Tense and Mood in English: A Comparison with Danish [Reprint 2013 ed.] 9783110860184, 9783110125818

Citation preview

Tense and Mood in English

Topics in English Linguistics 1 Editors

Jan Svartvik Herman Wekker

Mouton de Gruyter Berlin • New York

Tense and Mood in English A Comparison with Danish

Niels Davidsen-Nielsen

Mouton de Gruyter Berlin • New York

1990

Mouton de Gruyter (formerly Mouton, The Hague) is a Division of Walter de Gruyter & Co., Berlin.

@ Printed on acid-free paper which falls within the guidelines of the ANSI to ensure permanence and durability.

Library of Congress Cataloging in Publication Data Davidsen-Nielsen, Niels, Tense and mood in English : a coraparison with Danish / Niels Davidsen-Nielsen. p. cm. — (Topics in English linguistics ; 1) ISBN 0-89925-718-6 (cloth : acid-free paper) 1. English language—Tense. 2. English language—Mood. 3. English language—Grammar, Comparative—Danish. 4. Danish language—Grammar, Comparative—English. 5. Danish language—Tense. 6. Danish language—Mood. I. Title. II. Series. PE1301.D35 1990 425-dc20 90-13272 CIP

Deutsche Bibliothek Cataloging in Publication Data Davidsen-Nielsen, Niels Tense and mood in English : a comparison with Danish / Niels Davidsen-Nielsen. — BerUn ; New York : Mouton de Gruyter, 1990 (Topics in English linguistics ; 1) ISBN 3-11-012581-1 NE: GT

© Copyright 1990 by Walter de Gruyter & Co., D-1000 Berlin 30. All rights reserved, including those of translation into foreign languages. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any Information storage and retrieval system, without permission in writing from the publisher. Typesetting: Arthur Collignon GmbH, Berlin. — Illustrations: Thomas Lynnerup Jakobsen. — Printing: Gerike GmbH, Berlin. — Binding: Lüderitz & Bauer, Berlin. Printed in Germany.

To Clougha Pike

Acknowledgements

This book has been written intermittently over the last four or five years, interrupted by teaching and, particularly, administrative duties. Düring its preparation I have consulted many colleagues, to whom I would hereby like to offer my sincere thanks. Peter Härder and Torben Vestergaard, who have read the entire manuscript, and Carl Bache and Carl Vikner, who have read all of it except chapter 6, have provided me with extremely helpful and perceptive criticism. Without the liberal assistance of these linguists the task of completing this book would have been harder and the result poorer. I would also like to thank Helge Schwarz and Sten Vikner for careful criticism of substantial sections of the manuscript and Erik Hansen, Per Anker Jensen, Geoffrey Leech, Frank Palmer, and Mick Perkins for constructive comments on parts of it. The present work has greatly benefited from the help of these linguists as well. My thanks are due to several other colleagues and students as well for responding orally — with patience and apparent interest — to specific questions which arose as I was struggling to arrive at a clearer understanding of tense and modality in English and Danish, and which I needed to discuss with somebody then and there. Needless to say none of all these persons can in any way be held responsible for remaining errors and obscurities in the text. I am indebted to the Danish Research Council for the Humanities and to the Otto M0nsted Foundation for generously financing a research visit to the University of Lancaster in the autumn term of 1988 and to the Department of Linguistics and Modern English Language there for providing excellent working conditions in an extremely friendly and stimulating atmosphere. For assistance received in the late phases of the work I would like to thank Lillian Vindelbo, Anne Dorthe Staggemeier and Kirsten Eisborg for typing the manuscript and Alex KHnge for compiling the index and, in so doing, Spotting a number of errors. Finally, my thanks are due to my own department (of English) and the Copenhagen Business School for backing me up in several ways during the writing of this book.

Copenhagen, May 1989

Niels Davidsen-Nielsen

Contents

Acknowledgements Chapter 1 Introduction Chapter2 2.1. 2.2. 2.3. 2.4. 2.5. 2.6. 2.7. 2.8.

Auxiliaries

1 15

Introduction 15 Epistemic and non-epistemic modals 18 Criteria of auxiliarity 21 The semantic criterion 23 The criterion of functional dependency 25 The criterion of permanence of the lexical restrictions of V" 32 The criterion of direct attachment 35 Analysis proposed 37

Chapter 3 Analysis of mood 3.1. The concept of modality 3.2. Mood in English 3.3. Mood in Danish Chapter 4

Analysis of tense

4.1. Time and tense 4.2. Number of tenses and their implementation 4.3. Tense meanings 4.4. Eight or six tenses? 4.5. Other analytical possibilities 4.6. Tense in fiction Chapter 5 5.1. 5.2. 5.3. 5.4. 5.5.

VII

Mood usage

Possibility Necessity Probability Report Imperative

43 43 46 49 53 53 55 59 64 66 69 73 73 83 90 93 98

X

Contents

5.6. 5.7.

Subjunctive Mood in non-fmite verb phrases

Chapterö 6.1. 6.2. 6.3. 6.4. 6.5. 6.6. 6.7. 6.8. 6.9. 6.10. 6.11.

Present time meaning Future time meaning Fast time meaning Anterior time meaning Anterior past time meaning Anterior future time meaning Future of the past time meaning Anterior future of the past time meaning Summary Tense usage in indirect speech Tense in non-fmite clauses

Chapter7 7.1. 7.2. 7.3. 7.4. 7.5. 7.6. 7.7. 7.8. 7.9.

Tense usage

105 III 113 113 117 124 126 133 136 139 141 143 147 152

Modal usage of tenses

159

The present The future The present perfect The future perfect The past The future of the past The past perfect The future perfect of the past Summary

159 161 166 168 170 176 179 182 184

Chapter 8

Semi-auxiliaries and deontic modality

187

8.1. 8.2. 8.3.

Permission Compulsion Obligation

187 194 206

8.4.

Duty

209

Notes

213

Bibliography

215

Index

221

Chapter 1 Introduction

This book compares tense and mood in English with the corresponding categories in Danish. As tense and mood are grammatical categories, it thus gives a contrastive account of those time and modality expressions which are realizedgrammatically in the two languages. Its intended groups of readers are advanced students of English, Danish, and Hnguistics, teachers of English and Danish, and professional linguists. By focussing on grammar it obviously leaves aside a large number of time and modality expressions, namely all the lexical ones. Nevertheless, more ground will be covered than the title would seem to imply. In the first place tense is here interpreted as a broad category with eight members, so the perfect and future constructions — which are in most other studies handled in terms of other verbal categories (aspect or phase and modality respectively) — are included. Furthermore, mood is assumed to comprise not only morphologically signalled constructions like the subjunctive and the imperative but also syntactically signalled constructions with modal auxiliaries such as possibility and necessity (which may be illustrated by respectively You may be right and You must be right). In its account of tense and mood the present book has both an analytical and a descriptive objective. With respect to the former it seeks to contribute to the discussion of how tense and mood can be interpreted in the two languages (chapters 3 and 4) and — as a prerequisite to these analyses — how auxiliaries can be delimited from main verbs (chapter 2). The following chapters on mood usage (5), tense usage (6), modal usage of tenses (7), and semi-auxiliaries and deontic modality (8) are also interspersed with discussions of analytical problems. Predominantly, however, they give a — fairly detailed — description of English and Danish usage. One of the purposes of these chapters is to provide a descriptive basis on which reliable and efficient teaching material can be worked out, whether for Danish students of English or for English-speaking students of Danish. As they stand, however, they are probably too theoretically orientated to be directly useful to any but fairly advanced students. Before they can serve a pedagogic purpose with students on lower levels they will therefore have to be adapted and rewritten in simplified form. The analysis of tense proposed, according to which there are both future and perfect tenses, is decidedly non-mainstream and therefore

2

Introduction

bound to be controversial. However, a number of independent arguments can be adduced which make such an analysis possible and perhaps even preferable (see, for example, Vikner 1985; Davidsen-Nielsen 1985, 1988; Klinge 1988; and Härder forthcoming). In its recognition of analytically realized members of mood, similarly, this bock differs from practically all others on English and Danish grammar (a notable exception being Huddieston 1984: 164 ff.). The position adopted here is that tense and mood are implemented not only morphologically by means of inflections but also syntactically by means of auxiliaries (i. e. grammatical verbs) in the languages investigated. In the case of the perfect tenses, it should be added, the realization is both morphological and syntactical {have + V + -ed in English and have/vcere + Y + -et in Danish). These tenses are thus implemented in the same combined morpho-syntactic way as the progressive aspect in English (ie + V + -ing), the passive voice in English + V + -ed), and one of the two passives in Danish (blive/ vcere + V + -et). From a contrastive point of view and with regard to language teaching and learning there are obvious advantages connected with the broad analysis of tense adopted in this book. For example, a major difference between English and Danish concerns the use of the present perfect versus the past. If the present perfect construction is analysed as present tense 4- perfective aspect (or phase), the difference between this construction and the past involves two categories. Within this analytical framework a contrastive description will have to point out that in a number of situations where have/vcere + Y + -et is used in Danish (e. g. Dickens har skrevet mange romaner) not only another tense but also another aspect (or phase) is used in English {Dickens wrote many novels). And as for language teaching, a Danish Student of English will need to be told that in many cases he cannot transfer a combination of present tense and perfective aspect in L 1 to L 2 but must replace it by a combination of past tense and non-perfective aspect in L 2. An even more important difference between the two languages concerns the future vs. non-future constructions, the former being often required in English where Danish permits and in fact very frequently uses the latter; compare examples like You'll only forget and Du glemmer det bare. In this case too it is clearly more expedient from a contrastive as well as from a pedagogic point of view to describe the difference as pertaining to time and tense than to point out that in many situations where categorical utterances in the present tense are used in Danish, modalized utterances in the present tense (with will of prediction) are used in English.

Introduction

3

The contrastive and practical advantages connected with a broad analysis of tense are reflected, for example, in Johansson and Lysväg's grammar of English for Norwegian students. The authors State (1987, vol. 2: 117) that a system with eight tenses has the advantage of providing "a description which is as comprehensive and reveaüng as possible from a contrastive point of view". It should be emphasized, though, that the analysis of tense proposed in the present study is not motivated by contrastive (and pedagogic) considerations alone but is independently motivated as well. Arguments in favour of it will be put forward in chapter 4. As far as mood is concerned the contrastive and pedagogic advantages connected with adopting an analysis with morphologically as well as syntactically realized members are not nearly as obvious as in the case of tense. Nevertheless, such a broad analysis seems contrastively and didactically expedient in view of the fact that morphologically signalled modal expressions are semantically very similar to expressions signalled syntactically by modal auxiliaries and that there is a certain syntheticanalytic crossover between English and Danish. In English, for example, may, used for the expression of (logical) possibility, and the subjunctive sometimes perform the same communicative function and only differ in style; compare examples like Whatever the reason may be, we cannot tolerate his disloyalty and Whatever be the reason, we cannot tolerate his disloyalty. In examples of this type possibility is in Danish only expressed syntactically by kan or mätte: Hvad gründen end kanj matte vare, kan vi ikke tolerere hans illoyalitet. From the point of view of consistency it is clearly also preferable that tense and mood are analysed in the same way, i. e. that both of them are assumed to be realizable not only morphologically but also syntactically. It should be pointed out, though, that mood is here divided into a synthetic subsystem and an analytic subsystem. The reason for this is that its syntactically realized members are not assumed to have fused with its morphologically realized ones into one organic system. In this way mood is assumed to differ from tense (see chapter 3). The contrastive analysis presented in this book is parallel rather than directional, i. e. within the areas of tense and mood it investigates what structures in English correspond to what structures in Danish and vice versa, rather than what L 1 tense and mood constructions correspond to in L 2. This means that its results are more difficult to apply directly to foreign language teaching than the results of a directional analysis (whether from Danish to English or from English to Danish). On the other hand they are likely to be more useful to theoretically minded

4

Introduction

linguists, for example in their search for universals (see Jorgensen 1982: 58). Furthermore, they have the advantage of being equally suited for the preparation of English language teaching programmes for Danes and of Danish teaching programmes for speakers of English. Similarly, they will be equally useful for the preparation of machine translation systems from Danish to English and from English to Danish. While a directional analysis with, say, Danish as L 1 and English as L 2 is pedagogically more directly useful than a parallel analysis of the two languages for the preparation of material for Danish students of English, it is obviously of rather limited use to those wishing to work out material aimed at Englishspeaking students of Danish. It should be added, however, that though the description put forward here is basically parallel, the summaries which are included at the end of usage chapters or sections have a certain directional bias in tending to focus on transfer from Danish to EngUsh. This is due to the fact that there are more Danish students of English than English-speaking students of Danish and therefore a potentially larger number of readers with an interest in the Danish point of view. To some extent, furthermore, these summaries take the form of "trouble shooting", i. e. they focus on areas of difference where transfer from L 1 to L 2 would result in error. Elsewhere in the usage chapters, however, similarities are described as well as differences. What the book as a whole puts forward, therefore, is a parallel contrastive analysis of tense and mood in English and Danish in which both similarities and differences are accounted for. As regards contrastive analysis as such it will come as no surprise that the contrastive hypothesis — according to which foreign language learners transfer forms and meanings from the source language to the target language productively as well as receptively (see Lado 1957: 2) — is accepted in a book like this one (which would otherwise hardly have been written). We thus disagree with linguists such as Dulay and Burt (1974) and Krashen (1983), who are of the opinion that L 2 learning proceeds in basically the same way as L 1 learning (the L 2 = L 1 acquisition hypothesis), i. e. who basically reject the influence of L 1. As it is beyond the scope of the present study to enter into any detailed discussion of the contrastive hypothesis as such, we shall here restrict Ourselves to giving a brief sketch — largely based on Trosborg (forthcoming) and Härder (1988) — of its history and status. According to the contrastive hypothesis in its original strong version (see Wardhaugh 1970) it is possible to predict what errors will be made by foreign language learners by identifying the differences that exist

Introduction

5

between L 1 and L 2. While similarities in L 1 and L 2 are assumed to lead to positive transfer, differences are assumed to bring about negative transfer, also termed interference. In the late sixties and early seventies the contrastive hypothesis was investigated in a number of studies, synchronic error analyses as well as longitudinal studies. On the basis of these data Dulay and Burt (1974) concluded that negative transfer is a relatively unimportant factor in foreign language learning. Instead, it was assumed that learners follow a universal route in their acquisition of L 2, and that L 2 learning is quite similar to L 1 learning. In the wake of these empirical studies the contrastive hypothesis came under heavy attack in the seventies. Objections were raised against behaviourism as a theory of language learning, against structuralism as an adequate linguistic framework, and against the claim that difficulties and errors can be predicted from a contrastive analysis. On the practical level it was objected that if the majority of errors are not due to interference, contrastive analysis is pedagogically not very useful. According to the contrastive hypothesis in its weak version it is only possible to identify that subset of errors which is caused by interference, i. e. contrastive analysis here has to work hand in hand with error analysis. It is acknowledged that not all errors are due to interference and that the role of L 1 in foreign language learning is therefore less important than originally assumed. However, even if the original version of the contrastive hypothesis was too strong, it is clearly unsatisfactory for contrastive analysis to be stripped of all predictive power. There is not much point in carrying out extensive contrastive investigations only to confirm that errors which are assumed to be due to interference turn out to be so in reality. In recent years linguists have once again begun to look more favourably on the contrastive hypothesis. Since practical experience overwhelmingly supports it, this was bound to happen sooner or later. It has been shown that the absence of some of the errors predicted by the strong contrastive hypothesis can be explained as being due to a number of specific factors (see EUis 1986: 33 — 39). One of these is setting: in a naturalistic setting, e. g. the environment of immigrants learning a second language, there will be less negative transfer than in a classroom setting. Another is the degree of relatedness between L 1 and L 2: in those cases where the languages involved are highly different (e. g. English and Chinese) there will be fewer interference errors than if they are relatively similar (e. g. English and Danish). This is not surprising, for if the learner feels that

6

Introduction

the target language differs so much from his own that more often than not he cannot get away with transfer, he will tend to use this strategy sparingly. Thirdly, the absence of expected errors may be due to the fact that learners, in those cases where they have misgivings about transferring a specific structure or rule, may adopt the strategy of avoiding it. In other words: speakers do not mindlessly reproduce features of their mother tongue on all occasions. But the fact that they have these features at their disposal is of well-attested significance for their L 2 Performance, both with respect to correct and deviant features. Therefore, it will be of great help to learners of a foreign language to know that some L 2 phenomena correspond to L 1 phenomena and that others do not. In the former case this knowledge represents a short cut since a rule which is already firmly established in the learner's mind can be successfully used in the target language (the positive transfer). In the latter case it will serve as a useful warning against errors which the learner is likely to make (traditionally termed "pitfalls"). As argued by Härder, a "contrastive grammar should provide a path into the L 2, turning as much as possible of the L 1 into operational assets for the learner" (1988: 157). Not only similarities should be revealed, however, but also problematic differences, and in many cases "pointing out equivalences and warning against differences will probably go band in hand" (op. cit.: 158). Before leaving the question of contrastive analysis as such we should point out that accepting its relevance and usefulness does not imply that the psychological paradigm — behaviourism — with which it was closely connected in its early stages (the forties and fifties) is accepted as well. In this book there will be no mention of habits getting in the way of foreign language learning, and a term like "interference" will not be used either. On the other hand "transfer", which is not here assumed to refer to habits, will be preferred to a vaguer term like "cross-linguistic influence". In the contrastive description of tense in English and Danish given in this book the means of comparison, i. e. the tertium comparationis, is meaning (see e. g. Coseriu 1970). In chapter 6 on usage we therefore proceed from a temporal meaning to the tense forms used in clauses expressing this meaning in the two languages. For this purpose we operate with three semantic features: [THEN], [PREVIOUS], and [POSTERIOR]. These features will be explained in detail and discussed in chapter 4. At this point it will suffice to State that [-I-THEN] symbolizes past time meaning-, that [ +PREVIOUS] symbolizes anterior time meaning, i. e. time preceding as compared to a point in time which the Speaker selects as his time of orientation and which may differ from the moment of speech;

Introduction

1

and that [ +POSTERIOR] symbolizes future Urne meaning. By way of illustration it may be mentioned that an English sentence like The committee will have flnished its work next Monday has (non-past) anterior future meaning, i. e. its temporal meaning is [ —THEN, +PREVIOUS, + POSTERIOR]. The approach adopted for the comparison of tense usage in chapter 6 can be simply illustrated by two examples: a. Future time meaning, i.e. [ - T H E N , - P R E V I O U S , + POSTERIOR], is typically expressed by clauses in the future tense in English and by clauses in the present tense in Danish, e. g. You'll only forget and Du glemmer det bare. Here the same temporal meaning is expressed by two different tenses, and this constitutes an important difference between the two languages. b. Anterior past time meaning, i. e. [ + THEN, + PREVIOUS, - P O S TERIOR] , is typically expressed by the past perfect tense in both English and Danish, e. g. In 1980 I had lived in Copenhagen for ten years and I 1980 havde jeg boet i Kobenhavn i ti är. In this case a temporal meaning is expressed by the same tense in English and Danish, and this constitutes an important similarity between the two languages. While we have found it possible to propose a relatively simple semantic analysis of time which will serve as a tertium comparationis for a füll description of tense usage in English and Danish, we have found it very difficult to devise a reasonably clear semantic analysis of modality on which a füll description of mood usage in the two languages can be based. By means of the concepts "possible", "necessary", "epistemic" ( = what is concerned with the speaker's knowledge and his commitment to what he says), and "deontic" ( = what is concerned with the permissibihty and obligatoriness of actions and thus involves "morally responsible agents", see Lyons 1977: 823) four modal meanings can be differentiated. On the basis of such a semantic system it is possible to account for many mood uses (as well as non-grammatical modality uses). In English, for example, epistemic possibility can be expressed grammatically by may, can or the subjunctive, epistemic necessity grammatically by must or need, deontic possibility by the imperative, and deontic necessity by the imperative or the subjunctive. These uses are illustrated by the following examples: It may have happened that way. Can it have happened that way? Whatever the reason may be ...

epistemic possibiüty

8

Introduction

Whatever be the reason ... It must have happened that way. Need it have happened that way?

epistemic necessity

Take as many sweets as you like.

deontic possibility

Go to bed now. I suggest Smith be

deontic necessity

fired.

If a qualifier like "weak" is added, it is possible, furthermore, to account for the grammatical use of past tense forms like might, could, and should: It mightIcould have happened that way.

weak epistemic possibility

They should have reached their destination by now.

weak epistemic necessity

Nevertheless, there are several mood uses which do not fit into such a semantic system (with eight modal meanings). In English, for example, this goes for the Optative use of the subjunctive and the imperative {Long live Trotsky / Sleep well), the concessive use of the subjunctive {Be that as it may), and the hypothetical use of the subjunctive (If I were you...). In Danish, the use of skulle for the expression of report — illustrated by an example like Han skal vcere en dygtig pianist 'He is said to be an able pianist' — cannot be accommodated either, and the same goes for bürde used for the expression of probability — e. g. Den bor ligge her et sted 'It should/ought to be around here somewhere'. For the purposes of this book we have therefore chosen to base our description of mood usage in chapter 5 on an analysis (proposed in chapter 3) which is basically grammatical. In this description the members (of the category mood) possibility, necessity, subjunctive, and imperative are compared directly in the two languages. Furthermore, the use of the members probability and report in Danish is described and compared directly with the use of their nearest equivalents in English. Though a contrastive analysis of this type may by some linguists be regarded as theoretically less satisfactory than one based on a semantic tertium comparationis, it is still sufficiently semantic in its orientation (cf. the choice of terms) to ensure that apples in one language are not compared with pears in the other and thus to serve as a basis for a reliable description of usage. The approach adopted for the comparison of mood usage in chapter 5 can be simply illustrated by two examples:

Introduction

9

a. In positive declarative sentences possibility is realized by may in English and by kan in Danish, e. g. It may have happened that way and Det kan vcere sket pä den mäde. Here a member of the category mood found in both English and Danish is expressed by non-cognate and phonologically dissimilar verbs, and this constitutes an important difference between the two languages. b. In interrogative sentences possibility is realized by can in English and by kan in Danish, e. g. Can it have happened that way? and Kan det mre sket pä den mäde? In this case a member of the category mood found in both languages is expressed by cognate and phonologically similar verbs, and this constitutes an important similarity between the two languages. The contrastive description in chapter 8 of deontic modality expressed by semi-auxiliaries — an area of modality included in this book because of its dose affmity with mood (cf. 2.8) — selects as its point of departure the modal meanings permission, compulsion, Obligation, and duty. Although an analysis of modality in terms of features is absent here as well, the tertium comparationis is thus semantic as far as this part of the description is concerned. With regard to the method used in a contrastive analysis the following Progression has been recommended by Krzeszowski (see Krzeszowski 1967 and Jorgensen 1982: 69): 1) 2) 3) 4) 5) 6) 7)

Description of one language Description of another language Demonstration of comparability, establishment of basis of comparison Description of one language on the basis of comparison Description of the other language on the basis of comparison Comparison and demonstration of similarities Comparison and demonstration of differences

In this book we have found it neither necessary nor expedient to break down the account of tense and mood into such a large number of steps. As English and Danish are closely related languages, they can in many ways be compared directly, and a method like the one proposed by Krzeszowski — though generally sound — would therefore result in a rather repetitive presentation. For this reason steps 1 and 2 will be skipped here. On the other hand the remaining parts of the model will be retained, though not in any mechanical, point by point form.

10

Introduction

While mood is in this book interpreted broadly as a category which may be realized not only synthetically but also analytically, modality is interpreted more narrowly than in a number of other studies (particularly in the French tradition, cf. Herslund 1989) insofar as not all utterances are assumed to be modalized. If modality is understood as having to do with the speaker's commitment to the content of an utterance, a fundamental distinction may be drawn between categorical and modalized utterances, and this is in fact what we have chosen to do in chapter 3. By using utterances of the former type — for example Colonel Gaddafi is dead — the Speaker commits himself to the truth of the proposition. In utterances of the latter type no such commitment is signalled, cf. examples like Colonel Gaddafi may be dead and Perhaps Colonel Gaddafi is dead, by means of which the Speaker indicates that he considers it possible but cannot guarantee that a certain state of affairs obtains. Now if modality is understood differently as having to do with the speaker's attitude to the addressee, i. e. if modality is approached pragmatically rather than semantically, it is no longer meaningful to distinguish fundamentally between categorical and modalized utterances. As pointed out by Holmes, modality in this sense is directly related to the function of utterances in their social context and involves taking account of their illocutionary force (1982: 18). Since the Speaker conveys an attitude of some kind to the addressee whenever he produces an utterance, it follows that within such a pragmatic framework any utterance is modalized. This also applies to an example like Colonel Gaddafi is dead. Unless he pronounces it with marked prosody, the Speaker will by this utterance convey a confident attitude to the addressee. In the first paragraph of this chapter it was mentioned that the present investigation is limited to grammatical time and modality expressions and that lexical expressions are therefore excluded. At this point it may be added that as a result of this principle of delimitation prosodically signalled modality expressions are excluded as well. That Intonation has an attitudinal function has been known for a long time, and in a relatively early study like that by Schubiger it is stated explicitly that "It is akin to what in grammar is listed under the heading mood or modality'" (1958: 38). In the same book it is pointed out that besides expressing emotions by means of Intonation the Speaker in this way "also passes judgment upon the predication, he considers it as certain, probable, uncertain, improbable, negative" (ibid.). The use of Intonation for such epistemic purposes can be illustrated by the following examples (quoted from Holmes 1982: 14):

Introduction

a) b)

11

' / believe \ they are 'very in' volved. I be'^lieve \ they are 'very in^ volved.

Although both a) and b) are modalized lexically by means of believe, they do not present the State of affairs described by the noun clause as equally probable. By placing the falling-rising nucleus on the inherently stressed syllable of the modal verb in b) the Speaker brings the semantic element of supposition in believe into focus and thereby presents the strong involvement of the persons referred to as rather uncertain. In a), on the other hand, the qualifying verb is unstressed, and at the same time the placement of the nucleus on the personal pronoun I makes this utterance more assertive. As a result, the State of affairs described by the noun clause is here presented as more probable than in b). Although the Speaker does not commit himself to the content of this clause in a) either, he presents the Situation referred to as very Ukely to be true. With respect to the practical importance of investigating tense and mood contrastively, it may be mentioned that learning to express temporal and modal meanings is known to present serious problems to foreign language learners. For example, French and Dutch students have been shown to find it very difficult to modalize utterances appropriately in English, and contrastive work within the area is therefore considered sorely needed (see Robberecht — Van Peteghem 1982 and Aarts — Wekker 1987). In a paper by Holmes, similarly, it is pointed out that many learners have difficulty in expressing doubt and certainty, i. e. epistemic modality, appropriately in English and that "[w]ithout a command of such skills speakers tend to sound abrupt, rude or didactic in different situations" (1982: 24). As for tense, it appears from a study of learner English edited by Swan and Smith (1987) that e. g. the distribution of the present vs. future tense in English differs from that found in a large number of other languages with which English is compared and that the present tense is often used by speakers of these languages where the future should have been used instead. From the point of view of language teaching and learning the categories tense and mood/modality are therefore of primary importance. As far as English and Danish are concerned, however, only a few brief studies have been concluded so far (e. g. Lauridsen 1988, Davidsen-Nielsen 1985, 1986), a notable exception being a comprehensive investigation of modal verbs by Lauridsen (forthcoming). What the present book hopes to offer, therefore, is a — long overdue — contrastive description which will prove useful in English

12

Introduction

language teaching for Danes and Danish language teaching for speakers of English. Obviously, a book covering lexical expressions of time and modality in addition to grammatical ones would have been pedagogically preferable. For example, the use of modal particles such as vel 'presumably, I think', nok 'as likely as not', vist 'probably, I think', and mon 'I wonder' in Danish is bound to be an unnervingly high hurdle to Englishspeaking learners. Unfortunately, the amount of time such a complete investigation would require has made such a project impossible at this stage. As for the data presented in this book, it should be pointed out that the English examples have mostly been excerpted from recent books (particularly novels), newspapers, and periodicals and that the Danish examples have largely been extracted from the DANWORD corpus (University of Copenhagen). However, there are also many Danish examples which have been excerpted from books and newspapers and a few English examples which have been drawn from the Lancaster-Oslo/ Bergen (LOB) corpus. Furthermore, it has to a certain extent been necessary to make up examples in both languages, namely in those instances where the relevant data were not easily to be found in either Corpora or other texts. In such cases I have used my own intuitions about acceptability, not only with regard to Danish but also with regard to English. When in doubt, however, I have conferred with native speakers of English. Sometimes I have also found it useful to modify authentic examples, for example by changing declarative into interrogative sentences, thereby producing semi-authentic examples. Finally, the book includes a number of examples quoted from other works on English and Danish grammar. As far as these are concerned, the source is in each case indicated with page reference. In the remaining cases — where the data come from a variety of different sources — the benefit and interest to the reader of being informed about the source of each individual example have not been regarded as sufficiently great to necessitate an indication of sources (in the text and in the bibliography), which is after all both time-consuming and space-consuming. The view held here on data is that the examples should be the ones that are relevant to the investigation in all its aspects and that it does not matter whether they are derived from corpora, elicitation, or introspection as long as they are relevant. I wish to emphasize that an investigation like the present one could not have proceeded from authentic examples exclusively, for instance examples drawn from corpora, for in that case many relevant data might have been excluded (see Jensen, Sorensen, and Vikner 1989). For

Introduction

13

instance, sentences in the future perfect of the past are very hard to find outside conditional constructions in Danish, but this does not mean that examples of the type Jeg häbede, hans raserianfald snart ville have kulmineret 'I was Hoping his fit of rage would have culminated soon' can be ignored. Similarly, the use oineed for the expression of epistemic necessity in interrogative sentences in English is very difficult to attest by means of Corpora (cf Palmer 1979: 56), but this does not in our view justify ignoring examples of the type Need it have happened that way? The type of English investigated in this book is British English in its Standard version, and other varieties of English — whether determined regionally, socially, stylistically, or by age — will largely be left aside. Nevertheless, American usage will be commented on in a number of cases where it differs from British usage. In the contrastive description of tense in chapter 6, for example, it is pointed out that the tendency towards focussing on a past event rather than on its present relevance and accordingly to select the past tense instead of the present perfect is more pronounced in American than in British English, compare examples like Did you brush your teeth? and Have you brushed your teeth? (6.4). In this respect American English is even more different than British English from Danish, for in Danish the balance of focus relatively often inclines towards the current relevance of a past event rather than on the past event itself, cf an example like Dickens har skrevet mange romaner 'Dickens wrote many novels'. In the contrastive description of mood in chapter 5, similarly, the use of the so-called mandative subjunctive in American as compared with British English is mentioned (e. g. Peg suggested that he see a psychiatrist in Providence) and so is the epistemic use of the subjunctive after the conjunction lest in American EngUsh (e.g Lest he forget...) (5.6). In chapter 8 on semi-auxiliaries and deontic modality, furthermore, it is observed that the use of shall in questions with first person subjects about the will of the addressee — for instance Shall I dose the window? and Shall we go to the theatre? — is less common in American than in British English and that for the expression of this meaning Would you like mejus to ...? or Should Ijwe ...? is normally preferred (8.2). In the description of Danish we shall also basically restrict ourselves to the Standard language. While regional and social varieties of Danish will not be discussed at all, Variation determined by age will occasionally be mentioned. For example, it is pointed out in 2.7 that the modal verb behove 'need' requires a following Infinitive marker in conservative Danish but not in advanced Danish whereas the modal verbs turde 'dare' and

14

Introduction

gide 'feel like' combine with the bare infinitive in the conservative variety of Danish but are beginning to accept infinitives with at 'to' among younger speakers. In both English and Danish, furthermore, stylistically determined Variation will occasionally be commented on. For example, it is pointed out in 8.1 that deontic can and kan are often used to indicate a more casual and personally unobtrusive permission than may and mä. Finally it should be mentioned that by far the majority of examples from which the contrastive description of English and Danish proceeds are from written sources. We assume, however, that the grammar of time and modality expressions established on the basis of such data is to a considerable extent valid for spoken English and Danish as well. Before closing this chapter we should point out that the present investigation is not carried out as part of the exploration of any specific linguistic framework. Although the emphasis is on facts about English and Danish, it is naturally impossible to formulate such facts without some theoretical assumptions. The general view of language that is implied in this book is, I believe, dose to the mainstream of European linguistics, with both "structuralist" and "functionahst" aspects, as found also in e. g. the Prague school. In relation to the area covered in this book, one of the central assumptions is the view that both meaning and structure are necessary elements in the description of language — that these two factors must be understood in dose relation to each other, but that they should not be assumed to stand in a one-to-one relationship. As it makes use of very little formalism and explains many grammatical terms, the book ought to be comprehensible to a fairly wide group of readers, including students at universities and other institutions of higher learning. As the results put forward in it are largely new, it should at the same time be of interest to professional Unguists, no matter what their theoretical persuasion may be.

Chapter 2 Auxiliaries

2.1. Introduction In the course of history English and Danish have developed typologically from being predominantly synthetic (inflecting) to being predominantly analytic. In this process of morphological reduction the two languages have not become grammatically impoverished, for what was once expressed by means of inflections is to a large extent expressed syntactically nowadays, for example by means of word order and grammatical form words. To be sure, some of the work originally done morphologically is done lexically today. For example, there are lexical verbs — such as let in Enghsh and lade in Danish — which express what was formerly taken care of by the subjunctive. Similarly, in an Old English sentence like Se pe his gewealdes monnan ofslea, swelte se deape 'He who kills a man of his own free will is to die', the subjunctive form swelte 'die' expresses a command/necessity which is expressed with the lexical verb be to in modern English. Already a thousand years ago this type of substitution was beginning to take place, as is apparent from a sentence like Se pe slea his fader ... se sceal deape sweltan 'He who kills his father is to die', where instead of subjunctive swelte we find a combination of a modal lexical verb in the indicative and an Infinitive: sceal sweltan (Quirk and Wrenn 1957: 85). But over the centuries various syntactic devices have Seen the light. One of these is a system of auxiliaries, i. e. grammatical "helping verbs" such as have 'have', vcere 'be', blive 'become', ville 'will' (temporal) in Danish, which typically occur in conjunction with lexical verbs, which they premodify. In an Old English sentence like Ic hcebbe pone fisc gefangenne, for example, hcebbe is a lexical verb which means 'possess, hold' (I hold (have) the fish as caught). Subsequently, however, this verb developed into a mere grammatical Instrument — an auxiliary — when combined with a past participle which had lost its case inflection. As a result it began to be used also in cases where the original meaning was out of place, for example in I have lost the key (Jespersen 1909 — 49, IV: 29 f). Although some linguists question their existence, auxiliaries are recognized in most descriptions of modern English and Danish. According

16

Auxiliaries

to a Standard description of English, the class of auxiliaries comprises have, do, be (which are termed primary) and can, may, must, shall, will, as well as (marginally) dare, med, ought to, used to (all of which are termed modal). In Danish it is customary to recognize as auxiliaries have 'have', vcere 'be', blive 'become' (primary) and kunne 'can', matte 'must, may', ville 'will', skulle 'shall', turde 'dare', gide 'feel like', bürde 'ought to' (modal), see, for example, Aage Hansen (1967). Auxiliaries are separated from lexical verbs by a combination of morphological, syntactic, and semantic criteria. However, this Separation is by no means straightforward, as reflected by the fact that grammarians usually find it necessary to operate with "marginal" auxiliaries as well as "semi-auxiliaries" in addition to auxiliaries proper. A distinction may be drawn between auxiliaries and catenatives (Twaddell 1965: 22, Huddieston 1984: 142). Catenatives are lexical verbs like keep, promise, get, hope, help, like, want, etc. which resemble auxiliaries in combining with non-finite verb forms, as in They kept laughing, I promise to teil her. He got arrested by the police, which may be compared with They were laughing, I'll teil her. He was arrested by the police. In order to distinguish auxiliaries from catenatives in English, grammarians operate with the so-called NICE-properties: negation, inversion, code, emphatic affirmation (see Twaddell 1965 and Huddieston 1976). In negative, interrogative, and emphatic sentences, and with respect to "verbum vicarium", auxiliaries differ from other verbs in requiring no ^/o-support. This may be illustrated with the following examples: They were laughing. They kept laughing. Negation: Inversion:

They weren't laughing. They didn 't keep laughing. Were they laughing? Did they keep laughing?

Code:

They were laughing, weren 't they? They kept laughing, didn't they?

Emphatic affirmation:

They were laughing. They did keep laughing.

Furthermore, auxiliaries in EngUsh except be differ morphologically from lexical verbs in being defective, and semantically from most lexical verbs in expressing only general — temporal, aspectual, modal, or diathetic — meanings. (Such meanings are not expressed by catenatives like try.

Introduction

17

remember, enjoy, forget, etc.) In spite of the many criteria which have been proposed, the Separation of auxiliaries from lexical verbs is by no means easy, and in Danish, where the behaviour of auxiliaries in negative, interrogative, and emphatic sentences is no different from that of lexical verbs, and where auxiliaries are morphologically not very defective, the distinction is even more difficult to draw. In A Comprehensive Grammar of the English Language Quirk and his co-authors operate with a scale ranging from clear auxiliaries to clear füll verbs (1985: 137). Whereas a verb Hke may is assumed to belong unproblematically to the former category and a verb like begin unproblematically to the latter, it is not obvious how verbs such as need, ought to, had better, be to, have to, and seem to should be interpreted. Between the end points of their scale Quirk et al. operate with marginal modals, modal idioms, semi-auxiliaries, and catenatives. A sentence like / ought to go illustrates one such borderline case, where it is difficult to decide whether we are faced with one verb phrase containing an auxihary (as in / can go) or two verb phrases containing no auxiliary (as in / hope to go). In spite of their relativistic approach Quirk et al. do in fact propose a cut-off point, for on p. 120 they State that the English auxiliaries are the primary verbs be, have, do and the modal verbs can (could), may (might), will (would), shall {should), and must. This means that modal idioms, semi-auxiliaries, and catenatives are all excluded from the class of auxiliaries proper. The recognition of an auxiliary-full verb scale brings to mind the notion of imprecise categories, which is discussed by Osten Dahl in his book on tense and aspect (1985). In Dahl's opinion grammatical categories are typically imprecise. In this way they resemble everyday terms such as bald and bird, which are virtually impossible to defme in such a way that there will be no unclear borderline cases. Speaking of imprecise categories is practically the same as speaking offocused categories for both concepts imply that not all members of a category have the same status. Whereas the focus of a category is constituted by the "best exemplar", or "prototype", the entities that belong to the periphery have a more or less dubious membership. Thus although a penguin for example, is within the extension of the category "bird" (being a feathered warm-blooded vertebrate with two legs and two wings) it belongs to the periphery of this category (as a result of its inability to fly, sing, and be spotted in trees). The category "auxiliary" verb seems to be a good example of an imprecise linguistic category, for whichever set of criteria we decide to use to Single out its members, there are likely to be borderline cases. This

18

Auxiliaries

should not make us uneasy, for such borderline cases are interesting precisely because they are auxiliaries in some respects and füll verbs in others. As pointed out by Dahl, a linguistic notion may in fact be less useful if in our attempts to clarify it we eliminate genuine imprecision. If we look at the history of English and Danish, as well as that of many other languages, it is small wonder that auxiliaries do not make up a very precise category. In the slow process of change from a predominantly synthetic to a predominantly analytic language type a system of auxiliary verbs has gradually arisen, and some of these verbs have not yet acquired as secure a foothold in the category as others.

2.2. Epistemic and non-epistemic modals Before proceeding to the Separation of auxiliaries from lexical verbs in English and Danish, it is necessary to discuss the distinction between epistemic and non-epistemic (or root) modals, i. e. between modal verbs which judge the truth value of a Statement (as in The economy may get even worse) and those which do not have this property but have to do with what ought to, or is allowed to, happen ("deontic" modahty, as in You may leave the classroom) or with a quality or volition whose source is the referent of the subject noun phrase ("subject oriented", or "dynamic", modality as in Peter can read Greek). It will be argued here that just as verbs like have, do, be in English and have, vcere, blive in Danish differ from other verbs in belonging to two different classes, namely those of auxiliaries and lexical verbs, the verbs traditionally classified as modal auxiliaries belong to two different classes as well (see, for example, Anderson 1971). This division is justified for a number of reasons. In the first place, they are practically all ambiguous (polysemous), as illustrated by the following selected examples from Enghsh and Danish: She may come back at any moment. You may borrow my bike. You must be fifty at least. You must do it at once. You can't be fifty. You can 't borrow my bike. It needn't happen. You needn 't do that.

Epistemic

and non-epistemic

modals

19

Han km vcere tilbage hvert ojeblik. 'He may be back any moment.' Han kan sove hos mig. 'He can sleep at my place.' Du mä \cere mindst halvtreds. 'You must be fifty at least.' Du mä gore det omgäende. 'You must do it at once.' Peter skal vcere ekspert pä omrädet. 'Peter is said to be an expert in this field.' Peter skal vcere hjemme senest kl. 11. 'Peter must be back by eleven at the latest.' Den bor ligge her et sted. 'It should be lying here somewhere.' Vi ber starte straks. 'We ought to Start at once.' In each pair the verb in the first sentence is epistemic while that of the second sentence is non-epistemic (deontic in the examples given here). That the difference in meaning is not a function of the linguistic context in which the modal occurs, but that genuine ambiguity is indeed involved, is apparent from examples such as She may go Home tomorrow / Hun kan rejse hjem i morgen and They should be home / De bor vcere hjemme in which each modal has two meanings — one epistemic and one nonepistemic — in exactly the same context. As a distributional confirmation of this distinction, we can note that epistemic and non-epistemic verbs differ in their combinability with a following perfect infmitive with past time meaning. The former verbs show no restrictions whatever in this respect:

John

may must can't needn 't

have passed the exam.

John

kan mä skal .bor

have bestäet eksamen.

According to Palmer (1979: 36 f) epistemic modals are kept apart from all other modals in English by this criterion. It should be pointed out,

20

Auxiliaries

though, that deontic should and ought to do in fact occur in this context and that past tense dynamic could is also possible here: You ought tojshould have done it yesterday. John could have ßnished the job if he had wanted to. In Danish, the corresponding non-epistemic verbs also permit a following perfect infinitive with past time meaning: Du bürde/skulle have gjort det i gär. John kunne have afsluttet jobbet, hvis han havde villet. Another syntactic argument in favour of assigning the modal verbs to two classes is provided by conditional clauses. In this context nonepistemic modals occur freely, cf. examples like If I may borrow your bike... and Hvis jeg skal vcere hjemme kl. 11... 'If I have to be back by eleven ...', whereas epistemic modals are not normally permitted, cf. the unacceptability of examples like *If she must be abroad... and *Hvis hun mä vare i udlandet.... It should be added, though, that given specific larger contexts one occasionally comes across epistemic modals in conditional clauses. If A says He may come back at any moment, B may reply, epistemically as well, I f , as you say, he may come back at any moment, we'd better hurry. In Danish, similarly, epistemic modals cannot be Said to be ruled out in this context, as shown by the following example: Hvis Peter kan komme allerede kl. 3, sä mä vi se at komme i gang. 'If Peter may be here already at three, we'd better get started.' In English,thirdly, the epistemic modals differ from the non-epistemic ones in combining freely with the progressive aspect: He may be coming back at any moment. * You may be borrowing my bike. You must be approaching fifty. * You must be doing it at once. They can 't be leaving yet. * You can 't be borrowing my bike. It needn 't be happening. * You needn 't be doing that. As in the case of a following perfect infinitive, non-epistemic modals cannot be ruled out completely in this context, cf. the acceptability of a sentence like You must be working when I return.

Criteria of auxiliarity

21

That epistemic modals are free to combine with perfect and progressive constructions follows naturally from the fact that all they do is to judge truth values of various events. Such events may already be terminated or they may be in progress. For the same reason the epistemic modals in Danish combine freely with lexical verb phrases like vare ifcerd med and vcsre ved at 'be in the process o f . For example:

Hun

km mä skal .bor

•

• vare

ved ifcerd med

at skrive en roman.

'She may/must/is said to/should be writing a novel.' On the other hand, one cannot easily give oblige, permit, etc. the performing of events which are terminated or in progress. In Danish, fmally, epistemic modals differ from non-epistemic modals with respect to objects and place adverbials. As it appears from examples like Hun kanlvil/mälskal en masse 'She can/wants to/is allowed to/is about to do a lot of things' and Jeg vil/mälskallber hjem 'I want to/must/have to/ought to go home', the non-epistemic modals kunne, ville, mätte, skulle, bürde may occur by themselves and take objects (first four) or govern a place adverbial (last four). In the case of their epistemic counterparts both these distributions are ruled out. Furthermore the epistemic modals differ from the non-epistemic with respect to passivization. While the former require the Zj/?ve-passive, the latter require the -s passive, cf. examples like Peter bor blive udncsvnt snart and Peter bor udneevnes snart 'Peter ought to be appointed soon'. For these semantic and syntactic reasons we conclude that the modal verbs of English and Danish — with the exception of dare, shall and turde 'dare', gide 'feel like', which have no epistemic uses — belong to two classes, one of which is epistemic and the other of which is nonepistemic. '

2.3. Criteria of auxiliarity Even if we restrict ourselves to one language, the Separation of auxiliaries from füll verbs is no simple matter. In a contrastive analysis of two or more languages it is still more difficult to Single out auxiliaries, for in Order for the contrastive analysis to be reliable the criteria employed have

22

Auxiliaries

to be the same. Otherwise there is no guarantee that we will not be comparing apples in one language with pears in another. As the widely used NICE-criteria which have been proposed for English all involve a special use of the verb do, they are of no real value in a contrastive framework. This means that we have to search for criteria of a more general nature. In a paper dating from 1983 such criteria have been proposed by Ebbe Spang-Hanssen. The aim of this paper is to set up criteria which permit the analyst to Single out a class of grammatical verbs which are intimately connected with another verb and which are similar to verbal inflections.lt is not considered sufficient to propose a number of operational tests: the criteria should also isolate a class of verbs which is useful for the formulation of grammatical rules. Above all, it is considered important that auxiliaries — in so far as these occur in a language — behave like verbal grammatical endings. Among the eight criteria discussed by Spang-Hanssen there are four which are particularly helpful in a contrastive description of English and Danish and which we shall therefore select as our working tools: (i)

The meaning of an auxiliary is general and abstract.^ Apart from semantically empty words — like English do — its content is analysable in terms of temporal, aspectual, modal, or diathetic values. (ii) An auxiliary is functionally dependent in the sense that it is impossible to modify it without simultaneously modifying the lexical verb it Combines with. In other words, Aux + V can only be modified globally. (iii) The addition of an auxiliary does not affect the lexical restrictions of the verb it combines with. (iv) An auxiliary is attached to a lexical verb without any intervening Infinitive marker, i. e. it governs a bare Infinitive or a participle. Verbs which satisfy these four criteria are similar to verbal inflections since (i) the meanings of such inflections are general and analysable in terms of temporal, aspectual, modal, or diathetic values; (ii) inflections cannot be modified alone; (iii) inflections do not affect the lexical restrictions of the verbs they are attached to; (iv) inflections are attached directly to the verbs they modify. The reason why it is considered important that in Order for a form to be considered an auxiliary it should behave like verbal inflections is that these are indisputably grammatical entities and that what we are seeking to isolate is a class of grammatical verbs. In the following the above criteria will be examined one by one.

The semantic criterion

23

2.4. The semantic criterion Whereas temporal, aspectual, and diathetic meanings appear to be unproblematically general, this is not the case with all modal meanings, for there is a marked difference in generality of meaning between epistemic and non-epistemic modality. This is apparent from Danish examples like De kan vcere pä ferie 'They may be on vacation' and Mine studenter kan läse russisk 'My students can read Russian'. In the second sentence nonepistemic kan attributes specific properties to the agent referred to by the subject noun phrase Mine studenter, and the scope of the modal is restricted to this noun phrase. In the first sentence, on the other hand, the meaning of epistemic kan is as general as can be — logical possibility — and the scope of the modal is the whole proposition. According to Bybee and Pagliuca (1985) epistemic uses of modal verbs have developed from originally non-epistemic uses. For example, English must was used in the sense of Obligation in the earliest written documents as it is today, but in the 17th Century it took on the epistemic sense of inferred certainty. This type of development is seen as a semantic generalization which has taken place through metaphorical extension. As grammatical functions are necessarily abstract, such an emptying of lexical content is regarded as a prerequisite to grammaticization. As far as abstractness of meaning is concerned, deontic modality appears to be intermediate between epistemic and dynamic modality, cf examples like Can this be true? (most abstract), Can you swim? (least abstract), and Can I borrow your bike? (intermediate). In the course of history, an English modal like can has in some of its uses lost certain specificities in its meaning (see Bybee and Paghuca 1985: 68). Originally, it was used only dynamically in the sense of ability, i. e. it attributed certain physical or mental properties to an animate subject noun phrase referent. At a later stage, it developed the deontic meaning of permission, i. e. the properties expressed by it no longer had to be physical or mental but could be social as well. Moreover, can developed the epistemic meaning of possibility, i. e. its original semantic content was generalized in such a way that it no longer had to attribute properties to an animate subject noun phrase referent and that its scope could be extended to the whole proposition. Bybee and Pagliuca emphasize that inflectionally marked modalities are nearly always epistemic. On the basis of studies of a large number of languages they conclude that in inflection the rarity of non-epistemic

24

Auxiliaries

modalities is in striking contrast to the frequency of epistemic modalities. Although this is undoubtedly true cross-linguistically, it must be pointed out, however, that in English and Danish (and some other languages as well) inflectionally marked modalities are also deontic. This is apparent from the fact that compulsion and wish are expressed by - 0 and (in Danish) -e in the imperative and subjunctive, cf. examples like Go to bedj Gä i seng and Long live the Queen/Lange leve Dronningen. As it would not be reasonable to require that the meanings expressed by auxiliaries be more abstract than those expressed by inflection — the extreme form of grammaticization — we shall conclude that the deontic modals are sufficiently abstract to be recognized as auxiliaries, in spite of the fact that their meanings are less general than those of epistemic modals. According to the criterion of general meaning epistemic modal verbs are clearly included in the class of auxiliaries, and the same goes for verbs with temporal, aspectual, and diathetic meanings or which are semantically empty. Also included, though less obviously so, are the deontic modals. On the other hand, the dynamic modals are clearly excluded. In English this means that the following verbs are semantically auxiliaries: be, have, do, will (temporal), may, might, can, could of permission and epistemic possibility, must, need, should of compulsion and epistemic necessity, and shall, should of Obligation. In Danish the verbs singled out by the semantic criterion are vcere, have, blive, ville (temporal), kunne of permission and epistemic possibility, matte of permission, compulsion, and epistemic necessity, behove 'need' of compulsion and epistemic necessity, skulle of compulsion,obligation, epistemic report and (weakened) epistemic necessity, and bürde of duty and epistemic probability. The criterion of general meaning includes not only hard core epistemic and deontic modals but also some modals which are not normally regarded as auxiliaries. In English, for example, be bound to, be certain to, be sure to, have to, and have got to, all of which express necessity, are auxiliaries by this criterion. By way of Illustration the following examples may be given (Palmer 1979: 45 f): It's bound to come through, I think. If you've Seen all the old Frankensteins you've got to know all the jokes.

The criterion of functional

dependency

25

2.5. The criterion of functional dependency A verbal inflection does not constitute an independent lexical unit and therefore cannot be modified by itself. In a sentence like / handed her the tray carefully, for example,the adverb obviously does not modify the past tense inflection alone but also the verb to which it is attached. In order to eUminate from the class of auxiliaries those verbs which differ from verbal inflections in constituting independent lexical units, we shall therefore investigate whether the structure V + V" — in which V symbolizes a Potential auxiliary and V" a following non-finite verb — is modified globally or not. If the modification is global, V + V" behaves like V + inflection, and V may be analysed as an auxiliary. If V is modified by itself, on the other hand, it cannot be included in the class of auxiliaries. Consider first the following Danish sentences with non-epistemic modal verbs: Han vil absolut byde hende en drink. 'He insists on offering her a drink.' Han bor absolut byde hende en drink. 'He definitely ought to offer her a drink.' Han mä absolut byde hende en drink. 'He must definitely offer her a drink.' Han skal absolut byde hende en drink. 'He must definitely offer her a drink.' Han mä meget gerne byde hende en drink. 'He is welcome to offer her a drink.' Han kan let lobe 100 meter pä 11 sekunder. 'He can easily run 100 metres in 11 seconds.' Han gider sjceldent gore sit arbejde ordentligt. 'He seldom bothers to do his job properly.' Han tor aldrig sige nej. 'He never dares to say no.' In all these examples the adverbial modifies the modal verb alone,i. e. the modification is non-global and the verb phrase has the following structure:

26

Auxiliaries

That this is so is apparent from the fact that it is possible to focalize V" and thereby attach a value to it which is not simultaneously attached to V: Hvad han absolut vil/bor/mä/skal er at byde hende en drink. 'What he defmitely wants to/ought to/must is to offer her a drink.' Hvad han meget gerne mä er at byde hende en drink. Hvad han let kan er at lobe 100 meter pä 11 sekunder. Hvad han sjceldent gider er at gore sit arbejde ordentligt. Hvad han aldrig tor er at sige nej. The functional independency of these non-epistemic modals also appears from the fact that it is possible to convert declarative sentences in which they occur into specific questions where V" ( + complementation) is interrogated (see Vikner 1980): r vil -N bor mä skal Han < >• opfore sig bedre. mä gerne kan gider ikke tor ikke 'He wants to/ought to/must/may/can/can't be bothered to/daren't behave better.'

The criterion of functional

Hvad er det han

vil ber mä skal gerne mä kan ikke gider ^ikke tor

dependency

27

>?

'What is it he wants, etc.?' In the case of the epistemic modals and the verbs have, vcere, blive, (temporal) ville, on the other hand, V is functionally dependent: Hun er besternt gäet hjem. 'She has definitely gone home.' Hun har besternt gjort det. 'She has definitely done it.' Hun bliver besternt valgt. 'She will definitely be elected.' Det vil besternt ske snart. 'It will definitely happen soon.' Det mä besternt ske snart. 'It must definitely happen soon.' Det kan bestemt ske snart. 'It may definitely happen soon.' Den ber absolut ligge her et sted. 'It should definitely be lying here somewhere.' Peter skal ubetinget vcere en därlig mekaniker. 'Peter is said to be a thoroughly poor mechanic.' In these sentences V is not modified by itself, and V" cannot be focalized in any of them. Here, then, the verb phrase has the structure:

vil

bestemt

ske

28

Auxiliaries

That this is so can also be seen from the fact that it is not possible to form questions where V" is interrogated, cf. the ungrammaticality of, for example, *Hvad er det det mä? '*What is it it must?'. In Danish,the criterion of functional dependency thus includes in the class of auxiliaries the verbs with temporal or diathetic meanings (vcere, have, blive, ville) as well as the epistemic modals kunne of possibility, matte and beheve of necessity, bürde of probability, and skulle of report. On the other hand, it excludes not only the dynamic but also the deontic modals. In this connection an analytical problem has to be mentioned, though. When used in the sense of Obligation (promises, threats) skulle behaves syntactically like the epistemic modals in spite of the fact that it is semantically non-epistemic. In a sentence like Hun skal snarest fä sine penge tilbage 'She shall have her money back very soon' the adverbial modifies V + V" globally, cf. that neither focalization nor interrogation of V" is possible: *Hvad hun snarest skal er (at) fä sine penge tilbagej *Hvad er det hun snarest skal? Furthermore, this skulle resembles the epistemic modals — including skulle used for the expression of report — with respect to passivization: it requires the construction blive + V + -et (see 2.2 above). This is apparent from an (ambiguous) example like Forestillingen skal blive spület i morgen, which signifies either an Obligation (You shall have the Performance played tomorrow) or a report (The Performance is said to be on tomorrow). When used for the expression of compulsion or arrangement, on the other hand, skulle selects the -s passive. This may be illustrated by a sentence like Forestillingen skal spilles i morgen, which is ambiguous as well (The Performance must be given tomorrow/The Performance is arranged for tomorrow). In English the Situation is less clear than in Danish, for one thing because focalization and interrogation of V" — in those cases where they are possible — require pro-verb support of V . Nevertheless the Situation is largely the same as in Danish. In the case of nearly all non-epistemic modals, V is functionally independent: You may/can deflnitely flnish the pie. You must definitely visit my aunt. He can easily swim a mile. I deflnitely won't pay the money back. I definitely daren't publish that manuscript. What the adverbials express here is that there can be no doubt about the permission, compulsion, ability, willingness, and resolution expressed by

The criterion of funcüonal

dependency

29

V . That they modify V rather than V + V" is also apparent from the fact that V" may be focalized: What What What What What

you definitely may do is finish the pie. you definitely must do is visit my aunt. he easily can do is swim a mile. I definitely won 't do is pay the money back. I definitely daren't do is publish that manuscript.

As it appears, c/o-support of V is required here, but this does not affect the point that V and V" are separated by the focalization transformation. In these cases, then, the verb phrase has the structure:

While the deontic modals expressing permission {may, can, might, could) and compulsion {must, have to, need, should, ought to) and the dynamic modals expressing ability {can, could, be able to), volition {will, would), and resolution {dare) are in this way excluded from the class of auxiliaries, shall used in the sense of Obligation or insistence — like Danish obligational skulle — behaves differently syntactically. In a sentence like He shall soon have his money back the adverbial modifies V + V" globally, cf. that neither focalization nor interrogation of V" is possible. A possible reason why shall behaves like an auxiliary is that its deontic meaning is sometimes fairly weak as compared with its future time meaning, cf. , for example, You shall not catch me so easily next time and If you work well, you shall have higher wages. Here shall is closely related to temporal will. In the case of be, have, do, temporal will, and the epistemic modals the Overall picture which emerges is that they are modified together with V", cf. examples like the following:

30

Auxiliaries

She was definitely defeated. She has definitely received the parcel. She really didn't like it. We will definitely be safer when we reach the mountains. She may definitely be in London now. Can this really be the aase? She must definitely be in London now. Need this really be the case? She shouldjought definitely (to) be in London now. Here V is functionally dependent, i. e. the verb phrase has the following structure:

was

d^nitely

defeated

That this is so appears from the fact that it is not possible to focalize V". For example, *What she definitely may (be) is (to) be in London now is clearly ungrammatical. In some cases involving this group of verbs, however, adverbial modification is not evidently global: She has definitely finished the Job. She may definitely work in London now. She must definitely work in London now. It will definitely happen soon. That V is not here clearly dependent appears from the fact that focalization can hardly be ruled out as altogether ungrammatical: IWhat she definitely has done is finished (finish?) the Job. IWhat she definitely may do is work in London now. ? What she definitely must do is work in London now. IWhat it definitely will do is happen soon.

The criterion of functional

dependency

31

Interrogation of V" cannot be dismissed as clearly ungrammatical either: IWhat ? What ? What IWhat

is it she definitely has done? is it she definitely may do? (in the sense of epistemic possiblity) is it she definitely must do? (in the sense of epistemic necessity) is it it definitely will do?

Summing up, we arrive at the following picture in Engüsh: the criterion of functional dependency exciudes from the class of auxiliaries not only the dynamic but also the deontic modals (with the Single exception of shall). The epistemic modals, the verbs with temporal, aspectual, or diathetic meanings, and the semantically empty verb do are included, although not clearly and unproblematically so. Among the epistemic verbs some are still included which are not normally regarded as auxiliaries, such as be bound to, have to, and have got to. That these are functionally dependent appears from examples like the following, where V + V" is modified globally: It's definitely bound to come out. He's definitely bound to be joking. Before concluding our discussion of functional dependency we have to mention another analytical problem, namely modification by means of the negatives not/ikke (as well as a few other negative adverbs, such as hardly, scarcelyjnappe). In the case of some of the modal verbs it is evidently V which is modified in this way, cf. the following examples cited from Quirk and Greenbaum (1973: 189): You may not go swimming. You can't be serious. You can't go swimming. She can't ride a bicycle. You needn 't pay that fine. It needn 't always be my fault.

(non-permission) (impossibility) (non-permission) (inability) (non-compulsion) (non-necessity)

The main problem here is that epistemic can and need — which have been shown to be functionally dependent in sentences like Can this really be the case? and Need this really be the case? — are functionally independent as far as negation is concerned. In sentences like This cannot really be the case and This needn't really be the case, therefore, the structure is like this:

32

Auxiliaries

As it appears, can and need are functionally dependent with respect to really but functionally independent with respect to not. In Danish, the epistemic modals kume and behove are also functionally independent as far as negation is concerned: Du kan ikke mene det. 'You can't be serious.'(impossibility) Det behover ikke (at) vcere min fejl. 'It needn't be my fault.'(non-necessity) With respect to the epistemic modals can, need in English and kmne,beh0ve in Danish, then, application of the criterion of functional dependency does not yield a wholly clear result. And as pointed out above, another analytical problem is posed by the fact that be, have, do, temporal will, and the epistemic modals,in English are not straightforwardly included in the class of auxiliaries according to this criterion. Within the group of non-epistemic modals, thirdly, one verb in either language — obligational shall/skulle — is the odd man out in being functionally dependent.

2.6. The criterion of permanence of the lexical restrictions of V" A verbal inflection does not usually affect the lexical restrictions of the verb it is attached to. If sentences like / bring you flowers and / appoint you director are correct, then sentences like He brings you flowers, I brought you flowers and / appointed you director are also correct. As already pointed out, auxiliaries are grammatical verbs which resemble verbal inflections, and it is therefore natural to require that they do not

The criterion of permanence of the lexical restrictions of V"

33

affect the lexical restrictions of the verbs they combine with either. Since the context to the right of the verb phrase is dependent on V" exclusively (cf., for example, He has brought you flowers and / will appoint you director), the question to be asked is therefore whether V + V" admits the same subjects as V" alone. If this is the case, V may be interpreted as an auxiliary. If it is not the case, it has to be interpreted as a lexical verb. It has been noted by many grammarians that in order to isolate a class of auxiliaries it is particularly useful to investigate impersonal subjects. If V" requires this type of subject — as in the case of rain, expire ( = come to an end) / regne, udlebe — V + V" should also accept it in order for V to be analysed as an auxiliary. As illustrated by the following examples, such subjects are admitted in the case of have, be, do, (temporal) will/ have, vcere, blive, (temporal) ville and epistemic verbs: English It has rained. It is raining. It doesn't rain (much here). It will rain. It may rain. It must rain (a lot in Ireland). It's bound to rain (a lot in Ireland). It could/might be raining. It should!ought to be raining (now a few miles away). Surely it can't rain (a lot in Ethiopia). It needn't rain (all the time). Danish Det har regnet. 'It has been raining.' Kontrakten er udlobet. 'The contract has expired.' Fakultetet bliver ikke berort af nedskceringerne. 'The faculty will not be affected by the cuts.' Det vil regne (i de kommende dage). 'It will rain (in the next few days).' Det kan regne (i Jylland netop nu). 'It may be raining (in Jutland right now).'

34

Auxiliaries

Det mä regne (meget i Irland). 'It must rain (a lot in Ireland).' Det bor regne (i Jylland netop nu). 'It should be raining (in Jutland right now).' Det skal regne (en del i Irland). 'It is Said to be (quite) rainy (in Ireland).' Behßver det (at) regne (i Jylland netop nu)? 'Need it be raining (in Jutland right now)?' In the case of non-epistemic modal verbs, on the other band, V + V" does not normally accept impersonal subjects. This is apparent from the following selected examples: English *The contract can expire. (in the sense of ability) *The contract will expire. (in the sense of volition) *The contract daren't expire. *It shall rain. Danish ^Kontrakten kan udlebe. (in the sense of ability) * Kontrakten vil udlobe. (in the sense of volition) * Kontrakten tor ikke udlobe. * Kontrakten gider ikke udlobe. 'The contract doesn't feel like expiring.' *Det skal regne. (in the senses of Obligation and arrangement) According to Bybee and Pagliuca (1985) non-epistemic modality differs from epistemic modality in being agent oriented, more specifically in involving a wilful agent. In the case of modalities such as ability, volition, resolution, and inclination, where the agent is the referent of the subject noun phrase, impersonal subjects are clearly mied out. With deontic modality, on the other hand, the Situation is different because the source of modalities such as permission, compulsion, Obligation, and duty is the Speaker. In certain situations, the Speaker may permit, compel, oblige, or

The criterion of direct attachment

35

morally require an impersonal subject noun phrase referent to release a certain action: It can/may rain now for all I eure. Det mä/kan for min skyld godt regne nu. The contract must expire before the first of October. Kontrakten skaljmä udlobe senest forste oktober. IThe contract shall expire as you desire. ^Kontrakten skal udlebe som onsket. This contract ought tojshould expire before the first of October. Kontrakten bor udlobe senest forste oktober. This, however, is the exception and not the rule. In the case of epistemic modality, on the other hand, nothing prevents the use of impersonal subjects since all epistemic modals do is to judge the truth value of an event. The criterion of the permanence of V" 's lexical restrictions (impersonal subjects) thus clearly includes among the auxiliaries epistemic modals as well as the verbs have, be, do, (temporal) will, and have, vcere, blive, (temporal) ville. On the other hand, it excludes the subject oriented modals completely and the deontic modals in their normal uses. In Danish it also excludes vcere ifcerd med 'be in the process of as well as the verb fä 'get', which expresses necessity + future time in idiomatic expressions like Vi fär se 'We shall see', Det fär \cere 'It will have to be left as it is', and future result when followed by a past participle in examples of the type Fär du snart skrevet noget? 'Will you get something written soon?'. This is apparent from the fact that verbs which require impersonal subjects, such as udlobe 'expire', do not permit preceding er i fcerd med, fär: * Kontrakten er i fcerd med at udlobe, * Kontrakten fär udlobe, *Hvornär fär kontrakten udlobet?

2.7. The criterion of direct attachment The criterion of direct attachment of V to V" (that is, without any intervening Infinitive marker) excludes some verbs from the class of auxiliaries which are included by the other criteria.In English, have to, have got to, be boundjcertainjsure to, and ought to of epistemic necessity are auxiliaries not only with respect to their meaning — which is general and modal — but also with respect to functional dependency and permanence of the lexical restrictions of V". This is apparent from examples

36

Auxiliaries

like the following, the first three of which illustrate global modification and the last three of which illustrate combinability with impersonal subjects: He's definitely got to be joking. We're definitely bound to be late. She definitely ought to be in London now. There has to be a way out. There's bound to be a way out. There ought to be a way out. The verb be going to, which meets the first three criteria as well, is also excluded from the class of auxiliaries by the criterion of direct attachment, and so are the modals seem to and appear to. In Danish this criterion excludes a verb like voere nedt til 'have to' of epistemic necessity,which qualifies as an auxiliary not only semantically but also as regards the other two syntactic criteria: Det er besternt nodt til at ske ingen lange. 'It definitely has to happen soon.' In conservative Standard Danish the criterion of direct attachment excludes behove of epistemic necessity as well,cf.an example like the following: Behover dette overhovedet at vcere sandt? 'Need this be true at all?' In advanced Standard Danish, on the other hand, this verb requires no infinitive marker (see Hansen 1977): Behover dette overhovedet vcere sandt? In this type of Danish, behove is thus an auxihary according to all four criteria in the same way as epistemic need in English. In this book we shall base our analysis on the non-conservative variety of Danish and accordingly Interpret epistemic behove as an auxiliary. In addition to excluding some epistemic modal verbs,the criterion of direct attachment excludes a number of verbs which are excluded by other criteria as well, such as used to of habit, be able to of ability, ought to of duty, have (got) to and need to of compulsion, and dare to of resolution in English. Similarly, it excludes have (at) and vcere nodt til

Analysis proposed

37

'have to' of compulsion as well as vare i stand til 'be able to' of ability in Danish. On the other hand, the criterion of direct attachment obviously includes in the class of auxiliaries a large number of catenatives which are excluded by the other criteria. In English, keep, get, stop, hear, feel, etc. and most of the non-epistemic modal verbs are auxiliaries according to this criterion. In Danish, the same applies to verbs like fä 'get', lade 'let', se 'see', komme 'come', and the majority of non-epistemic modal verbs. The criterion of direct attachment differs from the other three criteria in relating neither to meaning nor syntactic function, and it therefore appears to be of secondary importance. Whether V and V" are separated by a particle or not might seem to be an accidental product of history and therefore not of great relevance to the Separation of auxiliaries from lexical verbs. On closer inspection, however, a different picture emerges. As pointed out by Erik Hansen (1977) Danish turde and gide, which are non-auxiliary according to the first three criteria, and which in conservative Danish combine with the bare infmitive, are beginning to accept a?-infmitives: Han ter ikke at sporge/Vi gider ikke at vare med 'He daren't ask/ We don't feel like participating'. Conversely, behexe,^hich is an auxiliary according to the first three criteria when used epistemically, is losing its Infinitive marker. It thus appears that presence vs. absence of the infmitive marker is in fact a non-trivial surface indicator of auxiliary vs. non-auxiliary status. This does not imply, it must be stressed, that any verb which permits a following bare infinitive is an auxiliary.

2.8. Analysis proposed Although the Separation of auxiliaries from lexical verbs in Danish and English is no simple matter, it seems well motivated to require that in Order for a verb to be analysed as auxiliary it should satisfy all the four criteria which are satisfied by verbal inflections. Obviously, it would be analytically satisfactory if application of each of these criteria singled out exactly the same class of verbs. That this is not the case is due to the fact that we are dealing with natural, changing languages, in which there are verbs that have acquired some but not all of the properties of auxiliaries. In particular, the non-epistemic modal verbs behave like auxiliaries proper, partly with respect to direct attachment and partly with respect to verbum vicarium/code (as well as the other NICE-properties in English discussed in 2.1 above). It is naturally for this reason —

38

Auxiliaries

and because of their morphological defectiveness — that these verbs are normally classified as auxiliaries in English. However, the fact remains that they differ from verbal inflections in a number of significant ways. The dynamic modals can/kunne of ability, willjville of volition, darej turde, and gide are clearly non-grammatical with respect to meaning, functional dependency, and permanence of the lexical restrictions of V", and the exclusion of these verbs from the class of auxiliaries is therefore based on solid arguments. On the other hand, the status of the deontic modals may, can/mätte, kunne of permission, must, med, shouldjskulle, mätte of compulsion, shall, should jskulle of Obligation, and bürde of duty is more problematic. Besides attaching directly to V" they are grammatical in nature by under certain circumstances permitting impersonal subjects and by expressing a type of meaning (deontic) which may also be expressed morphologically. However, the deontic modals are non-grammatical in three ways: they are modified by themselves (with the isolated exception of obligational shalljskulle), their meanings are relatively concrete (more so than those of the epistemic modals), and they do not freely take impersonal subjects.For these reasons they will here be excluded from the class of auxiliaries proper,and they will consequently be assumed not to be manifestations of (grammatical) mood. Owing to their affmity to auxiliaries, however, they will be dealt with in a special chapter (8) on semi-auxiliaries and deontic modality. According to the selected criteria, the class of auxiliaries in English is thus considered to comprise the following verbs: have, be, i/o,(temporal) will and the modals may/can/might/could of epistemic possibility and must/need/should of epistemic necessity. These verbs are recognized as auxiliaries in spite of the fact that their inclusion is not straightforward with respect to one of the criteria, namely that of functional dependency: as pointed out in 2.5 above they are perhaps not invariably dependent, and as far as modification by negative adverbs is concerned need and can are clearly independent. In Danish the class of auxiliaries is assumed to comprise the following verbs: have, \cere, blive, (temporal) ville and the epistemic modals kunne of possibility, mätte, behove of necessity, bürde of probability, and skulle of report/rumour and necessity. It may be added that the modal auxiliaries kunne, mätte, skulle, bürde, behove differ from their lexical counterparts and the remaining lexical modals not only with respect to some of the criteria discussed above and with respect to distribution (cf. 2.2), but also morphologically. they do

Analysis proposed

39

not normally occur in the past participle.^ That lexical modals accept this form freely is apparent from examples like these: Jeg har fer kunnet lose sädanne opgaver. Tve been able to solve such problems before.' Jeg har fer mättetjskullet lose sädanne opgaver. Tve had to solve such problems before.' Jeg har ikke gidet besoge dem. 'I haven't feit like visiting them.' Jeg har ikke turdet besoge dem. 'I haven't dared to visit them.' Jeg har aldrig behovet at forberede mig ret lange. Tve never had to prepare for very long.' The primary auxiliaries have, vare, and ville do not have past participles either, in contradistinction to their lexical counterparts: Jeg har Haft en ubehagelig oplevelse. T v e had an unpleasant experience.' Jeg har vceret hjemme hele dagen. Tve been home all day.' Jeg har Icenge villet besoge dig. Tve wanted to visit you for a long time.' Having neither past participle nor present participle form, the auxiliaries in Danish (except blive) are morphologically quite defective, their only non-fmite form being the Infinitive. Auxiliary blive, it should be added, occurs freely in the past participle, as illustrated by a sentence like Hm er blevet valgt som formand 'She has been elected chairman'. In conclusion, the auxiliary systems that we shall operate with in English and Danish may be summarized and exemplified in the following way: Auxiliaries in English Primary

have be do will

(perfect tense) (progressive aspect, passive voice) (empty) (future tense)

Modal

may, might, can, could must, need,should

(epistemic possibility) (epistemic necessity)

40

Auxiliaries

Examples The dog has run away. My foot is hurting. We were caught in a traffic jam. I didn't see him. Dimer will be at eight. The road may be blocked. The road might be blocked. Can spring be far behind? It could be true. You must be tired. It needn't be the only explanation. They should be back now.

Auxiliaries in Danish Primary

Modal

have vare blive ville kunne mätte, skulle bürde

(perfect tense) (perfect tense, passive voice) (passive voice) (future tense) (epistemic possibility) (epistemic necessity) (epistemic report and necessity) (epistemic probability)

behove

Examples Vi har kendt hinanden i mange är.

'We have known each other for many years.' Hunden er lebet vcek.

'The dog has run away.' Jeg er bundet af en uheldig

kontrakt.

T m bound by an unfortunate contract.' Han blev drcebt i krigen.

'He was killed in the war.' Denne afgorelse vil fä alvorlige

konsekvenser.

'This decision will have serious consequences.' Det kan vcere sandt.

'It may be true.' Det mä vcere sandt.

'It must be true.' Det behover ikke vare sandt.

Analysis proposed

'It needn't be true.' Han skal mre en udmcerket mekaniker. 'He is Said to be an excellent mechanic.' Peter skulle vare i London nu. 'Peter should be in London now.' Peter bor vcere i Loridon nu. 'Peter ought to be in London now.'

41

Chapter 3 Analysis of mood

3.1. The concept of modality By modal concepts are understood concepts of what is possible, what is necessary, what is probable, what is conceivable, and the like. The idea of modality is an old one, going back to classical Greek philosophy. Aristotle attaches particular importance to the notions of possibility and necessity. The emergence of such notions seems to be due to the fact that human beings frequently think and behave as if things might be other than in point of fact they are (see Perkins 1983: 6). In producing modalized utterances such as Perhaps Colonel Gaddafi is dead, It's possible that Colone! Gaddafi is dead, Colonel Gaddafi is possibly dead, Colonel Gaddafi may be dead, and Colonel Gaddafi could be dead, for example, the Speaker expresses that a certain State of affairs is conceivably real. With an unmodalized (categorical) utterance like Colonel Gaddafi is dead, on the other hand, he refers to a State of affairs which he considers factual. In recent decades philosophers and logicians have attempted to analyse modal notions by construing them as statements about possible worlds. Instead of saying, for example, "it is possible that such-and-such" they say "there exists a possible world where such-and-such". By a possible World they understand — besides the real world which constitutes one such world — any complete alternative to the way things actually are. In the modalized examples given above the Speaker operates with an alternative world in which Colonel Gaddafi is dead. In an utterance like You ought to be in bed, similarly, he operates with a possible world in which things are as he would like them to be. To regard something as being otherwise is thus to regard it as being true or real in some nonactual world. If we disregard the real world, possible worlds may be divided into different types (called modalities) according to the conceptual framework within which an event or proposition is considered real or true. A number of linguists are of the opinion that there are three basic modalities (see, for example, Perkins 1983: 10 ff):

44

Analysis of mood

1) A modality which is concerned with rational laws of inference and deduction. 2) A modality which is concerned with social or institutional laws. 3) A modality which is concerned with the relationship between empirical circumstances and the states of affairs which follow from them, that is, with natural laws. The first of these, which may be illustrated by examples Uke The economy may get worse and She must be fifty at least, as well as by the Colonel Gaddafi examples above, is termed epistemic (derived from the Greek Word for knowledge episteme). The second, which may be exemplified by sentences such as You may borrow my bike, You must do it at once, and He should go to bed, is termed deontic. The possible world involved here is one which is morally acceptable, either to the Speaker or according to a set of principles established by some legal authority or Institution. The last type of modality, which may be illustrated by examples like She can swim and / am capable of reading Greek, is termed dynamic and refers to "the relationship which exists between circumstances and unactualized events in accordance with natural laws" (Perkins 1983: 11). In this case, therefore, the possible world is one in which events are seen as following from certain empirical circumstances. Since the source of modality is here the referent of the subject noun phrase, the term subject oriented is sometimes used instead of dynamic (cf. 2.2 above). According to some linguists this type of modality includes — besides expressions of ability — volitional expressions such as will, be Willing to, and want (psychological laws being regarded as a subcategory of natural laws along with those of physics, chemistry, biology, etc.). It could also be argued, however, that they belong to the deontic type of modality. While it is fairly easy to understand that the notion of possibility lends itself to being construed in terms of possible worlds, it is less obvious that the notion of necessity is naturally accounted for in this way as well. Note, however, that although a modalized utterance like They must have arrived expresses necessity, it still differs from a categorical utterance like They have arrived in being non-factual and in this sense connected with some degree of uncertainty. Between the poles of a negative categorical utterance and a positive categorical utterance there is a modal area ranging from what is considered remotely possible to what is believed but now known to be necessary:

The concept of modality

De er ikke kommet frem.

De De De De De

45

They haven't arrived.

(low probability) kunne vcere kommet frem. They mightjcould have arrived. kan vcere kommet frem. They may have arrived. bürde vcere kommet frem. , ,,, , , • . , , . They should ought to have arrived. bor vcere kommet frem. mä vcere kommet frem. They must have arrived. (high probability)

De er kommet frem.

They have arrived.

Utterances such as They must have arrived/De mä vcere kommet frem thus express, not that the Speaker knows something for a fact, but, that he operates with a highly probable possible world in which their arrival has occurred. For the expression of even higher probability he would have to insert a reinforcing extra modal dement, such as definitely, undoubtedly, certainly / bestemt, garanteret, utvivlsomt. The place of necessity in a conceptual framework of possible worlds is also apparent from the fact that "possible" and "necessary" are logically complementary entities. This can be shown by investigating their meaning under negation. Saying that something is not possibly so is the same as saying that it is necessarily not so. And saying that something is possibly not so is the same as saying that is is not necessarily so. Briefly, these conceptual equivalences may be expressed as not-possible = necessarynot and possible-not = not-necessary. What the Speaker essentially does by modalizing an utterance is to qualify something which is categorical. As we have seen, this qualification will involve one of three conceptual frameworks: rational laws, social laws, natural laws. Since the scope of this book is restricted to tense and mood, which are verbal grammatical categories, the modal expressions examined below (in chapter 5) are: a) b)

those which are implemented by auxiliaries; those which are implemented by verbal inflections (the imperative and the subjunctive).

The modal use of tenses is described in chapter 7. As pointed out in 2.8, the deontic modals are excluded from the class of auxiliaries proper, but they share so many properties with auxiliaries that they are interpreted as semi-auxiliaries. For this reason these modal

46

Analysis of mood

verbs are dealt with in a supplementary chapter (8) on semi-auxiliaries and deontic modality. The grammatical approach adopted in this book obviously implies that a large number of modal expressions in English and Danish are excluded. For example, dynamic (subject oriented) modality is not dealt with at all, owing to the fact that all the dynamic modals, according to the selected criteria, are clearly non-auxiliary. More generally we have excluded all modal expressions — whether epistemic, deontic, or dynamic — which are implemented by verbs which are neither auxiliary nor semiauxiliary, by adjectival and participial constructions, and by adverbials. In English, for example, sentences of the following types will not be dealt with, in spite of the fact that they are all modal: You have to be joking. (epistemic verb) I permit you to smoke. (deontic verb) She can read Russian. (dynamic verb) She is certain to lose, (epistemic adjective) It is imperative that the job is done at once. (deontic adjective) Fm not able to attend. (dynamic adjective) He 's alleged to have resigned. (epistemic participle) You've urged not to proceed beyond this point. (deontic participle) Evidently they have slept here. (epistemic adverb) Hopefully this is enough. (dynamic adverb)

3.2. Mood in English English sentences are either categorical or modalized. In modalized sentences, modality may be expressed grammatically (whether morphologically by means of verbal inflection or syntactically by means of auxiliaries) or it may be expressed in various lexical ways (for example by füll verbs, adjectives, participles, and adverbs). Grammatically expressed modality is termed mood. If mood is expressed morphologically, we shall refer to it as synthetic. The subcategory synthetic mood has two members, namely the subjunctive and the imperative.'* Both of these are implemented by the 'inflection' - 0 (in case of the subjunctive often by be instead of is), but they can be told apart by their behaviour with respect to subjects. If mood is expressed syntactically by means of auxiliaries, we shall call it analytic (cf. Huddieston 1984: 164 ff.). The subcategory analytic mood has two members as well, namely possibility and necessity, which are

Mood in English

47

implemented by the epistemic auxiliaries mayjmight, canjcould and mustj should, need respectively. Diagrammatically, this analysis can be represented like this: Mood

synthetic

analytic

subjunctive

imperative

possibility

necessity

-0

-0

maylmight cantcould

must!should! need

Examples: Long live Trotsky. Stay as long as you like. This medicine maylmight eure your cough. This can't be true. Could it be that he is afraid? You must be hungry after your long walk. They should have arrived by now. It needn't be true. While the imperative is ahve and well in modern English, the subjunctive is largely restricted to formal style and set expressions. In the course of history most of its original functions have been taken over by other modal expressions, among these by the modal verbs. Although one does come across a mandative subjunctive, for example, in sentences such as The employees have demanded that the manager resign and People are demanding that she leave, it is more characteristic of American than of British English. In British English, where it is formal, the deontic modality involved may also be expressed by means of should or — since the modality is expressed by the verb in the superordinate clause as it is — by means of the indicative {resigns, leaves) (cf. Quirk et al. 1985: 156 f, 1014f, 1182 f). Theoretically, it would not be impossible to operate with one undivided category of mood whose members — which might be termed POSSIBIL-

48

Analysis of mood

ITY, NECESSITY, WISH, and COMPULSION - were implemented partly synthetically, partly analytically, in the same way as the category tense is implemented in both these ways. However, while in our view the analytic expressions of time have fused with the original synthetic ones into one relatively organic tense system, whose members may be analysed by means of the same semantic concepts (as described in chapter 4 below), the analytic expressions of modality have not fused with the original synthetic ones in a similar way. That it is not here the case of one organic system appears from the fact that while analytic mood — given the analysis of auxiliaries proposed in chapter 2 — has to do with epistemic modality, synthetic mood involves primarily deontic modality: the imperative is exclusively deontic {Go to bed), and the subjunctive is partly deontic {Long live Trotsky/Congress has voted that the present law be maintained), partly epistemic {If I were a Violinist, why would I write a Story about an elephant?). For this reason we have chosen to regard the category mood as subdivided into an analytic system and a synthetic system. Nevertheless, as implied by the terminology, the two subcategories are closely connected, for they are both used for the description of possible worlds, and they are mutually exclusive in the sense that a member of one of them cannot co-occur with a member of the other. This incompatibility is language specific. In Itahan, for example, it is possible for a modal auxiliary to combine with the subjunctive. This may be illustrated by a sentence like...e pensai quanto dovesse essere piacevole lavorare in quel luogo '... and I thought how pleasant it must be to work in such a place', in which epistemic dovesse — governed by the verb pensare — is past subjunctive (Bente Lihn Jensen, personal communication). Another analytical question which has to be raised is why the indicative is not recognized as a member of the category mood? This, after all, is the Standard Interpretation, and one possibility might be to operate with indicative as the unmarked (neutral) member of mood in the same way as present and non-progressive are by many regarded as unmarked members of the categories tense and aspect respectively. We have chosen not to do so for the following reasons: As already stated, we understand by mood grammatically expressed modality and by modality a qualification of an utterance whereby the Speaker operates with alternatives to the current actual world. By modalizing an utterance he superimposes notions such as possibility and necessity on it. This implies that modality — and consequently mood as well — is a non-obligatory category and that in indicative sentences such as Colonel Gaddafi is dead and The road is blocked the Speaker has chosen not to make use of it.

Mood in Danish

49

On the other hand, tense — by which we understand "grammaticalized location in time" (cf. Comrie 1985: 9) — is an obligatory category in English. Even in those cases where the present tense form is used for the expression of eternal truths, as in The sun sets in the west and Two and two make four , the Speaker temporalizes his sentences by describing a Situation which is true of the present moment (in the classroom, for example). As pointed out by Dahl (1985: 1): "in many languages the speakers are forced by the grammar to pay constant attention to time reference in order to choose correctly among the forms traditionally called 'tenses'." By using the present tense the Speaker gives the event described by the sentence in which it occurs a location in time in the same way as he does by means of the other tenses. That the present tense has a different status from the so-called indicative is apparent from examples like It mattered, it matters, it will matter and It may matter, it must matter, it matters. In the first of these, location in time is expressed not only in the first clause and in the last but also in the second, by means of the present tense form matters. In the second example, on the other hand, the third clause differs from the others in not expressing modality as understood in this book, and as modality is presupposed by mood we cannot recognize the indicative as a member of the category mood. Put briefly, the difference between the so-called indicative mood and the present tense is that the 'indicative' is never used for the expression of modahty whereas the present tense is always used for the expression of time. This is the case even when the present tense is used habitually. For example, a sentence like John goes to work at eight o'clock every day indicates that a certain habit holds at the present moment (cf. Comrie 1985: 39). It is for this reason that the category mood in English is assumed to have only four members — subjunctive, imperative, possibility, necessity — not five.

3.3. Mood in Danish In Danish, as in English, modality may be expressed grammatically (morphologically or syntactically) or it may be expressed in various lexical ways (by füll verbs, adjectives, participles, adverbs/adverbials, and modal particles such as vel, nok, etc.). The subcategory synthetic mood has the members subjunctive and imperative, which are implemented by the inflections -e and - 0 respectively. The subcategory analytic mood has four members, namely possibility, necessity, probability, and report, which are

50

Analysis of mood

implemented by the (epistemic) auxiliaries kume, mätte/behove, bürde, and skulle respectively. This analysis may be shown diagrammatically in the following way: Mood

synthetic

subjunctive -e

analytic

imperative -0

possibility necessity probability report kunne

mätte! behpve

bürde

skulle

Examples: Gud bevare Dronningen! 'God save the Queen!' Sid ned. 'Sit down.' Det kan vcere sandt. 'It may be true.' Det mä vcere sandt. 'It must be true.' Det behever ikke vare sandt. 'It needn't be true.' Det bor vcere tilstrcekkeligt. 'It ought to suffice.' Hun skal vcere en habil pianist. 'She is Said to be an able pianist.' The subjunctive is even rarer than in English, being restricted to the expression of wishes in solemn — often reUgious — language, set expressions, and curses: Komme dit rige. 'Thy Kingdom come.' ^ret vcere hans minde. 'All honour to his name.' Hun lange leve!

Mood in Danish

51

'May she live long!' Fanden tage mig! 'May the devil take me!' Krceft cede mig! 'Hell and damnation!' It is not used in noun clauses as in English (e. g. It is necessary that every member inform himself of these mies), nor for the expression of epistemic modality (e. g. Whatever be the reason, we cannot tolerate his disloyalty). The imperative, on the other hand, is used extensively as in English: Trced af 'Dismiss', Lad os gä 'Let's go', Kom ind 'Come in', etc. Theoretically, it would be possible to operate with one undivided category of mood, some of whose members are implemented by inflection (WISH, COMPULSION) and others of whose by auxiUaries (POSSIBILITY, NECESSITY, PROBABILITY, REPORT). We shall not do so for the same reason as the one given in the preceding section on mood in English. For reasons stated in the preceding section, similarly, we do not recognize a member of the category mood termed the indicative.

Chapter 4 Analysis of tense

4.1. Time and tense In many cases a linguistic category is evidently related to something found in the external world as this is reflected in the human mind (see e. g. Jespersen 1929). This goes for grammatical number, for example, which is clearly connected with a conceptual distinction between "one" and "more than one". In addition to linguistic categories there are thus conceptual categories. Human beings perceive and Interpret the external World by means of concepts like sex, quantity, time, etc., and it is therefore not surprising that such conceptual categories are reflected in language by categories like gender, number, tense, etc. A problem posed by some linguistic categories, however, is that they have no obvious conceptual counterparts. This is by some linguists assumed to be characteristic of case. For example, it is argued by Jensen that case is a purely syntactic phenomenon which has nothing to do with meaning (1985:104). It should be added, though, that other linguists are of the opinion that it refers to spatial relations (place and direction) and that a small number of spatial concepts can be isolated which are expressed in this way. Like case, mood does not have any immediately obvious extralinguistic counterpart. A good deal of thinking is in fact required before the idea of alternative worlds presents itself (cf. 3.1 above). In the case of tense we are not faced with this particular type of problem, for the corresponding extrahnguistic concept is evidently that of time. Unlike some other conceptual categories, on the other hand, time poses the special problem of being notoriously difficult to defme, and it has indeed been a challenge to philosophers ever since Aristotle first attempted to analyse it in his Physics. The difficulties involved are apparent from a dictionary definition of time as "all the years of the past, present and future" {Oxford Advanced Learner's Dictionary of Current English). This turns out to be circular, for "past", "present" and "future" are defmed as respectively "of the time before the present", "existing now", "coming after the present", and "before", "now" and "after", in their turn, are defmed as respectively "previous to the time when", "at the present time", "later in time". A more revealing definition

54

Analysis of tense

of time would be one involving change such as "progress of existence viewed as affecting persons or things". If human beings did not possess a concept of time approximately like this, any ordering of the stages in the development of a human being, for example, would be equally natural to them, and they would not be surprised "to see a certain individual first as a grown man, then as a baby, then as a corpse, then as an adolescent" (Comrie 1985: 3 f). As pointed out by Gale time has no reality independently of the changes that substances undergo, and what human beings do when they measure time is to select some cyclical physical process (seasons of the year, day and night, etc.) to serve as their dock (1968: 1,3). Initially, time may be represented as a straight line on which the past is located to the left and the future to the right of the present understood as a point with no duration:

PAST

PRESENT

FUTURE

Such a conceptualization of time must be assumed to be shared by all human beings. Even in societies that have a cyclical notion of time, the individual cycles are viewed as chronologically arranged, i. e. cyclicity is superimposed on a linear conceptualization of time (Comrie 1985: 5). All the languages of the world appear to have ways of locating events in time. Linguistic time expressions are either lexical (whether composite like "ten minutes ago" or non-composite like "tomorrow") or grammatical (whether inflectional or by means of auxiliaries). Although some linguists restrict the term "tense" to morphologically expressed location in time, i. e. to those distinctions in the forms of verbs which place the event described by a verb in time, we shall here use it more broadly to refer to any type of grammatically expressed location in time in the verb. This means that tense may also be expressed by auxiliaries or by a combination of an auxiliary and an inflection. Although most languages of the World have tense in this sense of the word, some do not have it, such as Burmese, in the same way as there are languages which lack certain other grammatical categories. Danish, for example, has no aspect, and English has practically lost gender (which only remains in part of the pronominal system). Finally, it should be mentioned that tense is a so-calied deictic category. The word "deixis" comes from a Greek word meaning "pointing" and is

Number of tenses and their implementation

55

used in linguistics to refer to those features of language which are relative to the place and time of the utterance, and which can therefore only be properly understood in relation to the "here and now" of the speech Situation. Locative adverbs such as here and there are clear examples of deixis ("at the place of the Speaker" : "not at the place of the Speaker"), and so are temporal adverbs like now and then ("at the time of speaking" : "not at the time of speaking"). Other examples are provided by the demonstrative pronouns thisjthese and that/those, which besides being spatially deictic (this book: that book) are temporally deictic {this morning: that morning). That tense is also deictic is apparent from an example like It mattered, it matters, it will matter. Here the past tense suffix expresses "prior to the time of speaking", the present tense suffix "at the time of speaking", and the future tense auxiliary "subsequent to the time of speaking". The function of tense is thus to locate an event in time relative to some other time, and as the example just given illustrates this other time is often the moment of speech. In the case of some tenses, however, the Situation is more complicated and involves other reference points than the moment of speech. This question will be taken up in section 4.3.

4.2. Number of tenses and their implementation In English as well as Danish tense is assumed to have the following eight members: present, present perfect, past, past perfect, future, future perfect, future of the past, future perfect of the past. If we disregard progressive forms, passive forms, and forms with modal auxiliaries, these members are — in regulär verbs — implemented in the following way: Enghsh Present Present Perfect Past Past Perfect Future Future Perfect Future of the Past Future Perfect of the Past

V + -j/0 hasjhave + V + -(e)d V +(e)d had + W + -(e)d will + V will + have + -(e)d would + V would + have + V + -(e)d

56

Analysis of tense

Examples: Present Present Perfect Past Past Perfect Future Future Perfect Future of the Past Future Perfect of the Past

travel(s) has/have travelled travelled had travelled will travel will have travelled would travel would have travelled

Danish Present Present Perfect Past Past Perfect Future Future Perfect Future of the Past Future Perfect of the Past

V + -er harjer + V + -(e)t V + -ede, -te havde/var+ V + -(e)t vil + V + -e vil + havejvare + V + -(e)t ville + V + -e ville + havejvcere + V + -(e)t

Examples: Present Present Perfect Past Past Perfect Future Future Perfect Future of the Past Future Perfect of the Past

rejser harjer rejst rejste havde/var rejst vil rejse vil havejvcere rejst ville rejse ville havejvcere rejst

As it appears, some tenses are formed exclusively with inflectional endings, others exclusively with auxiliaries, and others again with a combination of auxiliaries and inflectional endings. Tenses involving auxiliaries are traditionally called periphrastic. In Danish, the four perfect tenses may be formed with either of two auxiliaries: have, vcere. The rules governing the distribution of these two forms are given in 6.4 below. The following explanatory comments are called for on the Implementation of tenses as it is represented above:

Number of tenses and their implementation

57

(i) Future tense implementation in English by means of shall has been disregarded because this usage is by now quite rare. As pointed out in many works on English grammar the occurrence of shall with first person subjects to refer to (unmodalized) future is restricted to Southern Standard British English in which it is used to convey an air of formality in written language. As it appears from examples in which it is found, such as / feel I shall never get over it and We shall never be as we were, it can readily be replaced by will or 'IL Future time shall usage is connected with a strong prescriptive tradition in British English, but as noted by Quirk et al. "this prescription is old-fashioned and is nowadays widely ignored" (1985: 230) (ii) Only regulär tense forms are indicated. If irregulär verbs are taken into consideration as well, it is necessary to replace Enghsh -(e)d and Danish -ede, -te, -(e)thy abstract entities which may be manifested in a number of different ways. In descriptions of English grammar it is common practice to operate with a past tense morpheme {-ed,} and a past participle morpheme {-ed2} which besides being realizable as / d / ~ /t/~/id/ according to the regulär phonological rules may be manifested by a change of base vowel, by zero, or by irregulär suffixes, cf. examples like begin — began — begm, put — put — put, and shake — shook — shaken. The same approach can be adopted in Danish in order to include irregulär verbs such as synge — sang — sunget and hbe — leb — lobet in the description. (iii) English have, will (would) and Danish have, vcere , ville have a double status by functioning not only as auxiliaries but also as lexical verbs. The latter function may be illustrated by examples like She has the flu/Will (would) you lend me a cigarette? and Hun har influenzaj Hun er sygj Vil du läne mig en cigaret? That one phonological form may have two (or more) linguistic functions should come as no surprise; recall, for example, the discussion of epistemic and non-epistemic modals in 2.2 above. Such a double status is also characteristic of do, be in English and blive in Danish, which function not only as auxiliaries but also as lexical verbs (/ did the job, She is ill. Vi bliver hjemme). (iv) English sentences like He is retired, Mary is gone, Are you finished? and She was done with all that are with respect to tense analysed as nonperfect (unlike similar Danish sentences, e. g. Marie er gäet), for the "past participles" in constructions of this type with be are no longer fully verbal but rather of an adjectival or adverbial nature, compare, for example, gone with absent and finished with through.

58

Analysis of tense

(v) In passive sentences the eight tenses are in English implemented as V + -(e)d preceded by amlarejis (present), hasjhave been (present perfect), was/were (past), had been (past perfect), will be (future) , will have been (future perfect), would be (future of the past), and would have been (future perfect of the past). In Danish, they are implemented either periphrastically as V + -(e)t preceded by bliver (present), er blevet (present perfect), blev (past), var blevet (past perfect), vil blive (future), vil vcere blevet (future perfect), ville blive (future of the past), ville vare blevet (future perfect of the past) or inflectionally by V + -es (present), V + -edes, -tes (past). This impUes that a verb like slutte 'fmish', for example, may be passivized in ten different ways: sluttesjbliver sluttet, er blevet sluttet, sluttedes/blev sluttet, var blevet sluttet, vil blive sluttet, vil vcere blevet sluttet, ville blive sluttet, ville vcere blevet sluttet. (vi) In progressive verb phrases the EngUsh tenses are implemented as they are in passive verb phrases apart from the fact that V is followed not by -(e)d but by the suffix -ing. As progressive and passive may combine with each other there are in addition to constructions like is examining and was examining constructions like is being examined and was being examined. While progressive-passive verb phrases are of common occurrence in the present tense and the past tense, they are decidedly rare in the remaining six tenses. However, contexts do exist which may bring forth such constructions with two non-fmite forms of the auxiliary be, cf. an example like the following (Sten Vikner, personal communication): The guided tour of the cathedral, scheduled to take place this Saturday, is cancelled, as the cathedral will be being cleaned for the visit of the Pope. (vii) Finally, tenses may be implemented by verb phrases containing modal auxiliaries, as shown by examples like the following: It must happen soon. Det mä ske snart.

present

She may have been cheating. Hun kan have snydt.

present perfect

They should be home by now. past De bürde vcere hjemme nu. They might have been killed. past perfect De kunne vcere blevet drcebt.

Tense meanings

59

4.3. Tense meanings As pointed out in 4.1 tense relates the time of the event described by the verb to some other time, typically the moment of speech. If like Reichenbach (1947: 290) we symbolize event time as E and speech time as S, and if we indicate simultaneity with commas and Separation in time with dashes, the tense properties of examples like It matters, It mattered, and It will matter can be shown in the following way: Present Fast Future

E-

E,S S S—E

matters mattered will matter

In the case of the perfect tenses, the time of the event described by the verb is related not only to the point of speech but also to a second point of reference, symbolized as R. For this reason the perfect tenses have been called two-point tenses (Sorensen 1964: 74). In an example like She had signed the letter (when I returned) the event of signing takes place before this reference point (that of my return) which in its turn is anterior to the point of speech (what I say now). In She will have signed the letter (when I return) the event of signing takes place before the reference point (that of my return) and after the point of speech (what I say now). In She has signed the letter (now) the event of signing takes place before the reference point which is simultaneous with the moment of speech: Present Perfect Past Perfect Future Perfect

E —R,S E-RS S—E—R

has signed had signed will have signed

It has been objected to the analysis of the future perfect as S — E — R that although we know for sure that both speech time and event time precede reference time, it is not necessarily the case that speech time precedes event time (Vikner 1985). In a sentence like The President will have signed the treaty tomorrow, for example, the act of signing may already have been performed at the moment of speech. In that case the correct formula is E — S — R and a possible following sentence would be In fact, he has already signed it. On this basis it is argued that relations between the points in time used for the analysis of tense should be accounted for by comparing only two at a time, not three directly. For whereas the formula S — E — R, for example, does not reflect the only possible ordering of S and E in the case of the future perfect, the binary formula S — R/E — R is invariably correct. Although this argument is in principle correct, we

60

Analysis of tense

shall for the purpose of this book retain the more readily understandable analysis S —E —R. In partial support of this retention it may be argued that if the Speaker observes normal conversational rules and makes his contribution as informative as required (cf. Grice 1975, where this is termed the "maxim of quantity"), he will not use the future perfect in those cases where he knows or suspects that the event described by the verb is prior to, or simultaneous with, the moment of speech. Such a use would simply be misleading. The orderings E —S —R and S,E —R are therefore only relevant besides S — E — R in those cases where the Speaker does not know whether the event described by the future perfect verb form has taken place, is taking place, or is yet to take place but where the only temporal fact he is aware of is that it will have taken place before a certain future point in time (see Comrie 1985 : 71). In the future of the past and future perfect of the past the time of the event described by the verb is not related directly to the moment of speech but — as the terminology implies — to a time in the past which the Speaker selects as his point of departure. In the non-perfect sentence (She said) they would receive the letter the event of receiving takes place after this starting point, and in the perfect sentence (She said) they would have received the letter (by Wednesday) it takes place after the past starting point but before the point of reference (Wednesday). If we symbolize point of departure in the past as B (for basis time), the tense properties of these examples can be represented like this (cf. Maegaard et al. 1981): Future of the Past Future Perfect of the Past

B —E B—E—R

would receive would have received

As it appears from these and the above symbolic representations, future of the past and future perfect of the past are closely related to respectively future and future perfect, the only difference being that the two former tenses have a B where the two latter have an S. In this connection it should be pointed out that B and S belong to the same class in the sense that B may be assumed to replace S in the same position when future (perfect) of the past is called for in conversion of direct to indirect speech. Compare, for example, She said, "The meeting will take place tomorrow" and She said that the meeting would take place tomorrow. In the future of the past and future perfect of the past event time may precede, coincide with, or follow speech time, to which it is only indirectly related. For example, a sentence like The weatherman said it would clear

Tense meanings

61

up has the formula B —E —S in the context and so it did, B —E,S in the context and now it does, and B — S — E in the context so it probably will. All the eight tenses — in English as well as Danish — can thus be accounted for in the following way: Present Present Perfect E Past E Past Perfect E R Future Future Perfect Future of the Past Future Perfect of the Past

E,S - R,S S S S-E S- E- R B- E B- E- R

Reichenbach operates with a reference point R also in the non-perfect tenses, in each of which it is assumed to coincide with E. Although no direct evidence can be found for the existence of R in the non-perfect tenses, it can be inferred by analogy from the perfect tenses. Reichenbach notes that time adverbials correspond to R (1947: 294), as illustrated, for example, by Now I have eaten my supper (E —R,S) and Yesterday I ate my supper (E,R —S). Now if all tenses are assumed to have a reference point, they can be symbolized in the following way: Present E,R,S Present Perfect E - R,S Past E,R- S Past Perfect E-RS Future S - E,R Future Perfect S-E-R Future of the Past B - E,R Future Perfect of the Past B- E- R A Problem posed by such a representation in which the relation between points is ternary throughout is that the future tense might seem to be indeterminate with respect to S — E,R and S,R — E (and the future of the past indeterminate with respect to B — E,R and B,R — E). While the former is clearly the appropriate representation of wi/Z-sentences with future time adverbials such as The meeting will take place tomorrow, the latter would appear to be involved in sentences with present time adverbials, such as Now the negotiations will definitely be resumed. As argued by Härder (forthcoming), however, the future tense is not a mirror image of the present perfect tense (E —R,S ^ S,R —E), and it is only in the second

62

Analysis of tense

stage of semantic embedding, as it were, that the description of the future may be seen as true not only of the future but also of the present. In a sentence like Now the meeting will take place on Wednesday, for example, the primary point of reference is thus the one fixed by the future time adverbial. The appropriate formula of the future tense here and elsewhere is therefore S —E,R. On the basis of the relationship between E, R, S, and B we can now set up three semantic features: [ + PREVIOUS] serves the function of keeping the four perfect tenses apart from the four non-perfect ones. The defining characteristic of the perfect tenses is that their event time precedes their reference time. This is expressed by the notation [ + PREVIOUS]. [ ± POSTERIOR] serves the function of keeping the four future tenses apart from the four non-future ones. The defining characteristic of the future tenses is that their event time follows their speech time or — in the case of the future of the past and future perfect of the past — their basis time. This is expressed by the notation [ +POSTERIOR]. [ + THEN] (this term being used in the sense of "at that past time", i. e. = Danish da, not 50) serves the function of keeping the four past tenses apart from the four non-past ones. The definining characteristic of the past tenses is that their reference time or — in the case of the future of the past and future perfect of the past — their basis time precedes their speech time. This is expressed by the notation [+THEN]. An argument in support of the feature [ + THEN] is that it enables us to account for the so-called consecutio temporum principle, which can be observed in both English and Danish (as well as many other languages). This principle can be illustrated with the following English sentences, in which the verb of the first clause governs the choice of tense in the following clauses in such a way that we get either a sequence of [ + THEN] tenses or a sequence of [ —THEN] tenses. / had written the letter when John came and told me the news. When he has gone to the ball and the coast is clear you and I will go and search tili we find them. He said the operation would begin as soon as the liaison officer had phoned. He says the operation will begin as soon as the liaison officer has phoned. By means of these three binary features the English and Danish tense systems can be analysed in the following way:

Tense meanings

63

THEN

POSTERIOR

PREVIOUS

Present Present Perfect Past Past Perfect

— — + +

— — -

— + +

Future Future Perfect Future of the Past Future Perfect of the Past

— — + +

+ + + +

— + — +

It has been pointed out by Härder (forthcoming) that although a description of tenses with overlapping points like the one presented above is descriptively adequate, it is from a functional perspective unsatisfactory in being unduly redundant. It is simply not plausible that E,R,S, for example, reflects what the Speaker does when he selects the present tense. Härder therefore proposes an alternative model in which overlapping points are avoided altogether. In this model the starting point is the speech Situation: if the Speaker wishes a sentence to be valid for this Situation he chooses the totally unmarked present tense form. Furthermore, the choice between the present and the past is assumed to be the fundamental deictic alternative: valid for the speech Situation vs. valid for some time earlier than this Situation. A l l the remaining tenses are considered relative to the present or past tense and marked to a higher or lower degree, the future perfect of the past being maximally marked. Although this model is functionally and psychologically more adequate than the one with overlapping points it cannot be adopted here, as it does not lend itself to the contrastive description of English and Danish pursued in this book. However, the insight it provides can be captured at least partly by assigning T H E N , P O S T E R I O R , and P R E V I O U S marked and unmarked feature values. Instead of considering plus and minus equally natural, it is possible to regard one feature value as representing the expected State (unmarked) and the other as representing the unexpected State (marked). If it is assumed that the unmarked value of all three features is minus, and if "marked value" is symbolized as M and "unmarked value" as a blank, the following matrix — in which all eight tenses are kept apart — is obtained:

64

Analysis of tense

THEN Present Present Perfect Past Past Perfect Future Future Perfect Future of the Past Future Perfect of the Past

POSTERIOR

PREVIOUS M

M M

M M

M M M M M

M M

Although not identical with the markings proposed in Harder's model, the markings expressed by this matrix seem intuitively satisfactory in representing present tense as totally unmarked (cf. the Danish terms "utid", "altid") and assigning Ist degree of marking (complexity) to the past, present perfect, and future, 2nd degree of marking to the past perfect, future perfect, and future of the past, and 3rd degree of marking to the future perfect of the past. Finally it should be mentioned that the problem posed by overlapping points with respect to functional and psychological adequacy is in fact even more acute than would appear from the above discussion. As pointed out by Vikner (1985) we really need not one but two reference points for a fully adequate semantic representation of each of the eight tenses, i. e. we have to operate with four points of time, all of which coincide if the present tense is chosen. It is beyond the scope of this book, however, to enter into a discussion of this rather complicated analysis.

4.4. Eight or six tenses? In Order for a linguistic entity to be recognized as a member of a grammatical category it is usually considered a requirement that it should contrast semantically with the other members of that category. In the case of tense in English and Danish this might seem to throw doubt on the status of the future of the past and future perfect of the past. As pointed out in 4.3 these constructions are closely related to respectively the future and future perfect, which they typically replace in indirect speech (including style indirect libre). If the consecutio temporum principle is respected, there is no contrast between future and future of the past or between future perfect and future perfect of the past but complementary distribution instead, for in that case a reporting verb with the feature

Eight or six tenses?

65

specification [ —THEN] triggers off reported verbs which are also [ —THEN], for example future or future perfect, and a reporting verb with the specification [ + THEN] triggers off reported verbs which are also [+THEN], for example future of the past or future perfect of the past: Meteorologen siger, at vejret vil klare op. Meteorologen sagde, at vejret ville klare op. The weatherman says it will have cleared up by Wednesday. The weatherman said it would have cleared up by Wednesday. Consider next those cases where the consecutio temporum principle is violated — as it sometimes is in both languages — in such a way that a [ + THEN] reporting verb is permitted to be followed by reported verbs in the future or future perfect: Meteorologen sagde, at vejret vil klare op. The weatherman said that it will have cleared up by Wednesday. As these sentences are semantically quite similar to the corresponding sentences above with reported verbs in the future of the past and future perfect of the past, it looks as if there might be free Variation between the two sets of verbal constructions, i. e. once again absence of contrast. If this were so, future of the past and future perfect of the past could be regarded as variants — bound or free — of respectively future and future perfect, and an implementation rule could be set up which optionally manifests future and future perfect as respectively [ + T H E N , — PREVIOUS, + POSTERIOR] and [ + THEN, +PREVIOUS, + POSTERIOR] in indirect speech governed by reporting verbs with the feature specification [ + THEN]. On closer inspection, however, and as shown by the following examples, it turns out that there is in fact not free Variation between the two sets of verbal constructions in the case of consecutio temporum violation: The weatherman said it would clear up, and so it did. B-E,R-S *The weatherman said it will clear up, and so it did. S-E,R The weatherman said it would have cleared up by Wednesday, and so it did. B-E - R-S

66

Analysis of tense

*The weatherman said it will have cleared up by Wednesday, and so it did. S-E - R As it appears, the future of the past and future perfect of the past cannot be replaced by respectively the future and future perfect if S follows E. In these cases future of the past and future perfect of the past are simply required, and replacement of them by future and future perfect leads to ungrammatical sentences. As well-defined contexts exist in which the future of the past and future perfect of the past are obUgatory, we shall choose to recognize them as separate tenses. That an eight-tense analysis is more appropriate than a six-tense analysis is also apparent from conditional sentences, in which these verbal constructions occur outside indirect speech. While the future and future perfect are used for the expression of real conditions, as in If you ask him, he will be surprised and If Peter is in London, he will have heard about it by now, the future of the past and future perfect of the past are required for the expression of unreal and counterfactual conditions: If you asked him, he would be angry. If you had asked him, he would have been angry.

4.5. Other analytical possibilities The analysis of tense proposed here differs from most other analyses in operating with a relatively large number of tense members. In the majority of recent grammatical studies there are assumed to be only two tenses in English and Danish, namely present and past, which are implemented purely morphologically. In these studies the periphrastic constructions are assumed to implement modality in the case of will {would) and ville and aspect, or a third category, in the case of the perfect forms. As far as will and ville are concerned, it is argued, semantically, that although they provide the nearest approximation to neutral future they are always "tinged with modahty" (Leech 1987: 57), even in sentences like Perhaps he will be more peaceful when he wakes and Vi vil aldrig fä hende at se mere. As will and ville in examples of this type obviously do not express volition, the type of modality involved is assumed to be "prediction". In consequence of this analysis the modality of will {would) and ville must be subdivided into two markedly different types: volition vs. prediction. Furthermore, it is argued syntactically against recognizing

Other analytical possibilities

67

a future tense in English that will occupies a position in the modal paradigm where it blocks out occurrences of all other possible modals. It therefore seems well motivated to handle it with the other modals instead of together with the distinction of past and present time. The perfect forms are in some studies analysed as aspect (e. g. Quirk et al. 1985: 189 ff), but as different members of the same grammatical category cannot appear simultaneously in the same paradigm, and as in English any verb phrase is simultaneously perfect or non-perfect and progressive or non-progressive, this has to be an aspect of a different Order from the progressive aspect. Instead of operating with two sets of aspectual contrasts, other grammarians, for example Vestergaard (1985), prefer to account for the contrast between perfect and non-perfect forms by means of a category — different from both tense and aspect — termed phase (cf. Joos 1964: 138 ff.). In an example like The chair was pulled forward the effects of the event described by the verb are "in phase" with this event (current phase), whereas in The chair has beert pulled forward the effects are "out of phase" with the event, i. e. delayed {perfect phase). This really seems to be another way of saying that reference time coincides with event time in non-perfect constructions and is subsequent to event time in perfect constructions. The difference between event and effects of event can also be captured by saying that the function of the perfect is to describe an event at a time chosen (past, present, or future) while being true of an earlier time (Härder, forthcoming). Now if besides tense one has at one's disposal both phase and a concept of predictive modality, the eight verbal constructions we have analysed as tenses can be interpreted alternatively in the following way:

travel(s) has/have travelled travelled had travelled will travel will have travelled would travel would have travelled

TENSE

PHASE

present present past past present present past past

current perfect current perfect current perfect current perfect

MODALITY — — — —

predictive predictive predictive predictive

The analyst preferring to operate with aspect need only replace PHASE, current, and perfect by respectively ASPECT, non-perfective, and perfective in order to account for the eight constructions.

68

Analysis of tense

Let US now briefly explain why we have not adopted a simple twotense analysis where we operate with the past and the present only. There are two main reasons for rejecting the modality Interpretation (see Davidsen-Nielsen 1988 for a detailed discussion of this question): (i) Utterances with non-volitional will and ville differ from utterances with evident modals in being categorical. For example, Bill will finish his novel differs semantically from Bill may finish his novel in describing the actual future — not an alternative world in which the novel is fmished. (ii) Non-volitional will and ville are kept clearly apart from volitional (modal) will and ville by their syntactic behaviour: unlike the formally identical volitional modals they do not occur in conditional and temporal clauses, whereas they do occur before have + \ + -ed (English) and havejvcere + V + -(e)f. *If/when the meeting will take place, we can sign the treaty. *Hvis/när medet vil finde sted, kan vi underskrive traktaten. The guests will have arrived by that time. Gasterne vil vare ankommet til den tid. In English, furthermore, non-volitional will differs from the formally identical volitional modal in occurring in passivized sentences and before be + N + -ing: I will be brought back in disgrace. We will be flying at 20,000feet. In Danish, fmally, non-volitional ville is separated distributionally from the volitional modal ville by occurring in passive sentences formed with blive (cf. 2.2): Jeg vil blive fort vanaret tilbage. As far as the perfect forms are concerned, we have chosen to reject the aspectual interpretation because these forms express something which has nothing to do with aspect as usually understood. Aspect serves the function of presenting the event described by the verb either with or without reference to its internal temporal constituency. In Miss Smith was reading the paper and Miss Smith read the paper, for instance, the event of reading is presented respectively with and without reference to its internal temporal constituency. In the non-progressive it "is presented as a Single unanalysable whole, with beginning, middle, and end rolled into one" (Comrie 1976: 3). In the progressive example, on the other hand, the event is unfolded and looked at intemally. Now in both English

Tense in

fiction

69

and Danish, perfect vs. non-perfect constructions teil us nothing about internal event time, i. e. about the extension or temporariness of the verbal event. What they do teil us about is external event time, i. e. about the relative placing of an event described by a verb in a chronological Order. Since the perfect forms are in this respect similar to the evident tenses (past and present), the category tense need not be unduly stretched in Order to accommodate them as tenses. It is only by the fact that their reference time does not coincide with their event time that they might be considered "deviant". The analysis of the perfect constructions by means of phase is not open to objection in the same way as the aspectual analysis of them. Although it is probably wise to avoid proliferation of categories, the fact that reference time is in these constructions separated from — and posterior to — event time may be considered important enough to Warrant the introduction of a separate grammatical category. For the contrastive purposes of this book, however, an analysis based on phase (or perfective aspect) has the unfortunate consequence that the present perfect and the simple past, which are often in dose competition in both languages, and which are distributed somewhat differently in English and Danish, are represented as non-minimally opposed to each other, differing with respect to two grammatical categories. If the above discussion of how tense may be analysed has seemed fairly complex, this is due to the nature of the languages involved. Any seriously interested Student of the verb in English and Danish has to face the fact that tense, aspect, and modality are here tightly interwoven and very hard to separate.

4.6. Tense in fiction The above discussion of tense applies to what has been called the normal, referring mode of communication, in which events are conceived of as real, and in which the use of tense is governed by deictic principles: prior to, simultaneous with, or posterior to the moment of speech. Before concluding this chapter we should point out that in the fictional mode of communication characteristic of narration (novels, short stories, storytelling, etc.) other principles are at work (see Bache 1986). Since Danish students of English acquire their implicit knowledge of English grammar, including tense rules, largely by reading fictional texts, this difference is Worth noticing also from a didactic point of view. In narration, the past

70

Analysis of tense

tense is not used for the purpose of locating events as prior to the moment of communication (writing, speaking). That this is so is apparent from the following narrative passage: The Count suddenly leapt up from his quiet radio set and ran out into the kitchen. He took down from the wall Tim's picture entitled Three Blackbirds in a Treacle Well. He stared at it. It was a horrible picture. He was about to smash it into the rubbish bin when some unseen hand prevented this exhibition of blind rage. His heart was beating violently. (Iris Murdoch, Nuns and Soldiers) To the extent that temporality is involved in narrative texts it typically involves the feature [PREVIOUS]: the perfect tenses are used to locate events as prior to the remaining narrated events, which are described by means of non-perfect verb forms: The Count was sitting beside his radio set which he had just switched o f f . It was late at night. He had listened to a symphony concert.... Now all was silent. (Iris Murdoch, ibid.) Temporal ordering may also, however, involve the feature [ +POSTERIOR], as illustrated by the following passage: He went into the bedroom and opening a deep drawer thrust the picture away into the bottom of it. As he did so his hand came into contact with a soft roll of stuff which he recognized as being the Pohsh flag.... Ready to weep he prepared himself for bed. He would not sleep. (Iris Murdoch, ibid.) What seems to be typical of the fictional mode of communication, then, is that it is the Opposition between [+THEN] and [ — THEN] — by some linguists regarded as the fundamental deictic alternative (cf. 4.3) — which is eliminated. In narration it is unmarked for events to be described in the past tense, but for stylistic purposes they may also be described by present tense forms, not only in occasional sentences in order to create a dramatic effect (cf. the discussion of the so-called historic, or dramatic, present in 6.3 below) but also in longer stretches of fiction: What is happening upstairs is something that Howard will hear about only the following day. A window smashes in one of the

Tense in

fiction

71

small bedrooms, along a corridor where silence reigns; the cause is Henry Beamish, who has put bis left arm through and down, and slashed it savagely on the glass. Only a few people bear tbis, and most are beavily occupied; but someone is curious enougb to look into tbe little bedroom wbere be is, and see bim, and lift bim from tbe debris around bim, and call otbers. (Malcolm Bradbury, The History Man) As it appears from tbis extract, the present perfect {has put) and tbe future {will hear) are used to describe events as respectively prior and subsequent to tbe other narrated events. Tbe remaining, present tense forms, on the other band, represent a stylistic choice, and replacement of them by past tense forms would not affect tbe conceptual meaning of tbe quoted passage.

Chapter 5 Mood usage

5.1. Possibility By modalizing an utterance epistemically the Speaker indicates that he is not presenting what he is saying as a fact but rather, for example, that he is speculating about it, presenting it as a deduction, or that he has been told about it (Palmer 1986: 51), cf. examples like Han kanimälskal vare i London for ojeblikket 'He may/must/is said to be in London at the moment'. While in the case of epistemic necessity he expresses a high degree of confidence in the truth of what he is saying on the basis of deduction from facts, he indicates no more than that a certain State of affairs is conceivably real in the case of epistemic possibility. For this reason necessity and possibility might also be termed respectively deductive and speculative (ibid.: 60). Declarative Sentences In English, possibility is implemented by may in positive declarative sentences, as illustrated by the following examples: That may be the best light III ever appear in, to them. It's probably some weird piece of pure spite, something we may never mderstand at all. She may be irresponsible, a bit dotty. I mean, she may still change her mind. I hope that Gertrude may find a better husband. I really know so little about you and about what you may be feeling now. She may have feit possessive about Gertrude. In Danish, possibility is realized by the non-cognate auxiliary kunne: De ca. 33,000 unge, der har sogt ind pä landets universiteter, kan komme til at vente lange pä besked. 'The about 33,000 young people who have applied for admission to the universities of this country may have to wait a long time for a decision.'

74

Mood

usage

Der kommer nogen sol. Der kan ogsä komme en enkelt byge. 'There will be some sun. There may also be an occasional shower.' Om natten kan temperaturen komme ned pä omkring 10 grader. 'At night the temperature may drop to about ten degrees centigrade.' Meteorologisk Institut forudser dog, at solbaderne i de kommende dage af og til kan blive forstyrret af en mindre tordenbyge. 'The Met Office predicts, however, that in the next few days sunbathers may be disturbed now and then by a minor thundery shower.' Vi vil meget grundigt gennemarbejde CO Iis udspil, og se hvilke elementer og hvilke synspunkter i udspillet, der eventuelt kan bidrage til at pävirke vores forslag. 'We will study CO II's move very carefully and see what elements and what views in this move may possibly contribute to influencing our proposal.' Det kan mäske ogsä komme pä tale at lade bilister med mindst en passager bruge bus-sporet. 'Another approach which may perhaps be adopted is letting motorists with at least one passenger use the bus lane.' Excursus Not infrequently one comes across a can in positive declarative sentences which does not express (deontic) permission and which it would be somewhat strained to analyse as (dynamic) ability. As illustrated by the following examples, quoted from van der Auwera 1986; Vestergaard 1985; and Palmer 1979, it expresses a type of possibility which seems related to that expressed by epistemic may: Vampires can kill. Dogs can be snappish. The pound can be devalued. The road can be blocked. It can be cold in Stockholm. Who knows? It can go either way. The type of possibility indicated by can in examples like these has been characterized as theoretical (Vestergaard 1985: 65), neutral/circumstantial (Palmer 1979: 71 f), and contingent (van der Auwera 1986). It can be paraphrased as "It is possible for", and this can therefore differs semantically from epistemic may, for which the appropriate paraphrase is "It

Possibility

75

is possible that", or "perhaps". For example, there is a clear semantic difference between The pound can be devalued, The road can be blocked, It can be cold in Stockholm and The pound may be devalued, The road may be blocked, It may be cold in Stockholm, only the last three of which sentences are synonymous with Perhaps the pound will be devalued. Perhaps the road is blocked. Perhaps it is cold in Stockholm. In the examples Vampires can kill and Dogs can be snappish, to be sure, replacement of can by may does not result in such a change of meaning. This, however, is due to the fact that may does not in this type of generic context indicate epistemic possibihty ("perhaps") but theoretical possibility like can: in formal English may can be used synonymously with can to indicate "It is possible for" (see Quirk et al. 1985: 223). Finally, It can go either way differs semantically from It may go either way in that the former sentence describes a Situation arising from the nature of things (i. e. possible for) and the latter a Situation arising from the actual present State of affairs (i. e. possible that). As pointed out in 3.1, dynamic possibihty has to do with the relationship which exists between circumstances and unactualized events in accordance with natural laws. As it is this type of relationship which seems to be involved in examples with can like the ones given above, we shall conclude that the type of modality they express is not epistemic (rational laws) but dynamic (natural laws). Note in this connection the frequency of generic contexts in examples with theoretical can-, and recall that epistemic verbs in English are not normally permitted in conditional clauses whereas they combine freely with a following perfect infmitive with past time meaning and with progressive aspect (3.1): * If the road may be blocked ... The pound may have been devalued. It may be going either way. With respect to these distributional properties can used for the expression of theoretical possibility is non-epistemic: If the road can be blocked ... * The pound can have been devalued. * It can be going either way.

76

Mood usage

In Danish, theoretical possibility is expressed by kunne, and the English use of can for this purpose is therefore unlikely to create difficulty for Danish learners (unlike epistemic may)\ Hunde kan ycere bidske. 'Dogs can be snappish.' Pundet kan devalueres. 'The pound can be devalued.' Vejen kan afspcerres. 'The road can be blocked.' Det kan vasre koldt i Stockholm. 'It can be cold in Stockholm.' The use of kunne for the expression of both epistemic and theoretical (dynamic) possibility is a potential source of ambiguity. This is apparent from the last example, which corresponds not only to It can be cold in Stockholm but also to It may be cold in Stockholm: Det kan vare koldt i Stockholm. Sommetider faider temperaturen til minus 30 grader. 'It can be cold in Stockholm. Sometimes the temperature drops to minus 30 degrees centigrade.' Det kan \cere koldt i Stockholm, när du ankommer, sä du mä hellere tage varmt toj pä. 'It may be cold in Stockholm when you arrive so you'd better put on some warm clothes.' In other cases, however, there is no ambiguity in spite of the fact that the same verb is used for the expression of both types of possibility. As pointed out in 2.2, epistemic modals (including kunne) require the blive passive whereas non-epistemic modals (including dynamic kunne) require the -s passive. This means that The pound may be devalued and The pound can be devalued when translated into Danish are kept apart in the following way: Pundet kan blive devalueret. Pundet kan devalueres. The distinction between epistemic and theoretical possibility can also be expressed by a formal difference between kunne + vare + V + -(e)t and kunne + V + -es\

Possibility

77

Vejen kan vcere afsparret. 'The road may be blocked.' Vejen kan afspcerres. 'The road can be blocked.' It appears, then, that when used for the expression of theoretical possibility kunne is distributionally a non-epistemic verb like can used for this purpose in English. This is also evident from the fact that it is permitted in conditional clauses (Hvis vejen kan afspcerres ...) and that it does not combine with a following perfect Infinitive with past time meaning or with lexical verb phrases like vcere i fcerd med or vare ved at 'be in the process o f . Can and kunne used for the expression of theoretical possibility often indicate what is characteristic of somebody or something (dispositional). This is illustrated by some of the examples given above as well as by the following: Children can be very trying. The Bay of Biscay can be quite rough. When I first knew her she could be very catty. Tore var en af de fyre...der kan rejse sig op midt i timen og gore en masse ting...han kunne bruge en hei time til at lede efter en eller anden dodssyg blyant nede pä gulvet. 'Tore was one of those types...who would get up in the middle of a lesson and do all sorts of things...he could spend an entire lesson searching for some silly pencil on the floor.' Interrogative sentences Although the occurrence of epistemic may in interrogative sentences cannot be ruled out completely (cf. an example like May we be doing him an injustice? given by Quirk et al. 1985: 815), possibility is virtually always implemented as can in this sentence type, as illustrated by the following examples: 'Oh, my dear Anne,' said the Count, 'can it be true?' Can spring be far behind? Hut can she be right? Their accounts raised a pointed question: Can North be believed? Can it be that this new assistant, so assiduous and reliable, actually enjoys his labours?

78

Mood usage

From Chicago, a Ph.D. out of Bloomington, can I be as backward as I seem? Besides yes-no questions this implementation usually also characterizes w/j-questions: How can this be irrelevance? Where can they have got to? Whose beautiful antiques can these be? What can she mean? In Danish, possibility is implemented as kunne in questions as it is in other sentence types: Kan det virkelig vcere sandt? 'Can it really be true?' Men kan der vcere sne pä fjeldet ijuli? 'But can there be snow in the mountains in July?' Kan Situationen ikke blive ubehagelig for den gruppe der er pä vippen?

'Can't this Situation become unpleasant for the group which is threatened?' Kan denne gentagelse vare tilfceldig? 'I wonder if this repetition is accidental?' Hvis smukke antikviteter kan det mon vcere? 'Whose beautiful antiques can these be?' I hvilke tilfcelde kan en sädan gentagelse tcenkes? 'In what cases is such a repetition conceivable?' Negative sentences In English, possibility is implemented as can if it is the modality itself which is negated and as may if it is the following proposition which is negated: It can't be true. (It is not possible/that it is true.) It may not be true. (It is possible/that it is not true.) In Danish, possibility is implemented as kunne if it is the modality which is negated: Det kan ikke vcere sandt. (It is not possible/that it is true.) While Danish thus expresses "not possibly" in the same way as English (kunnejcan), it has no possibility modal for the expression of "possibly

Possibility

79

not" (corresponding with may). Since "possibly not" is logically the equivalent of "not necessarily", however, this meaning can be expressed by means of the auxiliary beheve: Det behover ikke vcere sandt. (It is not necessarily true = It is possibly not true.) In English, similarly, the necessity modal med can be used for this purpose: It needn't be true. Though this is logically the equivalent of It may not be true, the two sentences are not entirely synonymous, for there is a difference between making a judgment in terms of necessity and making it in terms of possibility. That epistemic may not and needn't are not always interchangeable is apparent from an example like the following (Palmer 1979: 54): He may be there, but he needn 't be. In the majority of cases, however, it makes relatively little difference if the Speaker selects may not or needn't, and this implies that Danish learners can usually get by with behove transferred to English in the shape of need if they have difficulty with the unfamiliar use of may in negative sentences. Another near equivalent in Danish of may not is mäske ikke, compare examples like It may not be true and Mäske er det ikke sandt. The Implementation of possibility and necessity in negative sentences in English and Danish can be summarized like this: not-possible = necessary-not not-necessary = possible-not

English

Danish

can may/need

kunne behove

The use of can and kunne for the expression of "not possibly" can be further illustrated by the following examples: They can 't have gone very far. Waldheim can 't have been unaware of the 1944 deportation of 40,000 Greek Jews to death camps. You can't be serious. Gentagelsen kan ikke vare tilfceldig, er de fleste politiske iagttagere enige om. 'The repetition can't be accidental, most political observers agree.'

80

Mood

usage

Waldheim kan ikke have vceret uvidende om deportationerne. Du kan ikke mene det alvorligt. 'You can't be serious.' The following examples provide further illustration of may in English used for the expression of "possibly not": Of course she may not have been lying on all these points. It may not matter in the least. You may not be happy, but at least you can quietly hide. After all, Gertrude may not be there at all. The use of need and behove for the expression of "not necessarily" ( = "possibly not") is exemphfied in 5.2 below. Before leaving negative sentences we should point out that epistemic can occurs in sentences which are not strictly speaking negative in the sense that not follows the auxihary, but which contain alternative negative forms: There can be no doubt about that. There can hardly be any doubt about that. Sentences in the past tense In English, possibility is implemented by the past tense forms might and could if the Speaker needs to indicate that he is relatively uncertain about the possibility of the State of affairs described by the utterance. This use of might and could to convey a "tentative" attitude (cf. Palmer 1979: 30) can be illustrated by the following examples: She might take it into her head to come round and bang at the door. That woman might attack you. It might even help both of them. What you suggest might be regarded as a recipe for folly and madness. He's not much here but he could arrive. Your act could make another person despair. But could she be right? In examples like these the past tense modals refer to non-past time and express weakpossibility ("It cannot be ruled out that ...")• As it appears, could is not restricted to interrogative and negative sentences like epistemic can. In negative sentences, could and might differ from each other with respect to the target of the negation like can and may:

Possibility

81

They couldn't be home. (not possibly) They mightn't be home. (possibly not) Might and could are also used without past time meaning in hypothetical conditional sentences: If I was a different person, it mightn't matter. If we instructed him carefully, Jones could be the right man for the Job. In indirect speech (including style indirect libre), it should be noted, the past tense forms do not usually express tentativeness — although it is possible for them to do so — but unweakened possibility, for in this type of context they normally represent backshifted may or can: She Said the time might come in my generation...when the educational system itself was mixed. {^—The time may come ...) I was wondering if it could have been fear. (f—Can it have been fear?)

In Danish, the Situation is basically the same as in English, apart from the fact that only one past tense modal is involved, namely kunne. The following sentences illustrate the use of this form for the expression of non-past weak possibility: Det kunne tcsnkes. 'It is conceivable.' Et mode med hojtstäende militcerpersoner kunne käste lys over den mystiske sag. 'A meeting with high-ranking military men might shed light on the mysterious case.' Udvalget har bevilget ham delvis immunitet. Dermed bliver han antagelig fri for spergsmäl, hvis besvarelse kunne bringe ham i fedtefadet. 'The committee has granted him partial immunity. In that way he'll probably be able to avoid questions the answers to which might get him into hot water.' En sädan omtale kunne Icegge hele boligsektoren for had. 'Such Publicity might cast odium on the entire housing sector.' En missil-aftale kunne danne forste trin i en ny afspcendingsproces med Sovjet.

82

Mood usage

'A missile agreement might be the first step towards a new process of detente with the Soviel Union.' The use of kunne in hypothetical conditional clauses can be exemplified by the following sentence: Hvis Jensen blev forberedt grundigt, kunne han vcere den rigtige mand til denne stilling. 'If we instructed him carefully, Jensen could be the right man for this job.' In indirect speech the past tense form normally represents a backshifted non-tentative modal, as in English: Han udtalte frygt for, at konßikten kunne brede sig. kan brede sig.)

Konflikten

'He expressed his fears that the conflict might spread.' Summary The implementation of possibihty in Enghsh and Danish by present tense auxiüaries can be summarized and exemplified in the following way: negative declarative interrogative modal propositional negation negation English Danish

may kan

can kan

can kan

may (need) (behover)

Boris may have committed this crime. Boris kan have begäet denne forbrydelse. Can Boris have committed this crime? Kan Boris have begäet denne forbrydelse? Boris can't have committed this crime. Boris kan ikke have begäet denne forbrydelse. Boris may not have committed this crime. (Boris needn 't have committed this crime.) {Boris behover ikke have begäet denne forbrydelse.) For weak possibility (tentativeness) the pattern of implementation in the different sentence types looks like this:

Necessity

English Danish

declarative

interrogative

negative modal propositional negation negation

might/could kuntie

could kume

could kunne

83

might

Boris mightIcould have committed this crime. Boris kunne have begäet denne forbrydelse. Could Boris have committed this crime? Kunne Boris have begäet denne forbrydelse? Boris couldn't have committed this crime. Boris kunne ikke have begäet denne forbrydelse. Boris might not have committed this crime. As it appears, there is in Danish no grammatical way in which to indicate weak possibility in sentences with propositional negation. The past tense form behovede — the natural candidate — does not seem to be used for the expression of tentative epistemic modality. In an example like Boris behevede ikke (at) have begäet denne forbrydelse, for instance, this form can hardly be assumed to express anything but deontic modality with past time meaning (recall that deontic skulle and bürde permit a following perfect infinitive (2.2.)), and sentences of the type Det behevede ikke vcere sandt/tilfceldet can scarcely be used epistemically with present time meaning. In Danish, tentativeness in sentences with propositional negation thus has to be expressed lexically (e. g. Det kunne tcenkes, at Boris ikke har begäet denne forbrydelse). This means that the Danish learner is faced with a construction in English whose unfamiliarity is due not only to the fact that it involves a non-corresponding verb form — like may and might in positive declarative sentences — but to the fact that it is grammatical.

5.2. Necessity Declarative sentences In English, necessity is implemented by must in positive declarative sentences. In this way the Speaker indicates that by inference he is confident that a certain State of affairs obtains: But you must be tired. I expect she hates me, why shouldn 't she, she must be sore as hell.

84

Mood usage

Oh all right, maybe you didn't write the letter, but you must have told all those foul lies to Anne Cavidge. You must have made her think we 'd been together. But he did feel that every intelligent person must be interested in Poland. He feit that it must be untrue. He must have run back instinctively the way he came, in order to return to the village. It must be the sadist in you taking over. In Danish, this member of analytic mood is implemented by the formally corresponding auxiliary mätte: IFK har jo lige mistet Glen Hysen, sä vi mä vcere stcerkest i ojeblikket. 'IFK have just lost Glen Hysen, you know, so at the moment we must be the stronger side.' Det mä vcere sket med henvisning til paragraf 6 om, at kommunalbestyrelsen kan give äbningstilladelse ved scerlige turistbesog for varer som souvenirs, antikviteter o.l. 'It must have happened with reference to regulation 6 that the local council may permit the sale of goods such as souvenirs, antiques, etc. after hours in the case of special tourist visits.' Mildere straffe mä vcere en medvirkende ärsag til, at anfallet af forbrydelser i 1986 rundede den halve million. 'More lenient sentences must be a contributory cause why in 1986 the number of crimes passed half a million.' Nedgangen i antallet af trafikdrabte mä skyldes den generelle hastighedsnedscettelse fra 60 kmjt til 50 kmjt, mener politiet. 'The reduced number of traffic casualties must be due to the general reduction of the speed limit from 60 to 50 kilometers an hour, the poHce believe.' Han mä vcere blevet pensioneret fornylig. 'He must have retired recently.' Der mä vcere flere andre gründe. 'There must be several other reasons.' Interrogative sentences As pointed out by Palmer "We seldom question epistemic modality" (1979: 56), and examples in which necessity is implemented in interrogative sentences are indeed quite rare. In yes-no questions one occasionally

Necessity

85

finds epistemic must, but only in those cases where a negative orientation has to be conveyed. For example, a Speaker may pose a question like Must he be on holiday? if he challenges the claim made by bis interlocutor that the person referred to is on holiday. Otherwise, necessity is implemented by need in yes-no questions (particularly frequently in challenging echo questions): Need this suggestion have any party political Need he be on holiday? Need it have happened that way?

implications?

In Danish, necessity is implemented by behove in this sentence type: Behover dette forslag have nogen partipolitiske Behover han vcere pä ferie? Behover det vare sket pä den mäde?

implikationer?

In wA-questions, necessity is normally implemented by must and mätte-. What must it have feit like? What must it have been like in the Middle Ages, I wonder? Hvad mä han have folt ved dette? Hvordan mä det mon have vceret at leve i Middelalderen? This Implementation is also the typical one in those cases where a whquestion represents a reaction to a preceding Statement with epistemic mustjmätte. If we consider the sentences He must have strangled his wife noiselessly in the kitchen last night, Han mä have kvalt sin kone lydlest i kokkenet i aftes, we see that in all the possible conversions of these to wA-questions must and mätte are retained: Who must have ...?IHvem mä have ...? Whom must he have ...?IHvem mä han have ...? How must he have ...?/Hvordan mä han have...? Where must he have ...?/Hvor mä han have ...? When must he have ...?IHvornär mä han have ...? In contexts like these need and behove are hardly possible as alternative forms. In others, however, they do occur besides must and mätte, particularly, it seems, in questions with whyjhvorfor: Why need/must John have killed his wife? Hvorfor behoverjmä John have släet sin kone ihjel?

86

Mood

usage

In cases of this type a fairly subtle semantic difference appears to be involved. By using needjbehove the Speaker challenges the deduction made by his interlocutor, i. e. a possible paraphrase might be "What makes you think you are right in inferring that ...?" By using must or matte, on the other hand, he asks his interlocutor to explain for what reason John is assumed to have killed his wife, i. e. a possible paraphrase could be "On what grounds do you infer that ...?" That this is not a far-fetched analysis is apparent from the fact that the adverbials at all and overhovedet can be added to the sentences with need and behove but not to the sentences with must and matte: Why need he have killed his wife at all? Hvorfor behever han overhovedet have drcebt sin kone? * Why must he have killed his wife at all? * Hvorfor mä han overhovedet have drcebt sin kone? Notice also that need and beheve, in the context we are discussing, differ semantically from must and mätte in being semantically very similar to should and skulle: Why should he have killed his wife (at all)? Hvorfor skulle han (overhovedet) have drcebt sin kone? Negative sentences In English, necessity is implemented as need if it is the modality which has to be negated ("not necessarily"): It needn't be true. If they are right, Pinturischio need not necessarily be the one to oblige. It need not affect the incidence of local taxation. This certainly could happen, but it need not happen. Survival öfter death and unending improvement need not mean perfect happiness. If this is the case, then it need not require more than time in the Programme.

Like can (cf. 5.1), epistemic need occurs also in sentences which are not strictly speaking negative, but which contain alternative negative forms:

Necessity

87

It need never affect the incidence of local taxation. It need have no effect on the incidence of local taxation. In Danish, similarly, necessity is implemented as behove in those cases where the modality is negated: Selv om han kalter lidt, behover han ikke vcere alvorligt skadet. 'Although he's limping somewhat he needn't be seriously injured.' Det behover ikke have vceret en fejltagelse. 'It needn't have been a mistake.' Det behover ikke vcere det, der gor sig gceldende. 'It needn't be that which manifests itself.' Selv om der skrives boger om det, behover det ikke vare vcerd at forstä. 'Although books are written on it, it needn't be worth understanding.' For the expression of "necessarily not" neither EngUsh nor Danish has any necessity modal available but, as pointed out in 5.1, this meaning can be expressed by means of the auxiliaries can and kunne: It can't be true. Det kan ikke vcere sandt. (It is not possibly true = It is necessarily not true) Sentences in the past tense Necessity is implemented by the past tense forms should and skulle if the Speaker expects the State of affairs described by an utterance of his to be real but does not feel absolutely certain, for example because the facts upon which his deduction is made may not be complete. According to Palmer epistemic should expresses extreme likelihood but implicitly allows for the Speaker to be mistaken (1979: 49), and this is true of skulle in Danish as well. The use of these past tense forms to indicate such a tentative certainty can be illustrated by the following examples: They should have reached their destination by now. This region should be just above the great face. You should be seeing my family tomorrow. It shouldn 't be difßcult to accomplish that. De skulle allerede vcere undervejs. 'They should be on their way already.'

88

Mood usage

Det hele skulle vcere fardigt i trykkeriet i morgen formiddag. 'All of it should be ready in the printing office tomorrow morning.' In examples like these the past tense forms refer to non-past time (in English, in fact, should never refers to past time) and express weakened necessity, which is the equivalent of high but not maximum probability. The fact that should and skulle indicate tentative inference readily explains their common occurrence in conditional sentences. In examples like the ones given above the addition of non-hypothetical conditional clauses would be very natural, e. g. If I remember right, this region should be just above the great face. Medmindre der sker uforudsete ting, skulle det hele vcere fcerdigt i trykkeriet i morgen formiddag. 'Unless something unexpected happens, all of it should be ready in the printing office tomorrow morning.' The following sentences give further Illustration of the use of shouldj skulle in conditional sentences: If you press that button, the engine should start. Hvis du trykker pä den knap, skulle motoren starte. The "provided that" element connected with these verb forms may even involve "double conditionality": Unless there's something wrong, the engine should Start if you press that button.

Medmindre der er noget galt, skulle motoren starte hvis du trykker pä den knap. Epistemic should and skulle with future time meaning are also of common occurrence in the subordinate clause of conditional sentences: Should you happen to be passing, do drop in. Skulle du komme forbi, mä du endelig kigge ind. If a crisis should arise, the public will have to be informed. Hvis en krise skulle opstä, mä offentligheden informeres. In English a special use is found of epistemic should in examples like these: Well, it surprises me that Eiken should be surprised. I can 't think why he should have been so angry.

Necessity

89

According to Palmer, who quotes the first of these examples (1979: 161), the modal is here semantically empty, whereas Quirk et al., who give the second (1985: 234), are of the opinion that it expresses "putative" meaning. Since "putative" means "commonly reputed to be", it would seem that this use of should is related to that of Danish skulle for the expression of "report" (to be discussed in 5.4). In Danish, a sentence like Det overrasker mig, at Eileen skulle vcere overrasket is clearly well-formed, and here the modal indicates "report" or "rumour". As shown by Jacobsson, however, the variety of should classified by Quirk et al. as putative does in fact sometimes convey factuality, cf. an example like I was...stupefied...that we should meet ort this beach (1988: 74); in this way it is clearly different from skulle of report in Danish. As pointed out in 5.1, could, might, and kunne sometimes represent a backshifted present tense epistemic modal (can, may, kan) in indirect Speech. Since shall and skal do not function as epistemic necessity modals, however, this use is ruled out in the case of should/skulle. On the other hand Danish matte frequently represents a backshifted epistemic mä in indirect speech: Han sagde, at de matte have reget hash inden forestillingen. (-(—De mä have ...) 'He said they must have been smoking pot before the Performance.' (