle are apparently unstable in the modern language, providing good examples of the need not just to incorporate new uses into descriptions, but of the impossibility of forcing new uses and newly developing categories into the framework of older descriptions: while diachronically these new uses can be seen as part of a single development with older usage, synchronically they have to be accepted as separate uses and descriptions have to be modified to accommodate them.
6 Conclusions 6.1 OUTLINE In this study we have looked at modern use of tener+past participle and at its history. We have compared this with the use and history of its Portuguese cognate and have considered cognates in other Romance languages. Such a study has not only provided an examination of TENERE+past participle, but also bears upon broader issues of linguistic description: what we mean by the term ‘auxiliary’, and how constructions—especially those on the periphery of ‘the system’—are to be classified; what effect changes in the role of a construction have on the structure of the language and the way it is described; how we deal with variation in language, and the part that such variation plays in diachronic studies. This final chapter aims to tie in these more general issues of language description and language change with the study of tener+past participle, and to suggest some possible directions for future research arising out of this work.
6.2 VARIATION IN LANGUAGE In chapter 2 I argued that speakers could be ordered according to the degree of tolerance they showed towards tener+past participle, and that sentences could be ordered according to their degree of acceptability. I also argued that variation in use of tener+past participle could be related to the historical development of the construction. These claims need to be examined further. Various people have pointed out that variation between speakers can be accounted for within linguistic theory. Weinreich et al. (1968, 99) state that one of their aims is ‘to reconcile the observed facts of linguistic heterogeneity with the theoretical desiderata of finding order and structure’. One of the conclusions they come to is that ‘the association between structure and homogeneity is an illusion. Linguistic structure includes the orderly differentiation of speakers and styles’ (pp. 187–8). Weinreich et al. are principally concerned with accounting for variation by reference to social factors such as class, and then linking this to possible diachronic developments. Others have concentrated more on variation as an internal linguistic phenomenon. Elliott et al. (1969) base their arguments on the premise that variation, particularly of the very subtle types which exist among speakers who apparently have the same dialect, must be considered part of our data, because variation is a fact, and any theory of language which ignores it cannot be as complete as one which does not. Furthermore, there are facts
Conclusions
149
both about linguistic theory and about the grammars of particular languages whose existence will be obscured unless variation is taken into account…. There are regularities underlying what appear to be chaotic disagreements. (P. 52) Elliott et al.’s survey of two cases of syntactic variation (deletion in subordinate clauses and replacement of a verb phrase by do so) indicates that variation between speakers is not random, but can be ordered on an implicational scale, according to the acceptability of different sentence types. Coleman (1973) also set herself the task of ascertaining whether or not syntactic variation follows any logical pattern. She found that for some of the items tested— complementiser restrictions, extraposition from NP, relative-clause formation, and some passive formations—it was possible to produce a continuum, with informants ranked from most conservative to most liberal, and with sentences ranked from most acceptable to least acceptable. She concludes that ‘the defining characteristic of all this variation is the difference of intuition as to where to draw the line of acceptability One finds very little disagreement on the order of the sentences’ (p. 82). Coleman also found some phenomena—sentences involving there-insertion or a definite pronoun referring to an indefinite antecedent—which could not be ranked as a continuum. Informants certainly differed as to which sentences they regarded as acceptable, but their judgements did not produce a ranking either of informants from least to most tolerant, or of sentences from most to least acceptable. Coleman concludes that while certain variations can be described by a continuum model, other variations, while systematic, require an assumption of dialect variation which does not produce a continuum. Returning to tener+past participle, we have seen that this is a construction which does lend itself to the continuum approach. Levels of acceptance can be hierarchically ordered, and variation between speakers is systematically accounted for. But this does not automatically mean that such variation corresponds to diachronic developments. We have ruled out factors such as age, social class, and geographical area. But Carden (1973) posits the existence of dialects in which different response patterns from speakers can be systematically ordered but which do not correlate with non-linguistic characteristics such as geographical area or social class, or by implication, diachronic developments. He claims his results support the hypothesis that ‘the difference in response is caused by a difference in internalized grammar, but the different grammars are distributed randomly among the informants’ (p. 6). While systematic variation between speakers may well be related to linguistic change, we cannot just assume this to be the case. As Weinreich et al. put it ‘Not all variability and heterogeneity in language structure involves change; but all change involves variability and heterogeneity’ (1968, 188). However, in the case of tener+past participle, we are not just assuming that synchronic variation must be related to diachronic change. Textual study has indicated that the stages through which tener +past participle passes are ordered in much the same way as the different degrees of tolerance for the construction found in the modern language. There is thus independent evidence to suggest that there is a link between the two. Where such a link can be found it seems sensible to use our knowledge of the modern language to help
Tener+Past participle
150
piece together our fragmentary historical knowledge. We should still proceed with caution, and continue to look for evidence within the texts; but where modern variation can be shown to be similar to historical developments, contemporary knowledge can suggest profitable lines of historical enquiry, and help us to interpret what we discover. There are two main ways in which we can try to relate synchronic variation and diachronic change. We could posit a causal link: we could assume that the variation found among speakers is the result of changes currently working through the language— as the changes move through the speech community, at any one point some speakers will have been more affected by these changes than other speakers. Or we could assume that the varying tolerance of contemporary speakers mirrors the historical development, without being caused by it: that perhaps there is some underlying factor which means that they pattern in the same way. If we take the link to be causal, we could argue that changes involving tener+past participle are currently moving through the language and through the speech community, and have affected different speakers to differing degrees. However, we saw in chapter 3 that tener+past participle had apparently reached its current stage of development by about the sixteenth century and has remained more or less stable since then. There is little evidence to suggest that it is currently undergoing change. We could then modify our hypothesis, and argue that in about the sixteenth century the change froze at the stage we see now, and that the effects of this change, which failed to reach completion, are preserved in the differences found among contemporary speakers. This is a possibility, but it is odd that the changes undergone by tener+past participle should not have affected all speakers by now, given the length of time since the changes first appeared. If the changes were to peter out before reaching all speakers, one would expect them to do so along geographical or social lines, but this is not the case here. Should we then go back to the idea of ‘randomly distributed dialects’? The answer to this question is both yes and no. The dialects are randomly distributed in that the distribution cannot apparently be related to extralinguistic factors. But we cannot get away from the fact that the different levels of tolerance shown by speakers parallel different historical stages in the development of the construction. Let us consider our second main hypothesis: that the reason for this synchronic/diachronic parallel is not causal, but relates to some more general underlying principle. We have seen how historically tener+past participle affects some categories before others; for example, transitive verbs are affected before intransitive ones, and within the transitive category, verbs with some notion of keeping something in a particular state are among the earliest to appear in the construction. So some categories are more susceptible to the changes than others are. We can list categories in chronological order of appearance in the construction. Those categories which appear first in the list are those which lend themselves most easily to the construction. Here the categories are ordered diachronically, but if some categories are more likely to appear in the construction than others, there is no reason why this should not also provide a synchronic order, indicating the relative acceptability of different categories by different speakers. The chronological order implies that some categories are more likely to appear in the tener+past participle construction than others. We have seen that synchronically in the modern language certain categories are more acceptable than others. We should not be surprised that the diachronic and the synchronic orders of acceptability should be the
Conclusions
151
same. Such an argument does not claim that current differences in tolerance among speakers are the direct result of historical change. Instead it suggests that both historically and synchronically certain categories are more susceptible to co-occurrence with tener. 1 The survey carried out in chapter 2 suggested that even within one narrow social group there is considerable variation in the use of tener+past participle. By concentrating on a small part of the speech community, certain social factors were eliminated. Other factors, such as the age of the informants, did not seem to influence the results. Other surveys have suggested that studies of linguistic variation and language change must take into account social and geographical variation (see, for example, Labov (1972b); Trudgill (1974); Trudgill and Chambers (1980)). The present survey suggests that we must also take into account the fact that people’s linguistic tolerance varies. It may be possible to a certain degree to link this to social or intellectual background, but partially at least it seems to be a question of individual variation. Some people are more prepared than others to ‘stretch’ constructions, to find suitable interpretations for them and to accept unusual or less frequent examples. It seems likely that it is through this sort of innovation (which may well be subconscious) that a construction develops new uses or extends old ones. And in this case individual tolerance and individual variation have to be taken into account. On the other hand, we could argue that differences between speakers do not imply any intrinsic order, but are the result of speakers having been exposed to different examples of the construction at a time when they were forming their own grammars. As some of the uses of tener+ past participle are quite rare, perhaps some speakers never came across them and consequently did not accommodate them in their idiolect. If we accept this view, we have to explain how learners, hearing a random selection of examples of tener+past participle, produce an ordered pattern of what is possible, for such a pattern implies a hierarchy of acceptability. We can answer this problem by suggesting that while some speakers come across both ‘conservative’ examples—such as tener entendido or tener previsto—and ‘extreme’ examples—such as tener mandado or tener dado—other speakers hear only the more conservative examples. This, however, is tantamount to saying that ‘conservative’ examples must be more common than ‘extreme’ ones. Otherwise speakers would stand an equal chance of hearing ‘conservative’ or ‘extreme’ examples, and there would be no systematic difference in usage between speakers of the sort which we have seen in this study. (If we reject the notion of an intrinsic order of acceptability we cannot argue that speakers who hear ‘extreme’ examples assume that ‘conservative’ examples are also possible, whereas if someone hears a ‘conservative’ example nothing leads them to posit more ‘extreme’ ones.) We therefore still have to explain the origin of the difference in relative frequency. Again we are back to the question of ordering the examples. 2 To sum up, the variation found in synchronic tener+past participle can be hierarchically ordered and this variation parallels the historical development. I have argued that such a similarity does not presuppose a causal relation, but reflects differences in the compatibility of certain verb types with tener, which can be used to postulate both a synchronic and a diachronic hierarchy of acceptability.
Tener+Past participle
152
6.3 WHAT DO WE MEAN BY ‘GRADUAL CHANGE’? Throughout this study I have talked of changes gradually moving through the language. But as Bennett (1981) has pointed out, we need to define what we mean by gradual syntactic change. Bennett distinguishes five parameters of change: syntactic, strategic, environmental, lexical, and sociogeographic. There is syntactic gradualness ‘where the syntactic status of some element changes gradually, e.g. from one part of speech to another or from one grammatical relation to another’. A strategy is defined as ‘a way of expressing a particular meaning’; Bennett gives as an example the use of will, which is one strategy for expressing ‘future time’ in English. Strategic gradualness ‘refers to a situation where there is no sudden switch from strategy x to strategy y, but instead there is an intermediate stage where a speaker uses both x and y, in whatever proportions’. As for environmental gradualness ‘a change is environmentally gradual if it occurs in some syntactic environments before others (e.g. in main before embedded clauses)’. A change is lexically gradual ‘if some lexical items are affected before others’. It is sociogeographically gradual ‘if it is adopted by speakers of some variety before others’. Bennett provides a general definition for gradualness which applies to all parameters: ‘a change is abrupt if logically possible mixed or intermediate stages are not attested, and gradual if such stages are found’. What picture do we get if we apply these parameters to tener+past participle? We have seen syntactic gradualness as measured by Lehmann’s parameters, which chart the progress of a form from one grammatical class to another. We saw from Lehmann’s criteria that syntactically the development of tener+past participle is clearly gradual. This needs no further discussion here. Bennett hypothesises that strategic change is always gradual, as ‘abrupt strategic change would involve a speech community’s switching “overnight” from one strategy for expressing some meaning to another. As communication between generations must be preserved, such a situation is out of the question’ (1981, 126). When comparing the developments of tener+ past participle and haber+past participle I pointed out the overlap in use between the two forms. Even in the modern language there is a partial overlap (see section 5.3). This dual strategy during a ‘changeover period’ is seen more clearly in Portuguese, where haver+past participle and ter+past participle coexist more closely than in Spanish, and where finally the newer strategy—ter+past participle—takes over from the older haver+past participle. So far we do not have enough evidence to test environmental gradualness. Textual examples of tener+past participle—particularly in the more ‘auxiliary-like’ roles—were too infrequent to be profitably classified according to syntactic environment. My initial impression is that syntactic environment does not influence use of tener+past participle, but more work needs to be done in this area. What Bennett describes as lexical gradualness has been discussed in earlier sections. We have seen that tener+past participle exhibits lexically gradual change. I have stressed that this survey does not attempt to cover social factors, and so it provides no information for the fifth of Bennett’s parameters—the sociogeographic one. In terms of geographical variation, we have seen how use of tener+past participle by the Oviedo informants (and almost certainly throughout north-western Spain) is different
Conclusions
153
from that of other speakers. However, this is not an indication of geographically gradual change: there is nothing to suggest that a change has begun in the north-west and will gradually spread into the rest of Spain. It seems that the differences observed in Oviedo are due to influence from other dialects, and that this influence is restricted to the northwest. I am not discounting the possibility of sociogeographic gradualness, merely pointing out the lack of evidence. It would be very difficult to spot social variation indicative of change in the texts, and the possibility of such variation in the modern language was not investigated. Changes in tener+past participle are gradual according to three of Bennett’s parameters: syntactic, strategic, and lexical. As yet, there is not enough information to evaluate changes by the two other parameters: environmental and sociogeographic. Work remains to be done in these areas to see how they fit in to the overall pattern of change.
6.4 OVERVIEW Tener+past participle—and in particular its usage in the modern language—has not previously been studied fully, and to this extent an examination such as the one presented here is justified in its own right. But throughout this study reflexes of TENERE+past participle—and in particular tener+past participle—have been used as an example of some of the problems faced in linguistic study. In summing up, I would like to point to three main areas: description and classification; the importance of a verb system and of the interplay between members of that system; the relation between synchrony and diachrony. In the area of classification, the study of tener+past participle supported the view that approaches which allow for graduated membership of classes are more satisfactory than ‘all or nothing’ approaches. I have argued against the assumption that if not all members of a category display all the appropriate defining characteristics, then that particular category should be rejected. Some members of a class may be more central than others, exhibiting more of the characteristics of the class than other more peripheral members do. Particular theoretical approaches may not require a category labelled ‘auxiliary verb’, but in more general terms it is still a useful notion, even if some verbs are more auxiliary-like than others. For those linguists interested in the history of the language there are additional diachronic reasons for having a class of auxiliary verbs. Those verbs in Romance which are usually labelled as auxiliaries differ semantically and syntactically from their Latin etyma, and as they move away from Latin usage they apparently follow similar patterns of grammaticalisation. It is useful then to group these verbs together because of the diachronic characteristics which they have in common and which set them apart from other verbs. (Though, once again, some verbs are more characteristic of this process than others.) Similar arguments hold synchronically. We have seen that in the modern language tener+past participle exhibits a range of uses, some of which are more grammaticalised and have more in common with the group of traditional auxiliaries than others. We need a framework which enables us to accommodate and to account for such differences. The main description of tener+past participle was put forward irrespective of whether particular uses exemplified ‘auxiliary’ or ‘main verb’ usage, and we saw how tener+past participle exhibited a range of different uses which did not behave in the same
Tener+Past participle
154
way syntactically or semantically. A descriptive framework has to be able to take these differences into account. It is not only the different uses of tener+past participle which demand a flexible classification: differences between speakers point to the same conclusions. While—as was noted in section 2.8—it is not theoretically necessary for the speech of different speakers to be incorporated into one description, where different speakers’ uses of a construction are based on the same principles and can be hierarchically ordered, it is advantageous if the uses can be accommodated within a single classification. We have also seen how the changes undergone by tener+past participle are gradual rather than abrupt. This too calls for a graduated classification. A form or construction does not suddenly move from one category to another: it moves along a scale, and there is not necessarily a clear cut-off point where it can be said to change categories. All these factors support a graduated system of classification. Turning now to the idea of a verb system, we have seen how newly grammaticalising forms enter the system on their own terms. Nothing is to be gained by distorting our description to make these forms fit a pre-existing system. We have to be prepared to accept new temporal, aspectual, or modal categories, or new combinations of tense, aspect, and mood. We have also examined how forms interact with each other; how they may create new oppositions as well as new categories. Through these oppositions, one form may affect the development of other forms: for example, once two forms are perceived to be in opposition because they contrast in some contexts, this may encourage the expansion of the less widely used form so that it begins to appear in more contexts where the commoner form is already found, thereby increasing the number of contexts where both forms are found and strengthening the opposition. We saw that this may have happened with tener and haber (see section 5.3.5). Or the expansion of a new form may strengthen a particular category, thereby encouraging the expansion of another form which has similar characteristics (see the discussion of TENERE and STARE in section 5.3.5). We need some notion of a verb system or paradigm to account for these types of influence by one form on another. But our system needs to be sufficiently flexible and open-ended to allow for the appearance of new forms and the rearrangement of existing ones. The interrelation of synchrony and diachrony has been a constant theme of this study. We have already discussed the introduction of a diachronic element into classification; while the classificatory needs of the synchronic and the diachronic linguist may be different, where it is possible to combine the two such a combination must be desirable, as it provides a more unified approach and removes the need for separate descriptions. This is one of the main advantages of a framework such as the one put forward by Lehmann (see section 2.8). This study has also examined variation between speakers, and considered how this synchronic variation relates to diachronic studies. Comparison of varying tolerance of speakers in their use of tener+past participle with the historical development of the construction has suggested that synchronic variation can not only provide an indication of changes currently working through the language, but can also be a useful parallel with previous diachronic development. In this way knowledge of the contemporary language can help in our study of the history of the language.
Conclusions
155
In all the areas which have been examined, specific examples have been brought to bear on more general problems. The whole study rests on detailed examination of the use of one construction. While we should be looking for theories which account for as much of our knowledge as possible and which provide as broad an explanation as possible, we must be prepared to test those theories with specific examples. Such testing may destroy the neatness of a particular theory, and some details may lead to a messy picture rather than a tidy theory. But while we need a descriptive framework within which to work and to order our knowledge, such a framework must be flexible enough to accommodate whatever ‘awkward’ examples we may come across. It is hoped that this study of tener+past participle has provided some of the sort of details with which such theories must deal.
6.5 DIRECTIONS FOR FURTHER RESEARCH Some aspects of this study call for further development. I shall highlight three of them here. Points one and two relate specifically to tener+past participle. Firstly, I have stressed that this is a study of what is possible, not how common different uses are; neither has it taken into account differences relating to social group or register. While there are problems attached to sociolinguistic surveys of syntactic variation, the study carried out here could provide the basis for a survey of this sort. Such a survey might provide more information about whether or not there are changes currently affecting the construction. It would be particularly interesting to investigate further the use of tener+past participle in Oviedo: usage in this region has been mentioned in this study, but a much wider survey is called for. In particular, the relation between standard Spanish tener+past participle and the regional construction needs to be examined. The informants I used all claimed to be monolingual speakers of Spanish and not to speak the local bable—although, as was pointed out in chapter 5, it is a moot point whether any sensible dividing line can be drawn between bable and regional use of Spanish. However, it would be interesting to see whether use of the construction by people who do consider themselves speakers of bable differs from the use I have recorded of non-bable speakers and, if it does, what its sociolinguistic significance is. The second point concerns the history of tener+past participle in Galicia and in Asturias. In this study I have proceeded on the assumption that developments in these areas have followed similar paths to that of Portuguese ter+past participle. But we have seen that the constructions in the different languages are not equivalent, and it would be interesting to discover how and when these differences arise. A thorough examination of the history of tener+past participle in these regions would provide more information on the development of the perfect in Romance. The third point is more general, and concerns the relation between synchronic variation and historical development. Labov and others have shown how synchronic variation (usually relating to social factors such as class) can give an indication of changes which are currently working through the language. In section 6.2 I put forward a possible link between variation and change where there is no indication that the situation is currently changing, suggesting that variation in speakers’ tolerance of a construction may parallel historical development without being caused by it. More examples of
Tener+Past participle
156
synchronic variation and diachronic development need to be compared. Further research in the first two areas I have outlined could help research on this point: if, as suggested above, a sociolinguistic study of tener+past participle were to provide more information as to whether or not changes are currently under way in this area, this would add to our knowledge in this third area to be considered; and more information about both the history and the current use of the construction in Galicia and Asturias would provide a useful comparison with what we have found out about standard Spanish. It is hoped that this study will provide a useful starting point for these paths of future research.
Notes Introduction 1 Roca Pons’ (1958) survey is far more detailed, but has no section devoted specifically to modern usage. He classifies examples by semantic and aspectual criteria, but not by date, and so examples from different periods are grouped together with little indication as to what is or is not possible in modern Spanish. 2 There are, of course, exceptions to this general picture: see, for example, Green’s (1982) work on Spanish passive constructions. 3 As Comrie points out, in giving such a definition he is following Holt (1943, 6), who groups under aspect ‘les manières diverses de concevoir l’écoulement du procès même’. 4 Aktionsart is often translated into French mode d’emploi or Spanish modo de acción. Gili Gaya (1955, 131–2) uses the term significado to indicate Aktionsart and aspecto for Aspekt. 5 As this study does not claim to provide an exhaustive survey of aspect in Spanish I shall ignore possible distinctions such as durative/progressive, or iterative/habitual, and consider only those aspects which are directly relevant to the verb forms and constructions to be studied here. 6 For a discussion of the various different approaches to the perfect and of the main exponents of each theory, see McCoard (1978). While McCoard concentrates on the perfect in English he includes illustrations from other languages, including Spanish, and much of his argument applies equally well to Spanish as to English. 7 He does not mention whether this example is unacceptable to any speakers, and so we are presumably to view it as being widely accepted.
1 The question of auxiliary status 1 See also Hamplová (1968, 209) and Carratalá (1980, 168–9). For semantic loss as a criterion for auxiliary status in French see Damourette and Pichon (1911, sec. 60), Gougenheim (1929, 1), Grevisse (1946, 463), Cohen (1960). For Portuguese see Dias da Costa (1976). For a critical summary of approaches to the definition of auxiliaries in Portuguese, including the criterion of semantic loss, see Pontes (1973, ch. 1). 2 Apart from the introductory remarks, Hernanz’s 1980 article is reproduced almost word for word in chapter 5 of her 1982 book, from which I shall quote. 3 For a criticism of Gili Gaya’s treatment of periphrastic constructions, see Lope Blanch (1962, 419–21) 4 On the multivalency of auxiliary verbs, and of lexical forms in general, see the comments in Cohen (1960, 435–6). 5 A similar problem is noted by Joos (1964, 30) in considering candidates for auxiliary status in English. 6 Most linguists do limit periphrases to purely verbal constructions, but some include other types of constructions. Dias da Costa (1976) includes other combinations in his discussion of periphrasis in Portuguese: he defines verbal periphrasis as ‘the expressions composed of a
Notes
158
grammaticalized verb, in the sense of verb with total loss of objective meaning, and a nominal form of a conceptual verb (even though an adjective, a noun or an adverb may occur in its place) which furnishes the meaning’ (p. 191). 7 For discussion of Palmer’s approach to defining auxiliary verbs, see section 1.5. 8 For a general summary of which verbs are classified as auxiliaries by the major Spanish reference works—including classification of tener—see Green (1982, 97–102). 9 For a similar view see Carratalá (1980, 171–2). 10 The position of the past participle—i.e. whether it comes before or after the direct object—is irrelevant for the purpose of these tests. 11 For a fuller discussion of the interrogatives see section 2.6. 12 Klein discounts selectional restrictions as a criterion, as he claims that this criterion is invalid in the case of verb-object restrictions. 13 As Klein rejects this criterion, his classification is of course not affected by this. 14 Although it is not directly relevant to the analysis of tener, it is worth pointing out that Fontanella and Klein disagree in their classification of the so-called ‘modal auxiliaries’. 15 For example, although Tesnière’s (1939) analysis effectively treats each element of a periphrasis in isolation—the auxiliary contains the grammatical information and the nonfinite verb form expresses the semantic value—Tesnière does recognise the unitary character of such constructions. Benveniste (1965) criticises Tesnière’s analysis for not concentrating on this unitary nature of the construction. For Benveniste both elements have a double function—lexical and grammatical—and the construction must be considered as a single unit; if one considers Il a frappé as Il a (grammatical element) and frappé (lexical element) it is impossible to explain how the present tense form Il a expresses a past or perfect value. 16 For a similar list of criteria, with exemplification and comment, see Quirk et al. (1985, 120ff.), who also include position of quantifiers such as all, both, each, and loss of selectional restrictions. Like Palmer, they also posit categories of semi-auxiliary and catenative. 17 Note that Palmer is not using the term ‘verb phrase’ in the way it is usually used in generative grammar; in Palmer’s usage ‘verb phrase’ refers only to the verbal element and does not include objects and adjuncts. 18 It should be remembered when considering any generativist treatment of auxiliaries that the category ‘Aux’ of such an analysis is not equivalent to ‘auxiliary verb’, although auxiliary verbs are frequently exponents of this category. As far as possible, I shall consider the general points which apply to the question of auxiliary verbs, rather than discussing the category ‘Aux’. 19 It is interesting, however, to note Huddleston’s rejection of the paradigm test (which, as indicated above, could also be used as a test for auxiliaries in Spanish). According to Huddleston, a construction such as has taken cannot be regarded as a single unitary member of a paradigm in the way that, for example, Latin amavi—a single inflected form—is a member of a verbal paradigm. He argues that while took is a single segment, has taken must be thought of as two segments, pointing out that the two segments need not even be contiguous (e.g. has never taken). A similar argument could apply in Spanish. For a discussion of the relevance of this point, and of the distinction between morphological and syntactic structures in paradigms, see section 2.8, and in particular the outline of Vincent’s (1987) arguments on pp. 90–1. 20 Criticism of the generativists’ failure to recognise the existence of nonsegmented continua has been voiced by others, outside the context of auxiliary verbs. See, for example, Bolinger (1961). 21 For a counter-argument, see Hadlich (1971). 22 Salvi (1987) makes a similar distinction between syntactic-semantic auxiliaries and semantic auxiliaries. The English auxiliaries are syntactic-semantic ‘because they differ in their syntactic behavior from (most) other verbs’. Romance auxiliaries are semantic auxiliaries
Notes
159
‘because their syntactic behavior is generally not distinct from that of the other verbs and, if some differences exist, these are…strictly correlated with the semantics of the auxiliaries’ (p. 234).
2 Tener +past participle in modern Castilian 1 For a general discussion of the way in which context affects judgements on grammaticality see Bolinger (1968). 2 It is a moot point as to whether syntactic variation can be quantified by traditional sociolinguistic methods of analysis. In particular, it is questionable whether one of the central elements in measuring sociolinguistic variation—the linguistic variable—can be transferred from phonetic to syntactic variation. For discussion of this point see Sankoff (1973), who argues that it can, and Lavandera (1978), who claims that, in its original form at least, it cannot. See also Romaine (1982). 3 Broadening the social base of the survey could yield some very interesting results. Informants frequently commented that some usages of tener+past participle were found in rural areas or among less educated speakers. An investigation of these groups would almost certainly produce greater variation in the results. These impressions coincide with the view of Criado de Val (1948, 118–20). 4 For a detailed description of the various interpretations of tener+past participle see Roca Pons (1958). Although his descriptions are detailed and accurate, not all the distinctions he makes are of equal importance in the development of the construction. 5 Most people would wish to argue that in examples such as these, tener is a full verb and does not form a unitary construction with the participle, which should be interpreted adjectivally. The examples are included here for the sake of completeness, and because I shall refer to them in my discussion of the grammaticalisation of tener. 6 Note that where the example expresses inalienable possession there is a purely imperfective state even where the verb is punctual, as in example (2). Inalienable possession seems to dominate all other interpretations and prevents any more grammaticalised interpretation. 7 Although the informant’s intuitions may seem rather vague, they are borne out more conclusively by results from another speaker which will be discussed in section 2.6. 8 To some extent at least the notion of repeated action is carried by tener+ past participle itself and does not rely on the addition of iterative adverbial phrases. For another informant the sentence Se tienen dichas las mismas reclamaciones muchas veces was tautologous: muchas veces was superfluous, as in this case the use of tener+past participle already indicated repetition. 9 Ramat (1987, 9–10) suggests that the first step towards auxiliarisation of periphrastic verbal forms may occur in cases of inalienable possession. The results of my survey suggest that cases of inalienable possession are indeed among the most widely accepted examples of tener+past participle, and it may well be here that tener is first regularly found with a participle. However, the results also indicate that such cases are clearly not the first to receive an auxiliary interpretation: the inalienable possession interpretation is dominant and the agent of the participle may well not be identical with the subject of tener. 10 Although he does not go into detail, Criado de Val (1948, 120) implies that some examples with particular verbs filter through into standard educated speech before others. He specifically mentions tener dicho. Alarcos Llorach (1947, 126) mentions decir and ver, and suggests that they are especially common in the ‘lenguaje popular’. 11 Various writers have commented on this. See Pottier (1964b, 63) and Barrera-Vidal (1972, 141).
Notes
160
12 Barrera-Vidal includes other quantitative adverbs (tanto, poco) in his discussion. Although these were not included in my survey, I have no reason to believe that they behave differently from mucho. There may be a reluctance to accept examples with poco, as tener+past participle tends to emphasise the frequency of an action rather the rarity of it. 13 Roca Pons (1958, 138–9 and 175–7) also comments on this point, and points out the differences between previous use of haber and tener with these verbs and the situation in the modern language. 14 Some approaches, such as that of the generativists for example, claim to deal with the competence or linguistic knowledge of an individual speaker. Such approaches would not aim to incorporate or describe variation between speakers. However, given the nature of the variation—i.e. that it can be hierarchically ordered—and the fact that different speakers’ use of a construction is based on the same principles and is clearly mutually intelligible, it would be an advantage if cross-speaker variants could be accounted for within a single framework. 15 The question of selectional restrictions can also be linked to semantic attrition. For treatment of this see Launay (1980). Launay’s account is not entirely satisfactory. According to Launay’s arguments any early stages of semantic loss occurring before loss of selectional restrictions would lead merely to metaphorical usage, while semantic loss occurring after loss of selectional restrictions would be a part of grammaticalisation. This view fails to emphasise the fact that this early semantic loss is the beginning of grammaticalisation, even if at this stage one might not want to refer to the verb in question as an auxiliary. It also destroys the unit of semantic loss by dividing it into two separate categories. However, in spite of its inadequacies, Launay’s account shows how loss of selectional restrictions is part of the more general process of semantic attrition. 16 In his analysis of the English go+-ing construction, Bolinger (1983, 164) reaches a similar conclusion:
By spreading the action over two verbs, the language captures nuances of aspect and modality that are not expressed with the principal verb alone…. But the semantic effect of these combinations is more general. They are the matrix from which modal and aspectual auxiliaries have developed and continue to develop.
3 Historical development of tener +past participle 1 For a different view, which questions the primacy of the oral language over the written, and regards written texts as an object of linguistic study in their own right, see Romaine (1982, 14–21). Rissanen (1986) also discusses the problems encountered and the methods needed in obtaining syntactic information from texts. 2 As the complete texts of Mil and STP are just under 25,000 words, these samples are slightly smaller. 3 The exact quotation is ‘ya tengo relucida…’ but this is one of Sancho’s many malapropisms, which is corrected by don Quijote. 4 Pountain (1985, 345) points out that tener+past participle sporadically shows non-agreement of the participle with the direct object in medieval Spanish, but overall there is a clear tendency for agreement to occur throughout this period. 5 For the purposes of counting and commenting on word order, only the first participle in multiparticiple constructions is considered; other participles must, inevitably, be separated from the finite verb. This is discussed more fully below.
Notes
161
6 Examples where había+past participle expresses possession have been excluded. Figures for later texts are not included as by this stage había+past participle has taken over completely from the -ra form as the pluperfect (although there is one vestigial -ra pluperfect in STP). 7 Once again, figures for later texts are not included as haber+past participle completely ousts ser+past participle. 8 Only examples where agreement can clearly be tested have been included in the figures: examples with a masculine singular direct object, a non-nominal direct object, or no direct object at all have been excluded. 9 Dik (1987) discusses the link between copulas and auxiliaries, claiming that the two are more similar than has often been thought. In particular he argues that the development of a copula into an auxiliary does not involve desemanticization (=semantic loss or erosion), as a copula does not have any independent meaning: the development is rather a case of expansion or ‘the acquisition of new grammatical functions by some grammatical element’ (p. 80). He also concludes that the change from copula to auxiliary ‘originates in innovative aspectual forms, which may later be reinterpreted in terms of Tense or Voice’ (p. 80). 10 Bolinger (1983) puts forward a similar argument in his analysis of the English go+-ing construction, in which he stresses the gradual nature of grammaticalisation. He concludes (pp. 166–7) that ‘Auxiliaries in English are not a class but a formal semantic AREA. It is not enough to argue…that “auxiliaries are main verbs”; main verbs are also auxiliaries. There is a probably a universal tendency for any two verbs juxtaposed and in tight construction, especially where there is no change of subject, to amalgamate.’ 11 A full survey of the development of haber+past participle is outside the scope of this study. For more detailed descriptions of its history see Lucot (1940) and Bassols (1951) for developments in Latin, and Alarcos Llorach (1947) for later developments.
4 Historical development in Portuguese 1 This example is questionable. The phrase is mandou deçepar huum homeem dos da uilla que tinha preso (RDJ 53) (‘he ordered a man to be mutilated from those of the town who he held prisoner’). The relative clause appears to apply to dos da uilla (‘from those of the town’), in which case the participle does not agree with the object. However, it could possibly be construed with huum homeem (‘a man’), in which case there is no such lack of agreement. 2 Entrado is not included in this list as it is used intransitively with haver (auia entrado na uill de Guimaraaes (RDJ 26) (‘he had entered the town of Guimaraaes’)), and transitively with ter (e diserom-lhe como os seus combaterom a uilla e que [a] tinham ja emtrada per força (RDJ 25) (‘and they told him how they had taken the town by force’)). 3 VME and CPA are not included in Tables 4.6, 4.7, and 4.8 as they contain no examples of haver+past participle. 4 Paiva Boleo (1936) also finds various instances of a past punctual use of ter/ haver+past participle which he tries to account for as exceptions caused by factors such as foreign influence. 5 Irmen disagrees with the traditional term for ter+past participle—the pretérito perfeito composto—preferring the term pretérito imperfeito actual. But his opinion on the usage of the construction broadly coincides with that of the other authors cited. 6 The examples given by Willis (1965, 211) suggest that the participle may be placed before or after an object with which it agrees. 7 Guiter does not indicate whether he includes all examples of ter+past participle, or whether the more lexical examples are discounted.
Notes
162
8 The replacement of haver by ter in Portuguese is discounted for the moment: HABEO FACTUM is used to indicate any compound present perfect. 9 Alarcos Llorach (1947) puts forward a similar chronology for Spanish haber +past participle. However, he includes a phase between Harris’s stages 2 and 3 in which haber+past participle expresses a punctual action occurring immediately before the present moment.
5 Comparison with other Romance languages 1 My study is of the Spanish spoken by informants from Oviedo, not of Asturian. It is arguable whether the speech of at least some of the informants is really a continuation of the bable or local dialect, which has become increasingly Castilianised, or whether they speak Spanish influenced by bable. This is probably an inappropriate question, to which there is no satisfactory answer. Over twenty-five years ago Grossi (1962) was commenting on the difficulty of finding native speakers of bable. He talks (p. 446) of ‘la extraordinaria castellanización que en los últimos tiempos…ha sufrido esta zona, y que hace que sea muy dificil encontrar a alguien que de una forma completamente natural y espontánea hable el dialecto’. He goes on to say that even elderly people ‘hablan una lengua híbrida, mezcla de castellano y bable, que mejor podríamos llamar “castellano contaminado por el bable”…y no a la inversa, ya que el fondo de la lengua y muchas tendencias, son típicamente castellanas’. (For a general discussion of the impossibility of making clear-cut distinctions between different varieties of a language, see Hudson (1980, chapter 2). For recent (although neither systematic nor unbiased) comment on the relation between standard Spanish and other languages or dialects in Spain, see the essays in Salvador (1987).) The point I wish to make here is that all the informants distinguished between Spanish and bable, and considered themselves to be Spanish speakers and not speakers of bable. However, differences in use of tener+ past participle between these informants and speakers of standard Spanish may well be due to the influence of the verbal system of Asturian on the speech of monolingual Spanish speakers in the region. Weinreich (1953, 11) points out that the speech of monolingual speakers may be influenced by developments originally introduced by bilingual speakers. 2 There could be an interesting parallel here with the use of the temps sobrecompausats (double compound tenses) in Occitan. According to Wheeler (1988, 265), one of the two main uses of this form is to express ‘present relevance of an event remote in time, in contrast to the present perfect, which expresses present relevance of a recently past event/state.’ Wheeler gives as examples Ai agut vist ‘Le Dernier Tango à Paris’ (‘I have seen “Last Tango in Paris” [i.e. already/once upon a time/several times in the past]’, and I siatz aguda estada en Arle, vos? (‘Have you ever been to Arles?’). These examples are very similar to some examples produced by the Oviedo informants, such as (7) and (8). This use is not apparently shared by the standard French temps surcomposés, which usually stresses the completion of an action, or the fact that something was the case but is no longer so (see Harris (1978, 155– 6; 1988, 229–30)). In his study of standard and nonstandard varieties of French, Cornu (1953) does not specifically mention the use noted by Wheeler, although some of his examples are compatible with such an interpretation. Unfortunately, lack of context in Cornu’s examples makes it difficult to decide whether this use is found in some regional varieties of French, and if it is, just how it is used. 3 Lack of context for the examples from Rohlfs and Seifert sometimes makes it difficult to ascertain the precise interpretation of a particular example; this is especially true of Seifert’s examples, as she does not supply glosses. When providing translations I have taken into account the comments made by Rohlfs and Seifert about each example.
Notes
163
4 It does not directly concern us here, but Rohlfs indicates that in Piedmontese tenere+past participle is also found, although not with a perfect function: it is used as a present durative construction, similar to standard Italian andare +present participle, e.g. u ten dic (compare standard Italian egli va dicendo); cosa tenstu fait? (standard Italian cosa vai facendo?). 5 I am grateful to Martin Maiden for help with the interpretation of examples (35)–(40). 6 This is similar to the situation in Portuguese before the split between agreeing and invariant constructions developed. At an earlier stage the agreeing construction came to be used in past action examples, then the participle ceased to agree in such examples, leading to the distinction between the invariant and agreeing constructions. In Spanish the agreeing construction may at times refer to past iterative action, but no split has occurred between invariant and agreeing constructions. 7 It is interesting to note that this type of development is apparently not limited to Romance languages. The English verb keep shares some of the characteristics of the Romance development we have been considering here. Latin TENERE usually meant to keep something/someone in a particular condition. This notion of keeping something is still found in the early examples we have looked at from Spanish. English keep occurs with the participles of transitive verbs which are similar in meaning to those found in the early Spanish examples, and usually with a meaning similar to that attributed to the early examples of tener+past participle; for example, He keeps it hidden, She keeps it a closelyguarded secret. Clearly English keep+past participle has not developed nearly so far as Spanish tener+past participle; there are no equivalents with keep of the Spanish tener previsto, tener entendido, etc. But it would seem that such a development is not simply a Romance phenomenon. It may be that verbs which express the notion of keeping or maintaining something have a propensity to develop a resultant state usage. These parallels can be taken a little further. In some Romance languages the reflex of TENERE+past participle has come to indicate repeated or durative action. English keep forms part of a construction with a similar meaning. But English uses keep (on) with the present participle rather than the past participle, e.g. He keeps (on) boasting about his success; They keep fussing about such minor details. This present durative construction is apparently similar in usage to the Piedmontese tenere+past participle construction (see note 4). It would be interesting to know whether such similarities are simply coincidental or whether there is an affinity between the expression of resultant and that of repeated or durative action. This possible link between verbs of keeping/maintaining, resultant state constructions and iterative/durative constructions looks at first sight to be an interesting one which merits further examination. 8 For the notion of the ‘extended present’ see section 0.3.1.5. 9 On the relation between action and resultant state as it applies to haber and tener, see Roca Pons (1958, 43–4). 10 The sentences marked with an asterisk in (49)–(50) are not in themselves ungrammatical, but they do not relate to their supposed counterparts. 11 Example (53) is from Roca Pons (1968), and example (54) appeared in El País: panorama semanal 29 February 1988. 12 Roca Pons (1968, 396–7) argues that (53) is still primarily a resultant state construction. His analysis of it seems to distort its interpretation in order to make it fit into his description of estar+past participle as purely stative. But he admits that this example is an indication of ‘la transición entre la expresión del estado y de la acción’, indicating that there are times when estar +past participle does not have a purely resultant state function. 13 This does not mean that the change in Catalan is due to Spanish influence. Badia suggests this influence as one possible reason for the development, but also suggests that it could be an internal development, similar to the development undergone in Spanish, but separate from it (1964, 68): ‘no hem de veure, en aquesta extensió desmesurada d’estar, un castellanisme
Notes
164
sistemàtic, sinó que molt sovint es tracta de les darreres etapes d’una evolució que el castellà realitzà en poc de temps, però que en català ha estat lenta i laboriosa’. 14 For example, Gili Gaya (1955) does not mention cases where the time of the action is referred to—as in (53)—although the implication is that such cases are inadmissible. Neither does he mention cases where the agent is included, as in (54). (In a later edition (1961, 125) he does, however, state that estar+ past participle does not appear in compound forms with haber, as in ha estado provocada, in (54).)
6 Conclusions 1 This argument has been couched in terms of verb categories—transitive versus intransitive, and different classes of transitive verb. It can also be applied, for example, to the order in which different object types (noun, que clause, infinitive, etc.) are likely to appear in the construction, or to participial agreement, or word order, or any of the factors which are involved in the development of tener+past participle and which affect current judgements on acceptability of examples. 2 The ‘language-learner’ hypothesis is, of course, not incompatible with the notion of hierarchical ordering. I am merely arguing that even if we accept such a hypothesis (and Carden (1973) has suggested that it is not always a convincing approach) we still have to posit a hierarchy of uses.
References Abuín Soto, M., 1971, El castellano hablado en las Rías Bajas Gallegas, Archivum , 21, pp. 171– 206. Alarcos Llorach, E., 1947, Perfecto simple y compuesto en español, Revista de filología española , 31, pp. 108–39. Alcina Franch, J., and J.M.Blecua, 1975, Gramática española , Barcelona. Alonso, A., and P.Henríquez Ureña, 1945, Gramática española: segundo curso , 5th edn, Buenos Aires. Alonso Montero, X., 1966, Ensayo de bilingüística: galego e castelán frente a frente, Grial , 13, pp. 314–33. Badia Margarit, A.M., 1962a, Gramática catalana , vol. 1, Madrid. ——1962b, Gramática catalana , vol. 2, Madrid. ——1964, Llengua i cultura als països catalans , Barcelona. Barrera-Vidal, A., 1972, Parfait simple et parfait compose en castillan moderne , Munich. Bassols de Climent, M., 1951, La cualidad de la acción verbal en español, in Estudios dedicados a Menéndez Pidal , vol. 2, pp. 135–47. Bello, A., and R.J.Cuervo, 1898, Gramática de la lengua castellana , 6th edn, Paris. Bennett, P.A., 1981, Is syntactic change gradual?, Glossa , 15, pp. 115–35. Benveniste, E., 1965, Structures des relations d’auxiliarité, Acta linguistica hafniensia , 9, 1, pp. 1– 15. Blansitt, E.L., 1963, The verb phrase in Spanish: classes and relations, University of Texas dissertation. Bolinger, D.L., 1961, Syntactic blends and other matters, Language , 37, pp. 366–81. ——1968, Judgements of grammaticality, Lingua , 21, pp. 34–40. ——1970, Modes of modality in Spanish and English, Romance Philology , 23, pp. 572–80. ——1983, The Go-progressive and auxiliary-formation, in F.Agard, G.Kelley, A.Makkai and V.Makkai (eds), Essays in Honor of Charles F.Hockett , Leiden, pp. 153–67. Bynon, T., 1977, Historical Linguistics , Cambridge. Canellada, M.J., 1944, El bable de Cabranes , Madrid. Cano Aguilar, R., 1981, Estructuras sintácticas transitivas en el español actual , Madrid. Carden, G., 1973, Dialect variation and abstract syntax, in R.W.Shuy (ed.), Some New Directions in Linguistics , Washington D.C., pp. 1–34. ——1976, Syntactic and semantic data: replication results, Language in Society , 5, pp. 99–104. Carratalá, E., 1980, Morfosintaxis del castellano actual , Barcelona. Castilho, A.T. de, 1967, Introdução ao estudo do aspecto verbal na língua portuguêsa, Alfa , 12, pp. 7–135. Catalán, D., 1974, Lingüística Ibero-Románica: crítica retrospectiva , Madrid. Chomsky, A.N., 1965, Aspects of the Theory of Syntax , Cambridge, Mass. Cohen, M., 1960, Quelques considerations sur le phénomène des verbes auxiliaires, Studii si cercetari lingvistice , 11, pp. 433–42. Coleman, L., 1973, Why the only interesting syntactic dialects are the uninteresting ones, in C.Corum, T.C.Smith-Stark and A.Weiser (eds), Papers from the Ninth Regional Meeting Chicago Linguistic Society , Chicago, pp. 78–88. Comrie, B., 1976, Aspect , Cambridge. ——1985, Tense , Cambridge.
References
166
Concepción Company, C., 1983, Sintaxis y valores de los tiempos compuestos en el español medieval, Nueva revista de filología hispánica , 32, pp. 235–57. Cornu, M., 1953, Les Formes surcomposées en français , Bern. Corominas, J., and J.A.Pascual, 1980, Diccionario crítico etimológico castellano e hispánico , Madrid. Coseriu, E., 198., Aspect verbal ou aspects verbaux? Quelques questions de théorie et de méthode, in J.David and R.Martin (eds), La Notion d’aspect , Metz, pp. 13–25. Criado de Val, M., 1948, Sintaxis del verbo español moderno , vol. 1, (Anejo de la Revista de Filología Española, 41), Madrid. ——1958, Gramática española , Madrid. Damourette, J., and E.Pichon, 1911, Des Mots a la pensée: essai de grammaire de la langue française , vol. 1, Paris. Dias da Costa, A., 1976, Periphrastic verbal expressions in Portuguese, in J. Schmidt-Radefeldt (ed.), Readings in Portuguese Linguistics , Amsterdam, pp. 187–243. Dik, S., 1987, Copula auxiliarization: how and why?, in M.B.Harris and P. Ramat (eds), Historical Development of Auxiliaries (Trends in Linguistics: Studies and Monographs, 35), Berlin, pp. 53–84. Elliott, D., S.Legum, and S.A.Thompson, 1969, Syntactic variation as linguistic data, in R.I.Binnick, A.Davison, G.M.Green and J.L.Morgan (eds), Papers from the Fifth Regional Meeting, Chicago Linguistic Society , Chicago, pp. 52–9. Esgueva, M., and M.Cantarero (eds), 1981, El habla de la ciudad de Madrid: materiales para su estudio , Madrid. Fente, R., J.Fernández, and L.G.Feijóo, 1972, Perífrasis verbales , Madrid. Fillmore, C.J., 1971, Some problems for case grammar, in R.J.O’Brien (ed.), Report of the TwentySecond Annual Round Table Meeting on Linguistics and Language Studies , Washington D.C., pp. 35–56. Fleischman, S., 1983, From pragmatics to grammar, Lingua , 60, pp. 183–213. Fontanella de Weinberg, M.B., 1970, Los auxiliares españoles, Anales del Instituto de Lingüística de la Universidad de Cuyo , 10, pp. 61–73. García Alvarez, M.T.C., 1960, Morfología verbal en el bable de Bimenes, Archivum , 10, pp. 405– 24. García de Diego, V., 1916, Dialectalismos, Revista de filología española , 3, pp. 301–18. ——1950, El castellano como complejo dialectal y sus dialectos internos, Revista de filología española , 34, 107–24. Gili, J. 1943, Introductory Catalan Grammar , Oxford. Gili Gaya, S., 1955, Curso superior de sintaxis española , 5th edn, Barcelona. Gougenheim, G., 1929, Étude sur les périphrases verbales de la langue française , Paris. Green, J.N., 1982, The Status of the Romance Auxiliaries of Voice, in N. Vincent and M.B.Harris (eds), Studies in the Romance Verb , London, pp. 97–138. ——1987, The Evolution of Romance Auxiliaries: Criteria and Chronology, in M.B.Harris and P.Ramat (eds), Historical Development of Auxiliaries (Trends in Linguistics; Studies and Monographs, 35), Berlin, pp. 257–67. Grevisse, M, 1946, Le Bon Usage , 3rd edn, Gembloux. Grossi, R., 1962, Breve estudio de un bable central: el de Meres, Archivum , 12, pp. 445–65. Guiter, H., 1980, Quelques particularités du verbe portugais, Arquivos do centro cultural português , 15, pp. 731–45. Hadlich, R.L., 1971, A Transformational Grammar of Spanish , Englewood Cliffs, N.J. Hamplová, S., 1968, Acerca de la manera de acción y el problema de su expresión mediante las perífrasis verbales en español, Philologica Pragensa , 11, pp. 209–31. Hanssen, F., 1913, Gramática histórica de la lengua castellana , Halle. Harris, M.B., 1978, The Evolution of French Syntax: a Comparative Approach , London.
References
167
——1982, The ‘past simple’ and the ‘present perfect’ in Romance, in N.Vincent and M.B.Harris (eds), Studies in the Romance Verb , London, pp. 42–70. ——1988, French, in M.B.Harris and N.Vincent (eds), The Romance Languages , pp. 209–45. Harris, M.B., and P.Ramat (eds), 1987, Historical Development of Auxiliaries , (Trends in Linguistics: Studies and Monographs, 35), Berlin. Hernández Alonso, C., 1984, Gramática funcional del español , Madrid. Hernanz Carbó, M.L., 1980, Las perífrasis verbales de infinitivo en español: hacia una posible solución transformacional, Revista española de lingüística , 10, pp. 411–43. ——1982, Elinfinitivo en español , Barcelona. Holt, J., 1943, Etudes d’aspect, Acta Jutlandica , 15, pp. 1–94. Huddleston, R.D., 1974, Further remarks on the analysis of auxiliaries as main verbs, Foundations of Language , 11, pp. 215–29. ——1976a, An Introduction to English Transformational Syntax , London. ——1976b, Some theoretical issues in the description of the English verb, Lingua , 40, pp. 331–83. ——1980, On Palmer’s defence of the distinction between auxiliaries and main verbs, Lingua , 50, pp. 101–15. ——1984, Introduction to the Grammar of English , Cambridge. Hudson, R.A., 1980, Sociolinguistics , Cambridge. Imbs, P., 1960, L’Emploi des temps verbaux en français moderne: essai de grammaire descriptive , Paris. Irmen, F., 1966, O pretérito composto em Português, Revista de Portugal , Serie A, 31, pp. 222–38. Joos, M., 1964, The English Verb , Madison. Keniston, H., 1937, Spanish Syntax List , New York. Klein, P.W., 1968, Modal Auxiliaries in Spanish , Seattle. ——1982, On defining auxiliary verbs in Spanish, Hispanic Journal , 4, pp. 113–28. Labov, W., 1972a, Some principles of linguistic methodology, Language in Society , 1, pp. 97–120. ——1972b, Sociolinguistic Patterns , Philadelphia. Lapesa, R., 1950, Historia de la lengua española , 2nd edn, Madrid. Lass, R., 1976, Variation studies and historical linguistics, Language in Society , 5, pp. 219–29. Launay, M., 1980, Acerca de los auxiliares y frases verbales, Lingüística española actual , 2, pp. 39–79. Lavandera, B.R., 1978, Where does the sociolinguistic variable stop?, Language in Society , 7, pp. 171–82. Lehmann, C, 1985, Grammaticalization: synchronic variation and diachronic change, Lingua e stile , 20, pp. 303–18. Lenz, R., 1925, La oración y sus partes , 2nd edn, Madrid. Lope Blanch, J.M., 1953, Observaciones sobre la sintaxis del español hablado en Mexico , Mexico D.F. ——1962, Sobre la oración gramatical: en torno al curso de sintaxis de Gili Gaya, Nueva revista de filología hispánica , 16, pp. 416–22. ——, 1972, Sobre el uso del pretérito en el español de Mexico, in Estudios sobre el español de Mexico , Mexico D.F., pp. 127–39. Lucot, R., 1940, Remarque sur l’emploi de habeo avec le participe en -to- , in Mélanges de philologie, de littérature et d’histoire anciennes offerts a Alfred Ernout , Paris, pp. 247–9. Lyons, J., 1968, Introduction to Theoretical Linguistics , Cambridge. ——1977, Semantics , Cambridge. McCoard, R.W., 1978, The English Perfect: Tense-choice and Pragmatic Inferences , (North Holland Linguistic Series, 38), Amsterdam. Macpherson, I.R., 1967, Past participle agreement in Old Spanish: transitive verbs, Bulletin of Hispanic Studies , 44, pp. 241–54. Martínez Alvarez, J., 1973, Las formas compuestas en el verbo del bable central, Archivum , 23, pp. 299–307.
References
168
Mattos da Silva, R.V., 1981, Um aspecto do auxiliar no português arcaico, Tulane Studies in Romance Linguistics and Literature , 10, pp. 93–109. Mattoso Câmara, J., 1972, The Portuguese Language , (trans. A.J.Naro), Chicago. Menéndez Pidal, R., 1906, El dialecto leonés, Revista de archivos, bibliotecas y museos , 10, pp. 128–72 and 294–311. Millán Urdiales, J., 1966, El habla de Villacidayo (León) , (Anejo del Boletín de la Real Academia Española, 13), Madrid. Moll, F.de B., 1952, Gramática histórica catalana , Madrid. Narbona Jiménez, A., 1981, ¿Verbos modales en español?, Verba , 8, pp. 171–86. Naro, A.J., and M.Lemle, 1977, Syntactic diffusion, Ciência e cultura , 29, pp. 259–68. Navas Ruiz, R., 1963, Ser y estar: estudio sobre el sistema atributivo del español, Acta salmanticensia: Universidad de Salamanca: filosofía y letras , 17, no. 3. Paiva Boleo, M. de, 1936, O perfeito e o pretérito em português em confronto com as outras línguas românicas , Coimbra. Palmer, F.R., 1965, A Linguistic Study of the English Verb , London. ——1974, The English Verb , London. ——1979a, Why auxiliaries are not main verbs, Lingua , 47, pp. 1–25. ——1979b, Modality and the English Modals , London. Peral Ribeiro, J.A., 1958, *ESSERE, SEDERE e STARE nas línguas românicas, Boletim de Filologia , 17, pp. 147–76. Pinkster, H., 1987, The strategy and chronology of the development of future and perfect tense auxiliaries in Latin, in M.B.Harris and P.Ramat (eds), Historical Development of Auxiliaries (Trends in Linguistics: Studies and Monographs, 35), Berlin, pp. 193–223. Pontes, E., 1973, Verbos auxiliares em português , Rio de Janeiro. Porto Dapena, X.A., 1972, Encol da estructura do verbo galego, Grial , 35, pp. 13–28. Pottier, B., 1961, Sobre el concepto de verbo auxiliar, Nueva revista de filología hispánica , 15, pp. 325–31. ——1964a, Introduction a l’étude des structures grammaticales fondamentales , 2nd edn, Nancy. ——1964b, Morphosyntaxe espagnole: étude structurale , 3rd edn, Paris. Pountain, C.J., 1982, *ESSERE/STARE as a romance phenomenon, in N.Vincent and M.B.Harris (eds), Studies in the Romance Verb , London, pp. 139–60. ——1983, Structures and Transformations: the Romance Verb , London. ——1985, Copulas, verbs of possession and auxiliaries in Old Spanish: the evidence for structurally interdependent changes, Bulletin of Hispanic Studies , 62, pp. 337–55. Pullum, G., and D.Wilson, 1977, Autonomous syntax and the analysis of auxiliaries, Language , 53, pp. 741–88. Quirk, R., S.Greenbaum, G.Leech and J.Svartvik, 1985, A Comprehensive Grammar of the English Language , London. Rallides, C., 1971, The Tense Aspect System of the Spanish Verb as Used in Cultivated Bogotá Spanish , The Hague. Ramat, P., 1987, Introductory paper, in M.B.Harris and P.Ramat (eds), Historical Development of Auxiliaries (Trends in Linguistics: Studies and Monographs, 35), Berlin, pp. 3–19. Ramsey, M.M., 1956, A Textbook of Modern Spanish (rev. R.K.Spaulding), New York. Real Academia Española, 1973, Esbozo de una nueva gramática de la lengua española , Madrid. Rissanen, M., 1986, Variation and the study of English historical syntax, in D. Sankoff (ed.), Diversity and Diachrony , (Current Issues in Linguistic Theory, 53), Amsterdam, pp. 97–109. Roca Pons, J., 1957–8, Verbs auxiliars afins a estar en català antic, Estudis romànics , 6, pp. 165–8. ——1958, Estudios sobre perífrasis verbales del español , (Anejo de la Revista de Filología Española, 67), Madrid. ——1964, Llengua i cultura als països catalans , Barcelona. ——1968, El aspecto verbal en español, Linguistica antverpiensia , 2, pp. 385–99. ——1971, Introducción a la gramática , vol. 2, Barcelona.
References
169
——1979, Sobre los verbos auxiliares en las lenguas románicas, in A.Várvaro (ed.), XIV Congresso Internazionale de Linguistica e Filologia Romanza: Atti , vol. 3, Naples, pp. 313–23. Rodríguez-Castellano, L., 1952, La variedad dialectal del alto Aller , Oviedo. ——1954, Aspectos del bable occidental , Oviedo. Rohlfs, G., 1969, Grammatica storica della lingua italiana e dei suoi dialetti: sintassi e formazione delle parole (trans. T.Franceschi and M.Caciagli Fancelli), Turin. Rojo, G., 1974, Perífrasis verbales en el gallego actual , Santiago de Compostela. Romaine, S., 1982, Socio-historical Linguistics , Cambridge. Ross, J.R., 1969, Auxiliaries as main verbs, Studies in Philosophical Linguistics , 1, pp. 77–102. Salvador, G., 1987, Lengua española y lenguas de España , Barcelona. Salvi, G., 1987, Syntactic restructuring in the evolution of Romance auxiliaries, in M.B.Harris and P.Ramat (eds), Historical Development of Auxiliaries (Trends in Linguistics: Studies and Monographs, 35), Berlin, pp. 225–36. Sankoff, G., 1973, Above and beyond phonology in variable rules, in C-J.N. Bailey and R.W.Shuy (eds), New Ways of Analyzing Variation in English , Washington D.C., pp. 44–61. Santamarina, A., 1974, El verbo gallego , Santiago de Compostela. Schogt, H.C., 1968, Les Auxiliaires en français, La Linguistique , part 2, pp. 5–19. ——1970, L’Auxiliaire, Actes du X e Congrès International des Linguistes, Bucharest, 28 August–2 September 1967 , vol. 4, pp. 345–52. Schroten, J., 1974–5, En torno a los verbos perifrásticos del español: un análisis sintáctico transformacional, Revista de filología española , 57, pp. 35–63. Seco, R., 1930, Manual de gramática española , Madrid. Seifert, E., 1930, ‘Haber’ y ‘tener’ como expresiones de la posesión en español, Revista de filología española , 17, pp. 233–76 and 345–89. ——1935, Tenere ‘Haben’ im Romanischen , Florence. ——1957–8, Die Verben Habere und Tenere im Katalanischen, Estudis romànics , 6, pp. 1–74. Spaulding, R.K., 1958, Syntax of the Spanish Verb , Liverpool. Sten, H., 1973, L’Emploi des temps en portugais moderne , Copenhagen. Stevenson, C.H. 1970, The Spanish Language Today , London. Stockwell, R.P., J.D.Bowen, and J.W.Martin, 1965, The Grammatical Structures of English and Spanish , Chicago. Tesnière, L., 1939, Théorie structurale des temps composes, in Mélanges de linguistique offerts a Charles Bally , Geneva, pp. 153–83. Trudgill, P., 1974, The Social Differentiation of English in Norwich , Cambridge. Trudgill, P., and J.K.Chambers, 1980, Dialectology , Cambridge. Vallcorba i Rocosa, J., 1978, Els verbs ésser i estar en català , Barcelona. Vázquez Cuesta, P., and M.A.Mendes da Luz, 1961, Gramática Portuguesa , 2nd edn, Madrid. Vincent, N., 1982, The development of the auxiliaries habere and esse in Romance, in N.Vincent and M.B.Harris (eds), Studies in the Romance Verb , London, pp. 71–96. ——1987, The interaction of periphrasis and inflection: some romance examples, in M.B.Harris and P.Ramat (eds), Historical Development of Auxiliaries (Trends in Linguistics: Studies and Monographs, 35), Berlin, pp. 237–56. Vincent, N., and B.M.Harris, (eds), 1982, Studies in the Romance Verb , London. Weinreich, U., 1953, Languages in Contact , New York. Weinreich, U., W.Labov, and M.I.Herzog, 1968, Empirical foundations for a theory of language change, in W.P.Lehmann and Y.Malkiel (eds), Directions for Historical Linguistics , Austin, pp. 95–188. Wheeler, M.W., 1988, Occitan, in M.B.Harris and N.Vincent (eds), The Romance Languages , London, pp. 246–78. Willis, R.C., 1965, An Essential Course in Modern Portuguese , London. Yates, A., 1975, Teach Yourself Catalan , London. Yllera, A., 1980, Sintaxis histórica del verbo español: las perífrasis medievales , Zaragoza.
Name index Alarcos Llorach, E. 10, 11, 194, 196 Alcina Franch, J. 10, 24 Alonso, A. 25, 83, 169 Badia Margarit, A.M. 160, 173, 174, 199 Barrera-Vidal, A. 71, 72, 194 Bassols de Climent, M. 196 Bello, A. 9, 12, 24, 83 Bennett, P.A. 5, 185, 186 Benveniste, E. 192 Blansitt, E.L. 48 Blecua, J.M. 10, 24 Bolinger, D.L. 42, 43, 193, 195, 196 Canellada, M.J. 155 Cano Aguilar, R. 17, 19 Cantarero, M. 122 Carden, G. 47, 182, 199 Carratalá, E. 191,192 Castilho, A.T. de 146 Chambers, J.K. 184 Chomsky, A.N. 33, 37 Cohen, M. 191, 192 Coleman, L. 181 Comrie, B. 6, 7, 8, 10, 11, 191 Concepción Company, C. 111, 117 Cornu, M. 197 Corominas, J. 100 Coseriu, E. 6 Criado de Val, M. 13, 24, 25, 58, 84, 193, 194 Cuervo, R.J. 9, 12, 24, 83 Damourette, J. 191 Dias da Costa, A. 147, 191, 192 Dik, S. 195 Elliott, D. 181 Esgueva, M. 122 Fente, R. 2, 48, 50, 53, 58, 84, 85 Fillmore, C.J. 69
Name index
171
Fleischman, S. 11 Fontanella de Weinberg, M.B. 3, 27, 29, 30, 31, 32, 33, 34, 38, 41, 48, 67, 80, 192 García Alvarez, M.T.C. 155 Gili, J. 160 Gili Gaya, S. 10, 13, 14, 24, 25, 83, 191, 192, 199 Gougenheim, G. 191 Green, J.N. 44, 191, 192 Grevisse, M. 191 Grossi, R. 155, 197 Guiter, H. 148, 149, 150, 151, 196 Hadlich, R.L. 193 Hamplová, S. 14, 191 Hanssen, F. 83 Harris, M.B. 3, 151, 165, 167, 196, 197 Henríquez Ureña, P. 25, 83, 169 Hernández Alonso, C. 26 Hernanz Carbó, M.L. 13, 15, 41, 42, 44, 45, 191 Holt, J. 191 Huddleston, R.D. 38, 42, 193 Hudson, R.A. 197 Imbs, P. 35 Irmen, F. 141, 142, 143, 144, 145, 146, 196 Joos, M. 192 Keniston, H. 84 Klein, P.W. 3, 13, 14, 15, 27, 30, 33, 34, 35, 38, 39, 40, 41, 42, 192 Labov, W. 47, 95, 96, 127, 184, 190 Lass, R. 44 Launay, M. 3, 13, 14, 15, 16, 17, 26, 27, 29, 30, 31, 33, 34, 35, 38, 41, 194, 195 Lavandera, B.R. 193 Lehmann, C. 4, 65, 86, 87, 89, 90, 91, 92, 94, 123, 126, 127, 135, 185, 188 Lemle, M. 129, 131, 145 Lenz, R. 13, 22, 83 Lope Blanch, J.M. 192 Lucot, R. 196 Lyons, J. 6, 7 McCoard, R.W. 191 Macpherson, I.R. 66, 68, 109 Martínez Alvarez, J. 155 Mattos da Silva, R.V. 129, 136 Mattosa Câmara, J. 146 Mendes da Luz, M.A. 147 Menéndez Pidal, R. 155
Name index
172
Millán Urdiales, J. 156 Moll, F. de B. 160 Narbona Jiménez, A. 43 Naro, A.J. 129, 131, 145 Navas Ruiz, R. 18, 19, 74 Paiva Boleo, M. de 146, 154, 156, 196 Palmer, F.R. 20, 21, 35, 37, 38, 41, 192 Pascual, J.A. 100 Peral Ribeiro, J.A. 174, 175 Pichon, E. 191 Pinkster, H. 124, 125 Pontes, E. 191 Porto Dapena, X.A. 154 Pottier, B. 20, 21, 22, 24, 168, 178, 194 Pountain, C.J. 7, 20, 22, 23, 172, 174, 177, 195 Pullum, G. 38 Quirk, R. 192 Rallides, C. 10 Ramat, P. 3, 44, 194 Ramsey, M.M. 23, 83 Real Academia Española 13, 14, 24 Rissanen, M. 195 Roca Pons, J. 3, 13, 15, 16, 18, 22, 25, 74, 76, 77, 161, 162, 168, 191, 193, 194, 198, 199 Rodríguez-Castellano, L. 155 Rohlfs, G. 162, 163, 164, 197, 198 Rojo, G. 13, 20, 22, 154, 155 Romaine, S. 193, 195 Ross, J.R. 37, 38 Salvador, G. 197 Salvi, G. 193 Sankoff, G. 193 Santamarina, A. 155, 166 Schogt, H.C. 12, 16, 19, 22, 23 Schroten, J. 40 Seco, R. 25 Seifert, E. 3, 96, 100, 119, 120, 121, 125, 156, 160, 161, 162, 163, 164, 165, 197, 198 Spaulding, R.K. 23, 48, 83, 84 Sten, H. 146 Stevenson, C.H. 11 Stockwell, R.P. 39 Tesnière, L. 192 Trudgill, P. 184
Name index
Vallcorba i Rocosa, J. 161, 162, 168, 173, 174 Vázquez Cuesta, P. 147 Vincent, N. 68, 69, 70, 71, 90, 193 Weinreich, U. 180, 181, 182, 197 Wheeler, M.W. 197 Willis, R.C. 140, 172, 196 Wilson, D. 38 Yates, A. 160, 174 Yllera, A. 99, 100, 101, 107, 120, 156
173
Subject index acabar de+infinitive 20 adjectival participles 24, 53, 57, 62 adverbs, frequentative 49, 63 quantitative 61, 63, 71, 72 temporal 49, 171 AGENT-subject 68–9, 71 agreement of past participle; historical development in Romance languages 164–7; see also past participle aller+infinitive (French) 23, 90 appena (Italian) 20 aspect 5–11, 140–6 definition of 6 durative 7–8, 54, 56, 58, 59, 76, 77, 141 imperfective 7, 74, 76–7 iterative 8, 48, 60, 62, 100, 141 lexical 6–7 morphological 6–7 perfective 7, 48 in Portuguese 140–6 punctual 8, 54, 141 Asturian 5, 110, 155–9, 166, 167, 190 compound forms 155–6 attrition, semantic 87, 88, 92, 94 auxiliary status 4, 12–46, 191–3 continuum of 25, 182 criteria 12, 34–6, 90 comparison of 34–6 formal 20–3 functional 19–20 paradigmaticity 23 semantic 13–17 syntactic 26–34, 35 word order 22 definition of 4, 13, 22, 23 position of tener+past participle 23–6 semantic loss as applied to tener+ past participle 17–19 tests for 39, 45–6 auxiliary verbs, classificatory criteria 85–93, 187 definition of 3, 4, 14, 36–42 diachronic justification 44–5
Subject index
175
English 3, 20–1, 37, 39, 41, 44 existence of 36–42 French 12, 19, 22 graduated class membership 85 –93, 187, 188 and main verbs 19, 38 Portuguese 3 selectional restrictions on subject and object 30–4, 89 avere+past participle (Italian) 90, 163 avoir (French) 22, 90 bable 189, 197 note 1 bondedness 87, 88 of tener+past participle 87, 89, 92, 126 Case Grammar 68–9 Catalan 5, 159–62, 165, 167, 173–4, 176, 177, 179 catenatives 20–1 clitic pronoun raising 23–4 coalescence 87 cómo questions 32–3, 78–83 condensation 87, 89 context 48–9, 84, 95 copular function of tener 74–7, 79, 99, 103, 120–3 current state 52, 53, 55, 56, 60, 62, 104 in Portuguese 141, 142 data collection 47, 48 deber 39, 40, 42 desear 77 diachronic change 3, 86, 87, 94, 182–5 diachronic descriptions 1–2, 16, 36, 43–6, 86, 93, 127–8, 188 direct object, with tener+past participle 2, 31, 50, 57, 65–8 nature of 110–11 see also past participle duration 19, 58, 59, 98, 105, 121, 122 durative resultant state 56, 77, 105 English 3, 20-l, 37, 39, 41 entender 56 ésser+past participle (Catalan) 173–4 essere+past participle (Italian) 175 estar+past participle 3, 24, 40, 72, 91, 168–71, 174, 176–8 estar+past participle (Catalan) 161, 173–4 estar+past participle (Portuguese) 172, 177, 178–9 estar+present participle 7, 14, 23 EXPERIENCER-subject 69 extended present 8, 10–11, 157, 158
Subject index
176
finite verb+non-finite verb form sequence 20–1, 34 fixation (diachronic process) 87 French 12, 19, 22, 146, 191 note 1 Galician 5, 20, 154–5, 165–7, 190 ter+past participle 154–5, 166 generative grammar 35, 38, 41, 43 German 37 gerund 3, 21, 26, 150 Golden Age 112, 113, 122 gradual change 185–6 graduated continua 4, 182 grammaticalisation, continuum of 25 and frequency 151 of haber+past participle 25, 45, 68, 113, 124–5, 148 processes 87, 88, 94 of tener+past participle 4–5, 15, 23–5, 56, 73–5, 119–21, 124–5, 148–53 of ter+past participle (Portuguese) 130, 133, 135, 14 8–52 grammaticalisation scale 86, 188 haber, lexical use 114, 123 as main verb 124, 125 haber+past participle 3, 9–11, 16, 23–7, 30, 40, 54, 80, 115–16 with adverbial phrases 91 comparison with tener+past participle 83–4, 104–5, 124–5, 168–70, 176–7 durative value 115 grammaticalisation 25, 45, 68, 113, 125, 148 historical development 113–24 non-agreement 66, 68 -ra pluperfect 115–17, 124, 148 relative frequency of, and tener+past participle 113–14 and ser+past participle 116–17, 168–70 word order 117–20 haber+past participle (Asturian) 155–6, 159 haber tenido 85 HABERE+past participle (Latin) 124, 148, 149, 151–2, 167 haver+past participle (Portuguese) 129, 136–9, 186 agreement of past participle and direct object 138 durative construction 144–5 grammaticalisation 135, 138, 148 historical development 129, 136–8 interpolation 139 havia/tinha+past participle 13 9–40, 150 -ra pluperfect 139–40, 148, 150 relative frequency of, and ter+ past participle (Portuguese) 136 and ser+past participle 140 word order 138, 150 historical development of tener+past participle 1–3, 94–128, 183, 195–6 and modern variation 126–7
Subject index
177
summary 119–23 textual study 95–6 infinitival complement of tener+past participle 51, 56, 64, 126 informants (native speakers) 4, 47–51, 83, 181, 184 analysis of judgements 52–64, 181 interpolation of elements, between haber and past participle 118–19, 123, 150 between haver and past participle (Portuguese) 139 between tener and past participle 112–13, 121, 126 between ter and past participle (Portuguese) 134–5 intransitive verbs, past participle with tener 2, 31, 57, 61, 63, 70–2, 126 past participle with ter (Portuguese) 129, 132 ir+gerund 31 ir a+infinitive 23, 31 Italian 5, 146, 162–4, 165, 174–5, 179 southern dialects 162, 164, 165, 175, 178 language change 14, 43, 50 Latin 44, 90, 100, 124, 148, 149, 151, 164–5, 187 Leonese 156 lexical value of tener 17–19, 74, 97, 98, 103, 104, 105, 106, 119, 121–2 LOCATIVE-subject 68–9, 71 loss of meaning criterion for auxiliary verbs 14–15, 18 as applied to tener+past participle 17–19 mantener 15, 74, 75, 79, 85 mental state 53, 63, 69, 103 modal auxiliaries 22–6, 39–42 modal periphrases 40 muchas veces 58, 61, 62, 71, 72, 91, 170 mucho 61, 71, 72 mucho tiempo 71 native speakers see informants (native speakers) negation 40 NEUTRAL 68–9, 71 north-western Spain 107, 110, 167, 186; see also Asturian, bable, Galician, Leonese, Oviedo dialect obligatorification 87, 89, 91, 92 oír 56, 122 Oviedo dialect 50–1, 156–9, 165, 166, 186, 189 paradigm test 37 paradigmatic variability 87, 88 paradigmaticity 23, 87–92 paradigmaticization 87, 88 passive 29–30, 168–71
Subject index
178
past participle, agreement with direct object, haber 117 tener 2, 50, 51, 57, 65–8, 89, 109– 10, 156–7, 159 ter (Galician) 154–5, 166 ter (Portuguese) 129–30, 132–3, 142, 145–6, 147, 148, 153, 167, 172 non-agreement with direct object, haber 66 , 130 tener 66 –8, 109–10, 126 ter (Galician) 154–5 ter (Portuguese) 142–3, 145, 146, 147, 153, 157–8, 165 postposed 64, 65 preposed 64, 65, 67 past participle+haber 111, 117, 118 past participle+tener 111, 119 past participle+ter (Portuguese) 133 perfect 8–11; in English 9, 11, 191 note 6; in Portuguese 11, 146, 147, 148–9, 152–3; in Spanish 9, 11, 84, 147, 148–9, 152–3 periphrasis 13, 20 periphrastic constructions 24, 26, 36, 91, 146 pluperfect 115–16, 124, 139–40, 148, 150 poder 39, 40, 42 polyvalency 36 Portuguese 3, 129–53, 165, 179 compared with Spanish 4–5, 148- 53 historical development 129–53, 196 see also ter+past participle possession 17, 18, 28, 51, 72–6, 95, alienable 17, 18 inalienable 17, 18, 51, 52, 56, 63, 103 expressed by haber+past participle 114 in historical development of tener+past participle 97, 105 present relevance 9–11, 48, 52 Principle of Convergence (Labov) 96 pronouns 23, 135, 150; see also interpolation qué questions 32–3, 55, 78–83 querer 30, 39, 77 question formation 30, 32, 78–83 questionnaires for informants 47–9, 51 -ra pluperfect in Spanish 115–17, 124, 148 in Portuguese 139–40, 148, 150 reflexive constructions 57–8, 63, 116, 148 regional variation 25, 50–1 resultant state constructions 8, 28, 54–6, 59–62, 70, 75, 83–5, 168–71 historical development 99–100, 105, 120
Subject index
179
Portuguese 141–5, 172 Spanish and Portuguese compared 152–3 Romance languages, constructions similar to tener+past participle 5, 44, 154–79, 197–9 Asturian 155–9 Catalan 159–62, 173–4 comparison of development 164–7 Galician 154–5 Italian 162–4, 174–5 Portuguese 172 STARE+past participle (Latin) 167–78 TENERE and STARE (Latin) 176–8 semantic loss 13–15, 17–19, 35, 88 semantics of tener+past participle 96–107 semi-auxiliaries 16, 99 sentential complement of tener+past participle 51, 57, 77, 126 ser+past participle 3, 24, 116, 124, 148, 168–70 ser+past participle (Portuguese) 140, 148–9, 172 Slav languages 6 sociolinguistics 49, 51, 186, 190 soler 16, 30 stare+past participle (Italian) 174- 5, 179 STARE+past participle (Latin) 5, 167–78 stare+present participle (Italian) 23 static/dynamic dichotomy 44 stative verbs 7, 56, 77, 141–2 strategic gradualness 185 subject of haber, and agent of participle 115 subject of tener, and agent of participle 68–72, 99, 107–9 survey methodology 47–51 data collection 47, 48 direct observation 47 informants 47–51 interviews 47 performance tasks 47, 48 questionnaires 47–9, 51 synchronic descriptions 1–2, 15–16, 36, 43, 45, 86, 93, 188 synchronic variation 3, 4, 5, 86, 94, 126–8, 182–5, 189, 190 syntagmatic variability 87, 89, 92 tener, as main verb 18, 19, 93, 96, 120, 124 as verb of possession 17, 18, 28, 72–6, 124 classification of tener+past participle 24–6, 34, 83–5 tener+adjective 22, 28, 53 tener+noun+participle 57 tener+past participle+infinitive 57, 64 tener+past participle+noun 57, 64 tener+past participle+que 57, 64 tener dado 51, 73, 184 tener dicho 103, 106
Subject index
tener enterdido 122, 184 tener oído 122 tener pensado 66, 77 tener usado 100, 101–2 tenere+past participle (Italian) 163–5, 174–5 TENERE+past participle (Latin) 2, 4, 124, 164–5, 167 Portuguese derivatives 148, 149, 152 and STARE+past participle (Latin) 176–8, 188 tenir+past participle (Catalan) 159–62 ter+past participle (Galician) 154–5, 166–7 ter+past participle (Portuguese) 3–5 +direct object+past participle 134, 135 durative construction 130, 141, 144–5, 146 grammaticalisation 130, 133, 135, 150–2 historical development 129–35, 186, 190, 196 invariant participle 130, 165, 166 iterative construction 130, 141, 143–5, 146 lexical/resultant state 134, 135 modern usage 146–8 position of weak object pronouns 135 relative frequency of, and haver+ past participle 136 word order 4–5, 133–4, 150 ter+past participle+direct object (Portuguese) 134 terminology 5–11 textual surveys 95–6, 182 two-participle constructions 113, 195 note 5 two-place verbs 70 Valladolid dialect 50, 51, 52, 157 variation among speakers 1, 2, 3, 4, 50, 51, 180–5 venir de+infinitive (French) 20, 90 verb paradigms 9 verbal participles 57, 62 word order 3, 22, 51, 64–5 haber+past participle 117–18, 119, 150 haver+past participle (Portuguese) 138, 150 historical development 111–12, 126 ter+past participle (Portuguese) 4–5, 133, 135, 150 written texts 95–6 Zaragoza dialect 50, 51, 52, 157
180