You are on page 1of 240

Verbs of Implicit Negation and their Complements

in the History of English


Verbs of Implicit Negation
and their Complements
in the History of English

Yoko Iyeiri
Kyoto University

John Benjamins Publishing Company


Amsterdam / Philadelphia
TM The paper used in this publication meets the minimum requirements of
8

American National Standard for Information Sciences – Permanence of


Paper for Printed Library Materials, ansi z39.48-1984.

Library of Congress Cataloging-in-Publication Data

Iyeiri, Yoko.
Verbs of implicit negation and their complements in the history of English / Yoko Iyeiri.
p. cm.
Includes bibliographical references and index.
1. English language--Verb. 2. English language--Verb phrase. 3. English language--Grammar,
historical. 4. English language--History. I. Title.
PE1319.I94   2010
425'.6--dc22 2010002767
isbn 978 90 272 1170 5 (John Benjamins B.V.) / isbn 978-4-8419-0540-3 (Yushodo Press)
(Hb; alk. paper)

© 2010 – John Benjamins B.V. / Yushodo Press


No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any
other means, without written permission from the publisher.
John Benjamins Publishing Co. · P.O. Box 36224 · 1020 me Amsterdam · The Netherlands
John Benjamins North America · P.O. Box 27519 · Philadelphia pa 19118-0519 · usa
Yushodo Press Co., Ltd. 3–42–3 Ohtsuka Bunkyo-ku Tokyo · Japan
Table of contents

Preface and acknowledgements ix


Abbreviations xi
List of tables and figures xiii

1. Introduction
1.1. Verbs of implicit negation and the aim of the present book 1
1.2. The Great Complement Shift 5
1.3. Some previous studies and the present research 8
1.4. Meaning and form 11
1.5. Data 19
1.6. Overview of the present book 23

2. Verbs of implicit negation and to-infinitives


2.1. Verbs of implicit negation and to-infinitives 27
2.2. The historical development of forbid 28
2.2.1. Preliminary remarks 28
2.2.2. The decline of that-clauses after the verb forbid 31
2.2.3. God forbid 35
2.3. The historical development of refuse 43
2.3.1. Preliminary remarks 43
2.3.2. That-clauses and/or to-infinitives as the complement of refuse 44
2.3.3. To-infinitives and/or gerunds as the complement of refuse 45
vi Table of contents

3. Verbs of implicit negation and gerunds


3.1. Verbs of implicit negation and gerunds 51
3.2. The historical development of forbear 52
3.2.1. Preliminary remarks 52
3.2.2. The historical rise of to-infinitives and gerunds with forbear 55
3.2.3. The preposition from and forbear in current English 59
3.3. The historical development of avoid 64
3.3.1. Preliminary remarks 64
3.3.2. The syntactic development of avoid in the history of English 66
3.3.3. The so-called horror aequi principle 71

4. Verbs of implicit negation and gerunds with prepositions


4.1. Verbs of implicit negation and gerunds with prepositions 79
4.2. The historical development of prohibit 80
4.2.1. Preliminary remarks 80
4.2.2. The rise of to-infinitives and gerunds 82
4.2.3. Gerunds with or without the preposition from 85
4.3. The historical development of prevent 87
4.3.1. Preliminary remarks 87
4.3.2. The establishment of gerundial constructions in the history of
English 89
4.3.3. Three types of constructions with gerunds in later Modern
English 91
4.3.4. Is “prevent + object + –ing” an elliptical form of “prevent +
object + from + –ing”? 97
4.4. The historical development of hinder 99
4.4.1. Preliminary remarks 99
4.4.2. The rise of gerunds and the decline of the other constructions 102
4.4.3. The contrast between nominal and pronominal objects 106
4.5. The historical development of refrain 112
4.5.1. Preliminary remarks 112
4.5.2. The rise of to-infinitives and gerunds 113
4.5.3. Gerunds with or without the preposition from 115
Table of contents vii

5. Verbs of implicit negation and subordinate clauses


5.1. Verbs of implicit negation and subordinate clauses 119
5.2. The historical development of fear 120
5.2.1. Preliminary remarks 120
5.2.2. Fear in Present-day English 121
5.2.3. Historical overview of the verb fear 123
5.2.4. The parenthetical use of fear in affirmative sentences 126
5.2.5. Fear and clauses of other types 133
5.2.6. Additional comments on that-clauses 135
5.3. The historical development of doubt 137
5.3.1. Preliminary remarks 137
5.3.2. Doubt in affirmative sentences and various types of subordinate
clauses 139
5.3.3. Doubt in affirmative sentences and subordinate clauses intro-
duced by whether and if 141
5.3.4. Doubt in affirmative sentences and subordinate clauses intro-
duced by that during the critical period 143
5.3.5. Doubt in affirmative sentences and the development of the
parenthetical use 146
5.3.6. Doubt in negative sentences and various types of subordinate
clauses 149
5.3.7. Doubt in negative sentences followed by that-clauses and but-
clauses 150
5.3.8. Doubt in negative sentences and the parenthetical use 152
5.3.9. The verb question 156
5.4. The historical development of deny 159
5.4.1. Preliminary remarks 159
5.4.2. The complement shift from that-clauses to to-infinitives 162
5.4.3. The parenthetical use and the ellipsis of the conjunction that 166
5.4.4. The rise of gerunds with deny 172

6. Summary and conclusions


6.1. Verbs of implicit negation and different types of complements in the
history of English 177
6.1.1. Verbs of implicit negation and to-infinitives 177
6.1.2. Verbs of implicit negation and gerunds 179
6.1.3. Verbs of implicit negation and gerunds with prepositions 181
viii Table of contents

6.1.4. Verbs of implicit negation and subordinate clauses 185


6.2. Historical shifts of complements in general 190
6.2.1. The rise of to-infinitives, gerunds, and epistemic parentheticals 190
6.2.2. The decline of expletive negation and the introduction of but 192
6.2.3. Lexical diffusion and historical timings 193
6.2.4. Meaning and form again 195
6.3. The OED as a historical corpus 197
Appendix 199
References 205
Index 219
Preface and acknowledgements

Recent scholarship has observed a growing interest in the patterns of verb


complementation in English, especially in the diachronic shift from infinitives
to gerunds. It is widely acknowledged about contemporary English that the use
of gerunds is on the increase in linguistic circumstances where the employment
of infinitives used to be favoured in former days. Potter (1975: 134) maintains
that “[t]he gerund continues to grow at the cost of the infinitive” in Present-day
English, although he admits at the same time the existence of slight differences
in meaning between infinitives and gerunds used as complements. He con-
tinues: “The gerund expresses an action as a process: the infinitive merely
names an action” (p. 134). There are also a number of more recent studies,
often within the framework of corpus linguistics, pointing out the expanded use
of gerunds in English today, e.g. Skandera (2003).
The present research is certainly an outcome in this scholarly tradition.
While the focus of this study is placed only upon a particular group of verbs
—verbs of implicit negation such as forbid and doubt—, the range of the
period dealt with is much wider than in many previous studies. Essentially, the
entire period from Old English to the present day will be discussed, although
some of the verbs explored in this book present only limited data or even no
data for some periods.
The aim of the present study is to elucidate that various syntactic changes
that occurred in the history of English are mutually linked. It shows that the
chronological developments of various verbs as described from Chapter 2 to
Chapter 5 all originate from a single phenomenon in the history of English, i.e.
the decline of that-clauses which took place from later Middle English onwards
and which is most probably related to the decline of the subjunctive.

The fundamental idea of the present book grew gradually in the past ten
years, and some portions of it have already appeared as short articles in various
x Preface and acknowledgements

academic publications. I am grateful to the following publishers and organisa-


tions for granting permission to re-use materials I have published in their
journals and books: Sage Publications, John Benjamins, Peter Lang, and the
Modern English Association. The original materials have been revised in the
present book, in some cases most extensively. Also some errors existent in
earlier versions have been corrected in the present book.
Finally, I would like to express my sense of gratitude to everyone involved
in the production of this book. First of all, my thanks go to those who kindly
read the entire manuscript and commented upon this book: Dr John Reed in
Manchester and Dr Graeme Davis in Brighton. I would also like to thank John
Benjamins and Yushodo Press for accepting to publish this book on a col-
laborative basis. The editorial staff of the two publishers have kindly helped me
in the editing of this book, and I am especially grateful to Mr Kees Vaes and
Ms Patricia Leplae at John Benjamins and Ms Yuko Ogino at Yushodo Press. I
would also like to acknowledge the Kyoto University Foundation, whose
generous financial support made this publication possible.

18 November 2009
Yoko Iyeiri
Abbreviations

BNC The British National Corpus


Brown The Brown Corpus
FLOB The Freiburg-LOB Corpus of British English
Frown The Freiburg-Brown Corpus of American English
LOB The Lancaster-Oslo/Bergen Corpus of British English
OED The Oxford English Dictionary, 2nd edition
List of tables and figures

Tables

1. The number of quotations in different centuries in the OED 63


2. The proportions of to avoid to the totals of avoid with complements in 74
the OED
3. Prevent and three gerundial constructions in previous studies 92
4. Prevent and three gerundial constructions in LOB and FLOB 93
5. Fear in affirmative sentences and clauses where the conjunction that 130
is expressed or unexpressed in the OED (raw frequencies)
6. Fear in negation and clauses where the conjunction that is expressed 130
or unexpressed in the OED (raw frequencies)
7. Fear in affirmative sentences and the proportions of dislocation to the 131
totals of clauses with that unexpressed in the OED
8. I fear and we fear in affirmative sentences in the OED 133
9. Doubt in affirmative sentences and different subordinate clauses in the 140
OED (raw frequencies)
10. Doubt in affirmative sentences and clauses where the conjunction that 147
is expressed or unexpressed in the OED (raw frequencies)
11. Doubt in negation and different subordinate clauses in the OED (raw 149
frequencies)
12. Doubt in negation and clauses where the conjunction that is expressed 153
or unexpressed in the OED (raw frequencies)
13. Deny followed by that-clauses and gerunds in LOB and FLOB (raw 161
frequencies)
xiv List of tables and figures

Figures

1. Forbid and three patterns of complementation in the OED (raw fre- 32


quencies)
2. Forbid and three patterns of complementation in the Helsinki Corpus 34
(raw frequencies)
3. God forbid that … and expletive negation in the Bible in English (raw 38
frequencies)
4. The ordinary use of forbid and expletive negation in the Bible in Eng- 39
lish (raw frequencies)
5. Refuse and its complements in the OED (raw frequencies) 47
6. Forbear and three patterns of complementation in the OED (raw fre- 56
quencies)
7. Forbear in negation and three patterns of complementation in the OED 57
(raw frequencies)
8. Forbear and gerunds with and without from in the OED (raw frequen- 62
cies)
9. Avoid and three patterns of complementation in the OED (%) 67
10. Forbear and avoid of all types in the OED (per 10,000 quotations) 70
11. Prohibit and three patterns of complementation in the OED (raw fre- 82
quencies)
12. Prohibit and gerunds with or without from in the OED (raw frequent- 86
cies)
13. Prevent and three patterns of complementation in the OED (%) 90
14. Prevent with personal pronouns and the three gerundial constructions 94
in the OED (raw frequencies)
15. Prevent with nouns and the three gerundial constructions in the OED 94
(raw frequencies)
16. Prevent with personal pronouns and three gerundial constructions in 96
the OED (%)
17. Hinder and four patterns of complementation in the OED (raw fre- 102
quencies)
18. Hinder and gerunds with or without from in the OED (raw frequen- 104
cies)
19. Hinder with personal pronouns and three patterns of complementation 107
in the OED (raw frequencies)
20. Hinder with nouns and three patterns of complementation in the OED 107
(raw frequencies)
List of tables and figures xv

21. Hinder with personal pronouns and gerunds with or without from in 108
the OED (raw frequencies)
22. Hinder with nouns and gerunds with or without from in the OED (raw 109
frequencies)
23. Hinder with personal pronouns and the three gerundial constructions 110
in the OED (raw frequencies)
24. Hinder with nouns and the three gerundial constructions in the OED 110
(raw frequencies)
25. Refrain and three patterns of complementation in the OED (raw fre- 115
quencies)
26. Refrain and gerunds with or without from in the OED (raw frequen- 116
cies)
27. Fear and three patterns of complementation in the OED (raw frequen- 124
cies)
28. Fear in affirmative sentences and three patterns of complementation in 124
the OED (raw frequencies)
29. Fear in negation and three patterns of complementation in the OED 125
(raw frequencies)
30. Doubt in affirmative sentences and the proportions of whether-clauses 142
and if-clauses to the totals of all types of subordinate clauses in the
OED (%)
31. The proportions of no doubt to the entire sample of doubt used as a 154
noun in the OED (%)
32. Deny and four patterns of complementation in the OED (raw frequen- 163
cies)
33. Deny and clauses where the conjunction that is expressed or unex- 168
pressed in the OED (raw frequencies)
34. Deny in affirmative sentences and clauses where the conjunction that 169
is expressed or unexpressed in the OED (raw frequencies)
35. Deny with first person subjects in the OED (%) 170
36. Deny with first person subjects in affirmative sentences in the OED 171
(%)
37. Deny and two types of infinitival constructions in the OED (raw fre- 174
quencies)
38. Deny and two types of gerundial constructions in the OED (raw fre- 175
quencies)
1. Introduction

1.1. Verbs of implicit negation and the aim of the present book

There is a series of verbs which imply negation in English. Verbs like prohibit
and doubt, for example, belong to this category. Huddleston and Pullum (2002:
835) call these verbs “covertly negative lexical items with clausal or clause-
like complements”, and give the following six subdivisions to them, depend-
ing upon their basic meanings: (a) failure, avoidance, and omission; (b) preven-
tion and prohibition; (c) denial; (d) doubt; (e) counter-expectation; and (f)
unfavourable evaluation. Needless to say, prohibit belongs to (b), whereas
doubt falls into (d).
The principal concern of the present study is to discuss the historical
development of these verbs, paying an especial attention to their syntactic be-
haviours. More specifically, the fundamental aim of the present study is to
depict the complement shifts that they have experienced in the process of their
historical development. For this purpose, I have selected the eleven verbs for-
bid, refuse, forbear, avoid, prohibit, prevent, hinder, refrain, fear, doubt, and
deny in particular, on which the discussions in the following chapters are
based.1 In this study, I will simply call verbs of this kind “verbs of implicit
negation”, although I am aware that there are various designations given to
these verbs in existing studies. Jespersen (1917: 75), for instance, gives the
appellation “verbs of negative import”,2 van der Wurff (1998) “adversative
predicates”, and Rissanen (1999: 273) “verbs with a negative implication”, to
the same series of verbs.
One of the characteristic features of these predicates is that in early

1
In addition to these eleven verbs, I will also discuss the complementation patterns of the verb
question under the section of doubt, since the former verb is similar to the latter in some syn-
tactic features.
2
This term is employed in various previous studies including Machida (1979: 15) and
Martinez (2003: 480).
2 Chapter 1

English they often dominated that-clauses3 which included unnecessary nega-


tion, as the example below illustrates:

They moche doubted that they shold not fynde theyr counte ne tale.
(1483 William Caxton, The Golden Legende)4

From the present-day perspective, the negative adverb not in the that-clause in
this example is not necessary. This phenomenon of including unnecessary
negation, which is often called “expletive negation”,5 is inclined to be the
central concern of previous studies dealing with verbs of implicit negation in
English. Traugott (1992: 270), for instance, refers to the optional and un-
necessary occurrences of negative items in the subordinate clauses of verbs of
implicit negation in Old English, whereas Fischer (1992: 282) discusses the
same phenomenon in Middle English. See also Kent (1890: 129–130), Jack
(1978c: 60, 63–64), and Baghdikian (1979: 676), all of whom discuss the same
phenomenon in Middle English.6
On the whole, expletive negation of this kind experiences a gradual decline
after late Middle English (see van der Wurff 1998). The obliteration of exple-
tive negation is often explained from the fact that the negative adverb ne,
which used to be typical of Old and Middle English, was frequently involved in
the same phenomenon. Van der Wurff (1998: 305–306), for example, maintains

3
Unless otherwise stated, that-clauses in the present study include those where the conjunc-
tion that is expressed and those where it is elliptical.
4
Unless otherwise stated, examples in the present study are cited from The Oxford English
Dictionary (OED), which is the main corpus analysed in the following chapters. Italics in the
citations are mine, except in the case of foreign languages which are originally italicised in
the OED.
5
I will use the term “expletive negation” to refer to unnecessary negation as illustrated by the
example here, although other terms like “paratactic negation” (cf. Jespersen 1917: 75) are
also employed for the same phenomenon in the literature.
6
Whether negation of this type is indeed expletive is another matter and disputed in previous
research. Smithers (1987: 111–112), in his edition of Havelok, takes the view that that-
clauses dominated by verbs of implicit negation are consecutive in function and therefore
negation in them is, in fact, necessary. Ishiguro (1998) takes the same view in his discussion
of the use of forbeodan ‘forbid’ in Old English. Expletive negation is, however, a widely
attested phenomenon in English, and to some extent in other languages as well. It is not
confined to the subordinate clauses dominated by verbs of implicit negation, but is on
occasions evidenced after conjunctions which imply non-fulfillment such as before, until,
and unless. See van der Wurff (1998: 298) and Iyeiri (2001: 91–96) among others for further
details.
Introduction 3

that the decline of expletive negation is ascribable to the disappearance of the


adverb ne in the history of English.
As the example quoted above illustrates, however, negative forms other
than the adverb ne freely occur as an item of expletive negation in earlier
English. Thus, I would take the view the decline of expletive negation needs to
be explained from a different perspective, although I admit at the same time
that the depletion of ne certainly contributed to the decline of expletive nega-
tion to some extent at least. This contention is supported by the fact that
examples of expletive negation are readily available in early Modern English
where the occurrence of ne is extremely marginal (see Rissanen 1999: 273 and
Blake 1988: 107). Furthermore, Denison (1998: 244) even provides two nine-
teenth-century examples of this phenomenon. Hence, the recession of expletive
negation in the strict sense is a process which took a relatively long time in the
history of English.
Although expletive negation itself is not necessarily a major issue of
complementation, it is relevant, and therefore treated in the discussion of the
present research. Something of particular relevance in this relation is the fact
that that-clauses which used to include expletive negation in earlier English
were often replaced by clauses introduced by but (or but that) in later periods.7
Warner (1982: 222–223) considers that but was in competition with that ne
during the Middle English period and that the former took the place of the
latter “probably during the fifteenth century”. The analysis of the Helsinki
Corpus of English Texts by López-Couso and Méndez-Naya (1998) also in-
dicates the rise of but-clauses of this usage after the Middle English period. In
the OED data of the present study, examples of but-clauses are indeed fairly
commonly attested during the Modern English period, as in:

He did not at all doubt but that they would find matter enough to shop the
evidence himself before the next jail-delivery.
(1771 Tobias Smollett, The Expedition of Humphry Clinker, 11 June)

This is an illustrative example of but-clauses dominated by doubt from the


eighteenth century.
Curme (1931: 241) notes the occurrence of but-clauses after what he calls
“verbs which though positive in form are negative in meaning”, and cites some

7
Unless otherwise specified, but-clauses in the present discussion include clauses introduced
by but and but that, the latter of which was often encountered in earlier English. This prac-
tice is followed throughout the present study.
4 Chapter 1

examples from relatively recent English.8 He claims that this is a phenomenon


after an interrogative or negative proposition. Furthermore, examples of but-
clauses of the same kind are available even in Present-day English, as the
following illustrations from the BNC show:

She did not doubt but that somewhere amongst the records and corporate memory
of the Society would be someone who knew Hereward Marr very well indeed.
(1991 Diane M. Greenwood, Unholy Ghosts, from BNC)

I can see no more reason to doubt but that these causes in a thousand generations
would produce a marked effect …
(1992 Peter J. Bowler, The Fontana History of the Environmental Sciences, from
BNC)

The availability of this construction in contemporary English is widely known


in existing studies including handbooks and dictionaries. See, for example, The
American Heritage Book of English Usage (1996: 91) for the employment of
but-clauses after doubt in Present-day English. When necessary, the discussion
in the following chapters deals with the occurrences of this construction in
historical terms.
Before closing the present section, I would like to make clear that the main
interest of the present study is not placed in the categorisation of verbs of
implicit negation per se, although such an attempt would certainly be in
accordance with the recent awareness that different groups of lexical items
display different syntactic behaviours. Like Vosberg (2006), who investigates
syntactic developments of various verbs during the Modern English period, the
present study takes the bottom-up process of scrutinising different verbs one by
one, although the ultimate goal is to clarify the historical development of
complementation patterns in English in general.9 Eleven verbs of implicit
negation have been selected merely for the purpose of investigating some
representative patterns of complement shifts which occurred in the history of
English. Accordingly, the contentions of the present study will in many cases,

8
Curme (1931: 241) also refers to the occurrence of but what-clauses in colloquial English, in
this relation.
9
When I obtained the publication by Vosberg (2006), my research project was well underway,
or rather very close to completion. Although some of the verbs explored in his study are also
discussed in the present study, the difference of corpora between his study and mine leads to
different conclusions on occasions. Also, the methods that we employ are different, which
may be related to the difference in some of our conclusions.
Introduction 5

though not always, be extendable to verbs of other types in English. The over-
all description of the complement shifts in the history of English in general is,
however, a separate piece of research, for which verbs of many other types
need to be analysed as well.
In addition, it is not exactly central to my argument whether a certain verb
is indeed a verb of implicit negation or not. This is essentially a matter of
definition and I am fully aware that other ways of classification than mine are
always possible. The verb fear, for example, is treated as a verb of implicit
negation in Chapter 5 of the present study, while it is a “liking verb” according
to Dixon (2005: 160) and a verb of “cognition, emotion, and attitude” accord-
ing to Huddleston and Pullum (2002: 170). Furthermore, Visser (1963–1973:
§1847) classifies the same verb into the category of “verbs of affection and
disaffection, fearing, desiring, wishing, etc.”
By the same token, the verb forbid, which is treated as a verb of implicit
negation in Chapter 2 of the present study, is discussed under the heading of
“the verbs of commanding and permitting” by Los (2005: 102). Once again,
semantic categorisation of this strict kind is of no necessity as far as the
principal aim of the present study is concerned, since the patterns of comple-
ment shifts revealed in the present discussion may be applicable or at least
extendable to various English verbs in any way, including those which are not
at all negative in meaning, although some adaptations are often necessary
depending upon the verb concerned.

1.2. The Great Complement Shift

The purview of the present research is restricted to the complementation of


verbs in English. Consequently, the term “complement” in this study is used for
subordinate clauses and non-finite clause-like forms which function as a core
and essential argument for the verbs concerned (cf. Hamawand 2002: 1–2,
Fanego 2004a: 322–324, and Dixon 2006: 15). The present study is particularly
interested in the relationship among the three complementation types of that-
clauses (including the elliptical use of that), to-infinitives (including for to-
infinitives), and gerunds with or without prepositions.10 I am, however, aware
10
See Note 3 to confirm that the elliptical use of that is included under the category of
that-clauses. Likewise, for to-infinitives, which were commonly attested in Middle English,
are also included under the category of to-infinitives in the present study. As for gerunds, I
will use the term “gerund” in the traditional sense, not including the present participle form
with –ing, although I am aware that some scholars are critical about making a distinction
6 Chapter 1

of the fact that other clause types like wh-clauses and whether/if-clauses are
commonly dealt with within the framework of complementation in English in
existing studies, which I also investigate in the discussion that follows wher-
ever they are relevant.11
In the contemporary setting, non-finite forms of complements like in-
finitives and gerunds are often considered to be compressed or reduced ver-
sions of that-clauses. Greenbaum (1988: 4), for example, refers to the contrast
between the following two sentences and states that the latter is the “full” form:
(a) They agreed to meet once a month; and (b) They agreed that they would
meet once a month. Quirk, Greenbaum, Leech, and Svartvik (1985: §14.8) also
discuss non-finite forms in general and argue: “Because nonfinite clauses lack
tense markers and modal auxiliaries and frequently lack a subject and a sub-
ordinating conjunction, they are valuable as a means of syntactic compression”.
When viewed from the historical perspective, by contrast, the main con-
cern of discussion will be somewhere else: it is a matter of significant interest
that diachronic shifts of complements are observed in language. Apparently,
English has experienced various complement shifts in the past several hundred
years, which Rohdenburg (2006: 159–160) summarises under the term the
Great Complement Shift. The Shift includes:

・ the rise of the gerund (both “straight” and prepositional) at the expense of in-
finitives (and that clauses)
・ the establishment of linking elements introducing dependent interrogative

between gerunds and participles. Quirk, Greenbaum, Leech, and Svartvik (1985: §17.54),
for example, state: “we do not find it useful to distinguish a gerund from a present
participle”. Also, Huddleston and Pullum (2002: 80–83) employ the term “gerund-
participle” to cover both for the same reason. They maintain: “even from the point of view
of syntax (as opposed to inflection) the distinction between gerund and present participle is
not viable” (p. 83). Furthermore, Duffley (2006: 8) follows Huddleston and Pullum (2002)
in using the term “gerund-participle”.
11
Hamawand (2002: 2) gives the list of that-clauses, wh-clauses, whether/if-clauses under the
category of finite complement clauses, and the list of for to-infinitives, bare infinitives, and
the –ing form (both participial and gerundial) under the category of non-finite complement
clauses. See also Dixon (1995: 184), who again gives a list of different complementation
patterns used in English.
While complements are often introduced by complementisers like that and to, gerunds
are not preceded by any. Fanego (2004a) discusses the complementiser use of the in late
Modern English in sentences like: The deceiving him was easy. She also considers that by
in sentences of the following type, which are increasingly frequent among American stu-
dents nowadays, has the same function as a complementiser: By trying to make her mother
happy proved unlucky for Paul.
Introduction 7

clauses (as in advice on how to do it)


・ the expansion of (subject-controlled, future-oriented) infinitival interrogative
clauses at the expense of finite wh-clauses (e.g. after verbs like hesitate)
・ changes involving the rivalry between marked and unmarked infinitives (e.g.
after the verb help)
・ the simplification of the relevant control properties resulting amongst other
things in the demise or obsolescence of unspecified object deletions with ma-
nipulative verbs like order.

While all the shifts listed here are relevant to the discussion in the following
chapters, the principal goal of the present study is to clarify the process of the
complement shift at the top of Rohdenburg’s list with respect to the historical
development of verbs of implicit negation in English: “the rise of the gerund
(both ‘straight’ and prepositional) at the expense of infinitives (and that
clauses)”.
In my view, this shift is most probably a combination of two different
shifts: (a) the shift from that-clauses to to-infinitives, which the present study
calls the first complement shift; and (b) the shift from to-infinitives to gerunds,
which the present study calls the second complement shift. To be more precise,
the first shift is a shift of complements characterised by the decline of that-
clauses. This can often lead to the rise of to-infinitives, but the rise of other
constructions is also possible to compensate for the loss, as I argue later in the
following chapters. In the case of the verbs doubt and fear, for instance, the
decline of that-clauses leads to the development of epistemic parentheticals (i.e.
I doubt and I fear used parenthetically, almost as a sentential adverb) instead of
the rise of to-infinitives. I would argue below that the development of these
parentheticals has to be understood in the context of the historical shift of
complements in English and particularly in relation to the first shift described
above. Where the first shift produces a sufficient number of to-infinitives, they
can later on lead to the second stage or the second complement shift, producing
gerunds with or without prepositions.
As far as the verbs of implicit negation investigated in the present study
are concerned, these two shifts are more or less clearly distinguished in terms
of when in the history of English they occur. The first shift or the decline of
that-clauses is most prominently observed from later Middle English to early
Modern English, whereas the second shift or the rise of gerunds is charac-
teristically evidenced during the later Modern English period, although there
are verbs in English which have escaped from the second shift or which are
still waiting for the change to take place. Forbid, for instance, is most typically
8 Chapter 1

followed by to-infinitives even today, although a slight rise of gerunds (i.e.


“forbid + –ing”) has been evidenced with this verb in recent years. This issue
will be discussed in detail in Chapter 2.

1.3. Some previous studies and the present research

The essential ideas hitherto outlined are not particularly new. A number of
previous studies, some of which go back to an earlier period of the last century,
notice the existence of various complement shifts that have occurred in the his-
tory of English, although, unfortunately, some are interested in the shift from
that-clause to to-infinitives only and others solely in the rise of gerunds. To
coordinate various existing arguments in the literature by finding some links
among them is also one of the essential purposes of the present study, although
more attention will be paid to the exemplification of various historical paths
which can possibly be taken by English verbs in general.
To illustrate some relevant contentions by previous scholars, one can refer
to Curme (1931: 457), who claims:

For centuries the to-infinitive and its modifiers have been developing into a
distinct subordinate clause of a new type, which has been crowding more and
more out of common use the older that-clause with a finite verb, so that the to-
infinitive has acquired functions unknown to the simple infinitive. Today the in-
finitive clause introduced by to is a form of expression which is felt and used as a
more convenient subordinate clause than the more formal clause introduced by
that …

This is certainly an illustration of the awareness that that-clauses used as


complements gradually receded and were replaced of by to-infinitives in the
history of English (i.e. the first complement shift in the present study). This
also implies that to-infinitives are increasingly a convenient alternative to that-
clauses.
In relation to the historical timing of this complement shift, Rohdenburg
(1995: 371) refers to Visser (1963–1973: §2059), who considers that the use of
that-clauses becomes marked after Thomas More. In analysing the com-
plementation patterns of the verbs make and forbid, however, Rohdenburg
himself feels that “this role reversal [i.e. the shift from that-clauses to
to-infinitives] was accomplished somewhat earlier than is suggested by Visser”
(p. 371). On the other hand, Rohdenburg (1995: 374) refers to verbs of
Introduction 9

ordering and instructing, and says that “the relatively mild charge remains
associated with the finite clause longest of all”, implying that different verbs
display different historical timings in this development.
The dates when the first complement shift took place indeed varies to
some extent depending upon the verb concerned, but it seems most likely that
later Middle English and early Modern English are often the key periods in
terms of this development. Fischer, van Kemenade, Koopman, and van der
Wurff (2000: 211) and Fischer (2003: 27–28) state that that-clauses started to
be replaced by to-infinitives in Middle English, while Manabe (1979: 4–5)
maintains that finite forms [i.e. that-clauses] were overtaken by non-finite
forms like to-infinitives in the sixteenth century.12
Apparently, the decline of that-clauses as a means of complementation
from later Middle English to early Modern English is related to, or even due to,
the decline of subjunctive forms in subordinate clauses during the same period.
Recent studies on verb complementation in English have shown an increasing
tendency to refer to this point. Los’s (2005: 297) contention that subjunctive
clauses were “under threat from the to-infinitive”13 has, for example, been
inherited by Fischer and van der Wurff (2006: 175), who maintain: “when this
form [i.e. the subjunctive] started coalescing with the indicative, the to-
infinitive provided a means to continue marking indirectness and/or non-
factuality of the complement clause”. This statement clearly refers to the link
between the decline of the subjunctive and the employment of to-infinitives.
The study by van Gelderen (2006: 169) is an additional example to illustrate
this scholarly tradition. She states: “Another development related to Early
Modern English verbs is that the Old and Middle English subjunctive endings
are being replaced by modal auxiliaries and infinitival complements”.14
Indeed, the recession of the subjunctive from later Middle English onwards
12
Manabe discusses under the term “non-finite forms” both infinitives and gerunds. He also
deals with non-finite forms of all functions, including the ones used as the subject as well
as the ones used as a verb complement. I have referred to this previous work of his,
however, since the gist here is that the use of finite clauses underwent a dramatic decline in
the course of the early Modern English period.
13
The key contention by Los in his study is that to-infinitives were in competition with
subordinate clauses in the history of English, and not with bare infinitives as often discuss-
ed in the literature. This argument of his is stressed in various other works published by
him as well. See, for example, Los (1998) and Los (2007).
14
Also relevant is the study by van Gelderen (2004: 268–269), though it is on American
English. She refers to the verb require that was typically followed by the subjunctive in the
past and mentions that it “has 8 clear subjunctives, 1 indicative, 47 infinitives …, and an
–ing complement” in a modern spoken corpus of American English.
10 Chapter 1

is most probably a key element in the occurrence of the first complement shift
in English whereby a significant expansion of to-infinitives takes place. For
practical reasons, however, I do not always intend to differentiate the sub-
junctive from the indicative in the discussion in the following chapters,
especially where examples of finite clauses are limited in frequency in any case.
Unfortunately, the formal distinction between the indicative and the subjunc-
tive is not at all clear in many of the examples that I have obtained for analysis,
and also many verbs, which used to dominate subjunctive clauses in earlier
English, now dominate indicative clauses, often showing a mixed state of
affairs in attested examples. The principal aim of the present study is not to
prove the direct link between the decline of subjunctive forms and the rise of
to-infinitives, but rather to trace the historical development of various verbs of
implicit negation in English in a concrete manner, having the link between the
decline of the subjunctive and the first complement shift already as a premise
in a way. I am fully aware of the link between the decline of subjunctive forms
and the instability of finite clauses in general from later Middle English to
early Modern English. The present study is, however, not a piece of research
into the decline of the subjunctive in English itself. Rather, the focus of this
study is placed upon the subsequent development of various syntactic forms in
the history of English.
As an additional point regarding the decline of that-clauses, I would like to
refer in the present study to the development of the parenthetical use of I doubt,
I fear, etc., which is not necessarily explored in the context of complement
shifts in the literature. As I discuss in what follows, this is a process in which I
doubt, I fear, etc. lose their matrix clause status and in which the original
subordinate clause becomes the matrix clause. Although this has often been
treated as a separate issue in existing studies, particularly in relation to the
matter of grammaticalisation, I would contend that the development of epi-
stemic parentheicals like this is also an aspect of the complement shift caused
by the decline of that-clauses from later Middle English to early Modern Eng-
lish. It was a different path from the rise of to-infinitives, a path which was
prepared particularly for limited types of verbs like doubt and fear. Interesting-
ly enough, these verbs also try to some extent to develop to-infinitives in their
historical development, although they are not always successful in this process.
Eventually, they found another way of solution, which was the development of
the parenthetical use. Further details on this issue are given in Chapter 5.
The shift triggered by the decline of that-clauses is then followed by the
second complement shift or the rise of gerunds in English, which is evidenced
slightly later than the first one. This phenomenon has also been noticed in a
Introduction 11

number of existing studies (e.g. Strang 1970: 100). Concerning the timing of
this development, Potter (1975: 134) points out that the shift from infinitives to
gerunds has been attested since the time of Shakespeare, while Blake (1996:
330) notices the same tendency in the Modern English period in general.
Similarly, Fanego (1996b: 49) argues about various verbs in texts later than
1710 that gerunds were “becoming gradually more common” than infinitives in
them. In the following chapters, I will show that the rise of gerunds is
evidenced with various verbs of implicit negation in the later Modern English
period and later, or at least later than the rise of to-infinitives.
In fact, the shift seems to be continuous even today, as Potter (1975: 134)
maintains that “[t]he gerund continues to grow at the cost of the infinitive” in
Present-day English,15 and the trend seems to be especially prominent in cur-
rent American English. Both Mair (2003) and Skandera (2003) point to a
possible trend of the increasing use of gerunds in American English in the
twentieth century, although the verbs analysed by them are rather limited in
number. The focus of the present study is, however, placed upon the historical
development of verbs of implicit negation in British English, although oc-
casional references are made to American English.

1.4. Meaning and form

One possible argument which can militate against the overall view hitherto
presented is that infinitives and gerunds yield slight differences in meaning. In
other words, the choice of complement types may be affected by the meaning
to be conveyed in the sentence. Wierzbicka’s (1988: 23) comment in the fol-
lowing is suggestive in this connection: “we seem to have reached a stage
when anybody wishing to seriously advance the view that syntax has semantic
foundations simply must meet the challenge of English complementation”.
Most of the discussions as to meaning differences of different comple-
mentation patterns are centred upon the contrast between to-infinitives and
gerunds. By contrast, discussions on the meaning differences between that-

15
This is simply a general tendency observed with a number of English verbs. As I argue in
the following chapters, the historical development of complementation patterns differs
significantly depending upon the individual verb. In fact, Denison (1998) notes that there
are verbs “which seem to have moved away from, rather than towards, complementation by
–ing” (p. 267) and gives the examples of fail, intend, propose, and purpose. Chapter 2 of
the present study also considers verbs which are still accompanied by to-infinitives instead
of gerunds in Present-day English.
12 Chapter 1

clauses and other forms of complementation are much less frequent, but
available. According to Hamawand (2002: 332), for example, the main differ-
ence between that-clauses and non-finite forms like to-infinitives and gerunds
is that the former is “grounded in time” and therefore linked to reality. Also,
Vespoor (2000: 217) considers: “when using a sentence with a that-clause, a
speaker construes the main event as an event in which the subject does not
directly interact with the event or state of affairs itself, but with the symbolic
representation thereof”.
Unlike these scholars, however, Noël (2003: 368–371) considers that the
difference between that-clauses and infinitives is perhaps only a matter of
emphasis. He states: “If it were true that the choice of a that-clause rather than
an infinitival complement, or the other way round, is made to convey a distinct
meaning, one may wonder why linguists should have such a hard time pin-
pointing the difference? Surely they are language users themselves, so that with
the help of a little introspection they should fairly easily be able to arrive at
some sort of agreement on the meanings involved?” (p. 369). Thus, this is
certainly a disputed area of research. For some scholars, that-clauses and other
forms of complementation convey different meanings, whereas for other schol-
ars they do not.
Turning to differences of meaning between to-infinitives and gerunds used
as a complement, existing studies are suddenly copious. This is, in fact, one of
the classic topics for linguistic discussions, especially in relation to the choice
of complementation patterns, in contemporary English, although it is also a
controversial area of research. As early as the end of the nineteenth century,
Sweet (1891–1898) notices that gerunds tend to be more general in meaning
than infinitives, which in his terms “bring out more strongly the attributes of
phenomenality” (§2326). A similar view is expressed by Wood (1956), who
states: “Both [i.e. infinitives and gerunds], of course, are non-finite forms, but
where the infinitive, although it does not specify an agent, usually implies one,
the gerund represents the activity as it were in vacuo, without reference to any
agent or occasion” (p. 11). In a similar vein, the famous contention by
Kiparsky and Kiparsky (1970: 145–146) stresses that gerunds are found with
“factive” predicates and that in this sense they differ from infinitives. 16
16
Although the argument by Kiparsky and Kiparsky (1970) is widely accepted to date,
Duffley (2000: 223) questions it by saying: “this approach is still unable to account for uses
of the gerund with verbs like consider (He considered giving her the keys)”. The example
here illustrates a non-factive use of gerunds. See also Egan (2008: 49), who remarks: “there
is no doubt that the –ing form occurs in many non-factive constructions. Indeed, out of a
total of 179 constructions in this study containing an –ing form of complement, as many as
Introduction 13

Furthermore, Tregidgo’s (1980: 45–48) observation in the context of English


teaching is also in the same line.
In respect of more contemporary studies, Quirk, Greenbaum, Leech, and
Svartvik (1985: §16.40) maintain that “the infinitive gives a sense of mere
‘potentiality’ for action” and that the –ing form “gives a sense of the actual
‘performance’ of the action itself”. Vespoor (2000: 215) also argues that
gerunds “prototypically denote co-temporal events and states of affairs” and
that to-infinitives “may have a future sense”. Likewise, Wierzbicka (1988: 69)
considers that gerundial complements “imply sameness of time” in relation to
the time expressed by the matrix verb. Leech’s (2004: 116) claim is also in
keeping with Kiparsky and Kiparsky (1970) mentioned above. He maintains
that infinitives normally express theoretical meaning and gerunds factual
meaning.
Previous studies in this area of research are indeed numerous. Hamawand
is another to note a number of semantic differences found between infinitives
and gerunds. He remarks, for instance, that “[t]he to-infinitive is forward-
looking and pointing ahead to some later time, whereas the –ing gerund is used
to denote activities entirely over at the time associated with the main verb”
(Hamawand 2002: 277). He also considers that “the to-infinitive signals
boundedness in time” and that “the –ing gerund signals unboundedness in
time” (Hamawand 2004: 453).
As a matter of fact, Hamawand, whose studies are based upon the
framework of cognitive linguistics, is of the strong contention that syntactic
differences are always linked to semantic differences. He clearly claims: “I
argue that each type of complement has a meaning of its own that cannot be
substituted freely. It is this meaning which motivates the structure of the
complement clause. In other words, the form of a complement clause is the
consequence of the meaning it evokes” (2002: 13). This is in a way a shared
view among functionalists and cognitive linguists in general. See also
Langacker (1992: 304–305) on this matter.
Since it is unending to refer to further additional previous studies on this
issue, it is perhaps wise to stop here by saying that the tabulated display of
contrast between infinitives and gerunds provided by Allerton (1988: 21, 1991:
474) is useful. See the following:

77 refer to complement situations located either in the future, vis-à-vis the time of the
matrix verb”. Further reference should be made to Duffley and Joubert (1999: 252), who
discuss various problems involved in previous studies on this issue.
14 Chapter 1

INFINITIVE GERUND
infrequent activity regular activity
intermittent activity continuous activity
interrupted activity continuing activity
uncompleted activity completed activity
contingent/possible event event presented factually
particular time and place neutral time and place
special subject non-specific subject
more verbal character more nominal character

Also for a manageable and neat summary of some previous studies on this
issue, see Duffley (2003), and for a further detailed summary of previous
studies in general, see Egan (2008: 45–88).
The possible existence of these differences in meaning may in part be
ascribable to the ultimate origins of the gerund and the to-infinitive in the
history of English. Gerunds are nominal in origin despite the existence of a
number of verbal features in them in today’s English. Although infinitives are
also nominal in a similar sense, the preposition to, which forms part of to-
infinitives, denotes the end-point or the goal in its original form.17 This is most
likely why, as Duffley and Tremblay (1994: 572) and Duffley (2006: 57) point
out, some verbs strongly associated in meaning with the “goal of the effort”
like struggle, strive, labour, and endevour display the to-infinitive construction
rather than the gerund. As Haspelmath (1989: 296) remarks, the preposition to
no longer has a long vowel, and the phonological weakening may be indicative
of the weakening of the original sense. Still, one would not be surprised if its
original meaning or at least its overtone is retained in some measure in current
English. Hence, the implication of the end-point or the goal is often existent or
even “felt” when to-infinitives are employed, as argued in various previous
studies.
In contradistinction to these contentions in the literature, however, there
are simultaneously a number of studies which claim that the difference of
meaning between infinitives and gerunds is minimal or slight. Jespersen
(1909–1949: V, 162), for example, considers that differences between them are
slight, although he does not necessarily deny the possible existence of some.18

17
In the contemporary context, Quirk, Greenbaum, Leech, and Svartvik (1985: §9.32) refer to
the association between the infinitive marker to and the spatial preposition to “through
metaphorical connection”. Both are in fact related in origin.
18
See, for example, Jespersen (1909–1949: V, 192, 199). He notes the directional implication
conveyed by to of the infinitive.
Introduction 15

Curme (1931: 491–492) also argues that “actual usage knows nothing of this
distinction” after making the comment that “[i]t has been claimed that the
gerund is preferred in stating a general fact, while the infinitive is used in
referring to special circumstances of a particular individual act”.
Quirk (1974) is another to be mentioned in this relation. As shown above,
Quirk, Greenbaum, Leech, and Svartvik (1985: §16.40) refer to the possible
difference in meaning between infinitives and gerunds, but at the same time, in
Quirk (1974: 166–167) one finds the following statement: “there ought to be a
big award for anyone who can describe exactly what makes him say ‘I started
to work’ on one occasion and ‘I started working’ on another”. Obviously, there
are indeed studies dealing with the meaning differences between these two
constructions of start (e.g. Řeřicha 1987), but the gist is that they are slight as
Quirk (1974) considers. Likewise, De Smet (2004) investigates the usage of
like, another verb which can be followed by to-infinitives and gerunds, and
reaches the conclusion that the difference of meaning is not absolute. Finally, I
would like to cite De Smet’s (2007: 69) words to the same effect: “motivating
principles [about the choice between infinitives and gerunds] exist but never or
rarely have an absolute, deterministic value”.
Presumably, there are differences to some extent between infinitives and
gerunds in meaning, and the differences are particularly important with some
verbs like try, remember, and forget. The meanings of these verbs differ to a
noticeable extend, depending upon whether they are followed by the to-
infinitive or by the gerund. 19 On the other hand, however, it is also an
undeniable fact that there are a far larger number of verbs with which the
meaning differences between the infinitival complement and gerundial one are
very slight or almost ignorable.20
Even with verbs like try and remember, which often display a fairly clear
distinction in meaning depending upon the complement types they employ, the
relationship between the construction they take and the meaning is not or has
not always been transparent in English. Haegeman (1980) discusses the

19
Here the question is whether the action concerned has been performed or not. In Hamma-
wand’s (2002: 192) words, “[t]he to-infinitive is forward-looking in the sense that it refers
to an action not yet performed …. The –ing gerund is backward-looking in the sense that it
refers to an activity completed”. For further details, see Palmer (1988: 202) and Duffley
(2006: 56–57) among abundance of existing studies on this issue.
20
The position differs significantly depending upon which school of linguistics a scholar
belongs to. A fairly comprehensive survey of various contentions in different linguistic
schools is provided in Chapter 1 of Hamawand (2002). The present study does not intend
to opt for any particular frameworks in this respect.
16 Chapter 1

difference of meaning between try followed by to-infinitives and the same verb
followed by gerunds, stating in the end that the two different usages of the verb
are not entirely separate. Also, it is perhaps relevant to mention that Duffley
(2006) refers to the verb attempt, which is in fact similar in meaning to try, and
says that “it can be construed with both the infinitive and the –ing, yet the
distinction in meaning between the two constructions is practically nil” (p.
58).21 The historical process of the development of try also implies that the
distinction between the two constructions has not always been strict in English.
Fanego (1997: 61–65) states that “try + to-infinitive” is witnessed from 1593,
while she refers to Visser (1963–1973: §1780) in saying that “try + –ing”
appears only “in the beginning of the nineteenth century”.
Remember is also worthy of note in this connection. It takes the gerundial
complement when it refers to an action performed, but again, even with this
verb, to-infinitives can serve the same function when it occurs in the perfect
form of “to have + past participle”, according to Jørgensen (1990: 147–148).
He states that the following two sentences are more or less synonymous, both
of which are again synonymous with: He remembered posting the letter.

He remembered having posted the letter.

He remembered to have posted the letter.

Moreover, Jørgensen (1990: 147–148) provides some quotations from twen-


tieth-century writings, which include:

He could not remember having possessed a thing like this


(1965 Patricia Wentworth, The Gazebo, from Jørgensen)

She remembered to have seen the bats flying low over a burnished pool at sunset
(1932 D. H. Lawrence, The Trespasser, from Jørgensen)

As in the case of try, the historical development of the complementation


patterns of remember also suggests that the meaning differences between the
two types of complements have not always been strict in English. Mair (2006),
whose argument is particularly detailed in treating the historical development
of the word remember, gives the interesting information that “[b]efore 1775,

21
Owen (1993: 181), however, points out that the employment of to-infinitives is more or
less regular with the verb attempt and that the use of gerunds is very restricted with this
verb.
Introduction 17

retrospective –ing complements are not attested [with remember], and after
1875 restrospective to is rare” (p. 219). In addition, Fanego (1996a: 77) dis-
cusses the relatively late development of the “remember + –ing” form in the
history of English. Moreover, it is also relevant to refer to Molencki (1999:
111), whose statement runs as follows: “The verb remember, which in modern
times is complemented by the gerund, was most typically followed by the per-
fect infinitive in the Early Modern English period”.
As hitherto discussed, it is important to be aware that a number of verbs
have experienced complement shifts of some kind or another in their historical
development, demonstrating how they have accepted different patterns of
complementation in the history of English. Accordingly, it may not necessarily
be wise to place so much focus upon meanings as to hinder the analysis of
syntactic developments, although at the same time one could lose the chance to
trace the change of meanings by paying too much attention to syntactic
developments only and forgetting the possible link between syntactic structures
and meaning. In this respect, the comment by Vosberg (2009: 213) in the
following is suggestive:

… in transitional stages of development semantic tendencies are very often


accompanied and even overshadowed by a number of other factors influencing
the variation of grammatical options, so that one of the two non-finite verb forms
provides a clearly preferred alternative in certain syntactic or morphosyntactic
environments.

About previous studies in general, I would consider that some tend to place
too much emphasis on meanings, almost assuming that everything can be
explained from the perspective of meanings. On the other hand, others are
inclined to place too little emphasis on meanings, almost assuming that mean-
ings are not at all related to syntactic structures. Both these directions are rather
one-sided in my view. The intention of the discussion in the present study is to
depict the historical development of some verbs of implicit negation in English,
knowing that there are always such complexities involved as described above.
The fundamental methodology employed in the present research is a syntactic
one, but I am aware, or at least try to be aware, that issues of meanings are also
relevant in some cases. The meaning of the verb doubt in the citation below, for
example, is something close to ‘hesitate’, which I presume is a natural con-
sequence of the particular syntactic structure with which it occurs:
18 Chapter 1

We need not doubt to take away and freely to coerce that improficuous matter of
hair. (1650 John Bulwer, Anthropometamorphosis).

As illustrated by this example, the meaning conveyed by a particular con-


struction cannot be ignored in analysis even though the essential focus of the
discussion in the following chapters is placed upon syntax. Simultaneously,
however, I do not intend to place too much emphasis upon the issue of mean-
ings, either.
To summarise the gist of the above argument, detailed analyses of syn-
tactic developments can help to trace some possible shifts in meanings of verbs
of implicit negation as well, although the latter is not always the principal goal
of the present study. The standpoint of the present study is to investigate com-
plicated matters as they are in complicated forms so long as the complication is
tolerable in discussion. In other words, I will directly plunge into the discussion
of the syntactic development of individual verbs, although I am aware that
different shades of meanings are always existent in different constructions,
which will be mentioned in the process of discussion when necessary. In the
end, language is a phase of complex systems. It is an entity which represents a
system of complexity and which is most ideally treated as a whole.
Finally, the essential theme echoed throughout the following chapters is
centred upon the matter of complement shifts in English in general. This is one
of the characteristic features of the present study, while most previous studies,
though not all, have often paid attention to the development of to-infinitives or
gerunds per se, and not really to the major structural reorganisation which took
place in the history of English. Accordingly, the argument in the following
chapters is to provide hints as to various possible historical paths that different
English verbs have followed or can follow in their development, although the
chronological timing may not necessarily be the same with different verbs. In
the light of the argument by Fanego (1996b: 57) that verbs of negative im-
plication were the first in experiencing the clear shift from to-infinitives to
gerunds, the verbs analysed in the present research may not necessarily be
representative of English verbs in general, as far as the timings of their
historical developments are concerned. Still the directions of development il-
lustrated by the eleven verbs discussed in the present study may be relevant to
some other verbs in English. This book shows that verb constructions are
always susceptible to change in the English language.
Introduction 19

1.5. Data

In the chapters that follow, I will analyse the entire dataset of the quotations in
the second edition of The Oxford English Dictionary on CD-ROM (OED) as
the main corpus. Citations in the present study are, therefore, essentially from
the OED quotations, unless otherwise stated (cf. Note 4). When necessary, I
will also use the Helsinki Corpus of English Texts and the Bible in English on
CD-ROM 22 as supplementary historical corpora, and the Lancaster-Oslo/
Bergen Corpus of British English (LOB), the Brown Corpus (Brown), the
Freiburg-LOB Corpus of British English (FLOB), and the Freiburg-Brown
Corpus of American English (Frown) as supplementary corpora of con-
temporary English. I will also use the British National Corpus (BNC) to collect
some Present-day English examples, particularly when relevant examples are
not copious enough in LOB and FLOB.
It is true that the OED is in its original form intended for use as a historical
dictionary, and not as a corpus for linguistic analyses. However, there is an
increasing awareness among scholars as to how it provides an excellent and
useful collection of historical data for linguistic explorations so long as it is
handled with care (cf. Jucker 1994, Fischer 1997, and Mair 2001, 2004, 2006).
One obvious drawback in using the OED as a historical corpus is that the
distribution of the quotations is “skewed” in a noticeable manner. Mair (2004:
123) mentions: “the editors’ decisions as to what type of texts should be con-
sulted for quotations are not always in line with what today’s linguist would
wish for”. Obviously, this is unavoidable, since it was compiled for more
general use, and not compiled for linguists alone. On the whole, emphasis of
the OED quotations is placed upon literary works as Schäfer (1980: 13) ob-
serves. It is, therefore, not at all recommended for linguists to conduct any
genre analyses or stylistic investigations by using the quotations data in the
OED (see also Mair 2006: 219).
The imbalance or the skewed distribution of quotations is particularly
noticeable in chronological terms. As often noted in previous studies, the num-
bers of quotations in different centuries are not constant in the OED. Schäfer
(1980: 52) demonstrates in the form of a graph that the number of sources
themselves differs significantly depending upon the century, which inevitably
leads to a biased chronological distribution of quotations as well. Especially,
the period before 1580 is underrepresented (cf. Brewer 2000: 48), whereas
Shakespeare, who belongs to a slightly later period, is particularly over-

22
The Bible in English on CD-ROM (Copyright © 1996 by Chadwyck-Healey Ltd.).
20 Chapter 1

represented as Schäfer (1980: 6) remarks. In fact, a number of previous studies


point out that many of the first citations ascribed to Shakespeare can relatively
easily be antedated (cf. Schäfer 1980: 4, 44), which may simply be attributed to
the sheer volume of quotations from his works and perhaps to the under-
representation of pre-Shakespearean periods.
In respect of more recent centuries, quotations in the OED are suddenly
abundant from 1800 onwards, although there is some notable drop at the
beginning of the twentieth century (cf. Hoffmann 2004: 24–25). It is perhaps
most reasonable that editors of the dictionary paid much attention to the
nineteenth century, since it was the most recent century for them when the
project of the OED was launched. On the other hand, the shortage of examples
from the eighteenth century may partly be due to the accidental failure in the
organisation of the OED project whereby some slips assigned to American
volunteers did not eventually arrive, as Murray (1977: 169) mentions. Further-
more, the length of quotations in the OED differs, in some cases to a notable
extent. It tends to be significantly longer in examples from recent years than
those from earlier dates.
Certainly, it is necessary to take these characteristics of the OED into con-
sideration while handling the data extracted from it. In other words, linguistic
projects utilising the OED need to be designed so that these drawbacks should
not affect the results. This does not, however, hinder the use of it for linguistic
purposes. As mentioned above, the usefulness of the OED as a dataset has
increasingly been acknowledged by various linguists in recent years. Hoffmann
(2004: 26) himself states, after mentioning some of its shortcomings, that “it
can nevertheless provide the linguist with a wealth of useful information”.
Moreover, Mair (2001: 608) mentions how substantial the dataset of the OED
is as a historical corpus, and concludes his paper by saying: “I am convinced
that neither the present study nor Denison 1998, illustrating a far more sys-
tematic use of the computerized OED in historical linguistics, have even begun
to exhaust the potential of this fascinating new resource [i.e. the OED] at the
disposal of students of the history of English”. Mair (2004: 124) also considers
that various disadvantages of the OED as a corpus are “offset by one crucial
advantage, namely the sheer mass of material”.
As a matter of fact, it was one of my central aims to test the usability of the
OED as a historical corpus when I embarked upon the present project. I am
now fully convinced of it after analysing eleven verbs of implicit negation by
using the quotations data in the OED. The discussion in the following chapters,
I hope, will reconfirm this point, corroborating the contentions by Hoffman
(2004) and Mair (2001, 2004). After demonstrating this by the analyses of
Introduction 21

different verbs in different chapters, I will return to the present issue in the final
section of Chapter 6.
Despite the excellence of the OED as a corpus, I would, however, consider
that there is, in respect of the twentieth century, dearth of quotations in the
OED, at least for the purpose of the present study. Since the linguistic be-
haviours of the eleven verbs under consideration in the century are often
referred to as a departing point of discussion in the following chapters, it is
most important to have a fairly clear picture of their usage in the twentieth
century, which unfortunately cannot be fulfilled by the use of the quotations in
the OED alone. 23 I will, therefore, mitigate the dearth by the additional
investigation of LOB, Brown, FLOB, Frown, and BNC as I mentioned above.
Despite this methodology, however, the essential problem is in a way in-
surmountable. Considering the difference in the nature of language, I would
feel inclined to avoid mixing the data of the OED and those of the other
corpora in statistical analyses. Hence, the data of the contemporary corpora
under discussion are used only for the sake of reference in the present research.
In other words, a fuller discussion of the complementation patterns of verbs of
implicit negation in current English is still called for, which I will leave for
future studies.
Among the other corpora used in the present research, the Bible in English
is another dataset compiled for other purposes than linguistic analyses. It is a
collection of twenty-one different versions of English Bible translations from
the Old English period to the present day, of which the following are explored
wherever necessary in the present study:24

1. West Saxon Gospels (c990)


2. John Wycliffe (c1395)
3. William Tyndale (1530, 1531, and 1534)
4. Miles Coverdale (1535)

23
Despite the ongoing expansion of the online version of the OED, I will stick to the use of
the CD-ROM version in this study. This is for the reason that it is essential to have a
clearly-defined boundary with a fixed number of examples when dictionaries are used as
corpora in linguistic analyses. An additional reason for the use of the CD-ROM is that the
present research was launched when the online version was not yet available. See also Mair
(2006: 227), who also utilises the CD-ROM version of the OED in his linguistic analyses.
24
For the West Saxon Gospels, the Bible in English on CD-ROM includes two different ver-
sions, of which I have selected the one of MS Corpus Christi College 140 (Cambridge)
only. Also, in the case of the Wycliffite Bible, I have investigated the later version alone,
which is a freer translation, although the earlier version is also included in the Bible in
English.
22 Chapter 1

5. Great Bible (1540)


6. Thomas Matthew (1549)
7. Bishops’ Bible (1568)
8. Rheims Douai (1582, 1609–1610)
9. Geneva Bible (1587)
10. King James Bible (1611)
11. Daniel Mace (1729)
12. Richard Challoner (1750–1752)
13. John Wesley (1755)
14. John Worsley (1770)
15. Noah Webster (1833)
16. Leicester Ambrose Sawyer (1858)
17. Twentieth Century New Testament (1904)
18. New English Bible (1970)
19. Good News Bible (1976)

Here again, the use of this corpus is both advantageous and disad-
vantageous for the purpose of the present study. A natural consequence of
translation in general is that a translated text is never free from the linguistic
influence of the original text, which may be disadvantageous from the
perspective of linguistic studies aiming to analyse the natural behaviour of
language. Side by side with this problem, however, the Bible in English has the
obvious advantage of making a fairly reasonable linguistic comparison and
contrast possible in chronological terms, since the content matter is more or
less constant in different versions of the Bible in different ages. In the
following discussions, I will investigate front matters and side notes attached to
translations as well as the main text of the Bible, since they often provide more
useful data than translated texts themselves, which may be archaic or con-
servative linguistically for their dates.
The remaining corpora explored in the present study are well-known and
frequently used for linguistic analyses in general. The Helsinki Corpus of Eng-
lish Texts, LOB, Brown, FLOB, and Frown have all been compiled with a
relatively strict corpus design by organising how to collect materials from
different genres and different periods, whereas additional emphasis has been
placed upon the quantity in the compilation of the BNC, which is enormous in
size. Since the present study, which refers to them only as supplementary
corpora, does not make the full use of their genre differences in discussion, I
will simply mention their dates and sizes here, and will not describe the genre
structures in them. The Helsinki Corpus of English Texts is a corpus of about
1.6 million words and includes sample data from the Old English to the early
Introduction 23

Modern English period. It is a historical corpus with a relatively well-balanced


collection of texts from different ages, and subdivided into the following
eleven periods: OE1 (–850), OE2 (850–950), OE3 (950–1050), OE4 (1050–
1150), ME1 (1150–1250), ME2 (1250–1350), ME3 (1350–1420), ME4
(1420–1500), EMod1 (1500–1570), EMod2 (1570–1640), and EMod3 (1640–
1710).
The other five corpora are, by contrast, ones of contemporary English.
LOB, Brown, FLOB, and Frown are a group of corpora, though produced at
different places in different dates, in that they all include about one million
words and possess the same kind of genre structures in compilation. LOB
contains written materials in British English from 1961, whereas Brown
written materials in American English from 1961. FLOB and Frown include
the same amount of updated materials from 1991 and 1992, from British Eng-
lish and American English respectively. Finally, the BNC is a large-scaled
corpus of 100 million words—90% from written materials and the remaining
10% from spoken ones—and its data have largely been collected from the last
quarter of the twentieth century.25 For further details of these corpora, see
Kytö (1993), Aston and Burnard (1998), and Mair (1998) among others.

1.6. Overview of the present book

In what follows, the selected eleven verbs of implicit negation are classified
into different types and discussed in appropriate chapters. First of all, Chapter 2
delves into the historical development of forbid and refuse, both of which are
most frequently followed by to-infinitives in contemporary English. Much
attention of this chapter is placed upon the verb forbid, which experienced
interesting shifts of complements in its historical development. While the
occurrence of to-infinitives is relatively stable with refuse throughout the
history of English, the verb forbid was in the past often followed by that-
clauses, which were later superseded by to-infinitives, and occasionally by
gerunds. At least, the first complement shift, which I consider is triggered by
the decline of that-clauses as mentioned above (see 1.2.), is clearly visible with
the development of forbid. The discussion in this chapter traces this process.

25
Although I hold the CD-ROM version of the BNC as an authorised user, I fully benefited
from the BYU-BNC created and maintained by Mark Davies at Bingham Young Uni-
versity, to whom my thanks are due. See <http://corpus.byu.edu/bnc/> (accessed 14 No-
vember 2009).
24 Chapter 1

On the other hand, the construction with that-clauses still remains with this
verb in the fixed or fossilized form God forbid that … in Present-day English,
which is also discussed fairly extensively in Chapter 2. Since the verb forbid is
encountered in abundance in biblical translations, I will make a full use of the
Bible in English on CD-ROM in this chapter as well as the OED dataset, which
is the main corpus of the present study.
Chapter 3 analyses the historical development of forbear and avoid, both
of which are often, though not always, followed by –ing in Present-day English.
The discussion in this chapter elucidates how in the history of English they
have experienced the complement shifts from that-clauses to to-infinitives, and
from to-infinitives to gerunds. In other words, the verbs discussed in this
chapter have experienced the first and the second complement shifts in the
history of English. While the use of gerunds is well-established with avoid in
Present-day English, however, the complementation patterns following forbear
are still unstable today, displaying not only gerunds but also to-infinitives. This
may be due to the fact that forbear is polysemous in nature, as I discuss in this
chapter. Furthermore, the verb is also followed on occasions by the preposition
from plus gerunds as verbs discussed in Chapter 4 are. I will still deal with
forbear in Chapter 3, however, since the employment of from is not as constant
with it as with the verbs treated in Chapter 4. Before closing Chapter 3, I will
further make a brief survey of a phenomenon often called horror aequi. This is
a phenomenon to avoid the repetitive use of the same forms, which is observed
in various different phases of English. In the case of complementation patterns,
the forms often avoided are: (a) the to-infinitive followed by the to-infinitive,
and (b) –ing followed by –ing. The –ing form in (b) refers not only to the –ing
gerund but also to the –ing participle. While this phenomenon can most
appropriately be discussed under the section of avoid as far as the verbs
selected in the present study are concerned, it is a phenomenon applicable to
various English verbs in general.
Chapter 4 investigates the four verbs prohibit, prevent, hinder, and refrain
in succession. The most characteristic feature of these verbs is that they are
usually followed by the preposition from plus gerunds in contemporary English.
Like the verbs discussed in Chapter 3, they have also experienced the historical
shifts of complements, which have ultimately led to the rise of gerunds.
Apparently, the construction with “from + gerunds” represents a step further
forward from the development of simple gerunds. The historical data extracted
from the OED show that the development of “from + –ing” is a later stage that
follows the expanded use of gerunds alone, although these different stages are
not equally transparent with all of the verbs investigated in this chapter.
Introduction 25

Prohibit is a fairly representative verb in this respect, displaying a clear shift


from to-infinitives to gerunds and a clear shift from simple gerunds to gerunds
with from. By contrast, the two stages are not so clearly differentiated in the
case of the other verbs discussed in this chapter. Furthermore, Chapter 4 pays a
special attention to the verb prevent and makes an extensive treatment of the
relationship between “prevent + possessive + –ing” and “prevent + object +
–ing”, since it is an issue which has attracted much attention of previous
studies.
Chapter 5 discusses the three verbs fear, doubt and deny. They also
experienced the first complement shift or the decline of that-clauses in the
history of English. Interestingly enough, however, it did not really lead to the
rise of to-infinitives, which is the usual path of development as the verbs
discussed in the preceding chapters illustrate. The discussion in Chapter 5
demonstrates that the verbs under consideration are characterised instead by the
historical development of the parenthetical use as in I fear and I doubt, where-
by the original matrix verb loses the matrix status. Consequently, original
subordinate clauses are treated almost like matrix clauses. Once fixed as a
parenthetical form, expressions like I fear and I doubt freely move within the
newly-defined matrix clause and take medial and end positions as well as the
initial one, and function almost as a sentential adverb. This development is
most outstandingly observed with fear and doubt, while it is not so obvious
with deny, which tries various possibilities in the history of English, often
displaying mixed types of developments. An additional point to note in this
chapter is that the path leading to the development of the parenthetical use is a
characteristic feature of fear and doubt in affirmative sentences only. It is not
really evidenced with the same verbs when they occur in negative sentences,
although there are some possible examples of this use in negation in the data
analysed.
Finally, Chapter 6 summarises the discussions from Chapter 2 to Chapter 5,
and reconsiders the historical shifts of complements of English verbs in general.
The chapter also considers how the use of the OED as a historical corpus
should be evaluated. It reaches the conclusion that it is an excellent source for
linguistic analyses, although one also needs to be aware of its drawbacks,
which are often related to the process of its compilation.
2. Verbs of implicit negation and to-infinitives

2.1. Verbs of implicit negation and to-infinitives

As discussed above in the Introduction, English verbs have experienced vari-


ous types of complement shifts in their historical development, of which the
present study is particularly interested in: (a) the decline of that-clauses (often
leading to the rise of to-infinitives), and (b) the rise of gerunds (with or without
prepositions) at the expense of to-infinitives. There is, in usual cases, a time lag
between these two types of shifts, the first typically occurring from later Mid-
dle English to early Modern English and the second from the later Modern
English period onwards. Accordingly, (a) is called the first complement shift
and (b) the second, in the present study.
While many English verbs have undergone both of these shifts in their
historical development, there are also some which stick to the employment of
to-infinitives in Present-day English. They have not experienced to any sig-
nificant degree the rise of gerunds or the second complement shift. The present
chapter is particularly concerned with the chronological development of such
verbs. The verbs analysed in the following sections are forbid and refuse. They
are most characteristically followed by to-infinitives in Present-day English,
although the use of gerunds is not altogether absent with them in their
complement position. In other words, the second complement shift is evi-
denced only to a minor extent with these verbs.
In the discussion below, much emphasis is placed upon forbid (2.2.), whose
historical development illustrates the decline of that-clauses and the rise of to-
infinitives in a fairly clear manner. Since the same verb has a long history of
existence in English, providing abundant examples from the Old English
period onwards, it functions as an excellent illustrative verb to show the com-
plement shift under consideration. The verb, though it is usually followed by
to-infinitives in Present-day English, displays the remnant use of that-clauses
when it occurs in the fixed or fossilised form God forbid …, which will be
discussed in some detail in the following sections.
28 Chapter 2

Since relevant examples are copiously attested in English Bible translations,


I will make a full use of the Bible in English on CD-ROM in the present
chapter, especially when discussing the verb forbid, side by side with the data-
set of quotations in the OED or the main corpus of the present study. Unless
otherwise stated, all citations from the Bible are from the Bible in English on
CD-ROM and the other citations from the OED. This practice is followed
throughout the present book. See also Chapter 1 on this matter.

26
2.2. The historical development of forbid
2.2.1. Preliminary remarks

The verb forbid is in usual cases followed by to-infinitives in Present-day


English, as the following examples illustrate:

Additionally this clause forbids the seller to use the order or his connection with
the buyer as a reference sale or generally for advertising or publicity purposes.
(Richard Christou, Drafting commercial agreements, from BNC)

… it forbids us to pretend to do something which in fact cannot be done


(FLOB, G)

At the same time, however, examples followed by gerunds with or without the
preposition from are also available on occasions with this verb. Some il-
lustrative examples of forbid followed by gerunds in contemporary English are:

As an example of how punctuation rules work, consider the rule which forbids
you inserting a comma to separate the subject and the verb.
(Nigel Fabb and Alan Durant, How to Write Essays, Dissertations and Theses in
Literary Studies, from BNC)

It is grossly unfair to forbid Chileans from taking over these investments


(FLOB, A)

The first of the above examples illustrates the use of simple gerunds and the
second the use of gerunds with from.
Although the constructions with gerunds have increasingly been admitted
26
An earlier version of 2.2. appeared as Iyeiri (2003). Sage publications kindly granted
permission to use it in the present study.
Verbs of implicit negation and to-infinitives 29

under the entry of forbid in recent dictionaries (e.g. Longman Dictionary of


Contemporary English, Macmillan English Dictionary for Advanced Learners,
and Oxford Advanced Learner’s Dictionary of Current English, s.v. forbid), it is
safe to state that to-infinitives are still dominant with forbid in Present-day
English in the light of the fact that FLOB provides seven examples of to-
infinitives as against only a single example of the gerund.27 This is confirmed
by the analysis of the BNC by Egan (2008: 356), who shows that the examples
of gerunds used with forbid count less than five in the entire corpus, which is
enormous in size.
In the history of English, one finds an additional construction, namely
forbid followed by a subordinate clause introduced by the conjunction that. See
the two examples below where forbid is followed by that-clauses:28

A lawe whiche did forbidde that they shoulde not woorshippe images.
(1569 James Sandford, tr., Agrippa’s Of Vanitie and Uncertaintie of Artes and
Sciences)

27
As the table below shows, infinitives are dominant with forbid, but gerunds also occur in
British and American English of the 1990s (FLOB and Frown):

that-clauses infinitives gerunds (from + gerund)


FLOB 0 7 1 (1)
Frown 0 9 3 (1)

While gerundial phrases are either directly dominated by forbid or preceded by the pre-
position from, the latter construction is much more frequently discussed in Present-day
English grammars and dictionaries. Partridge (1994: 254), for example, states that the use
of “from + gerund” after forbid is incorrect and that the use of infinitives after forbid is
correct. By contrast, Burchfield (1998: 306), revising Fowler’s Dictionary of Modern
English Usage, admits “from + gerund”, which, he argues, is on the increase these days,
perhaps under the influence of prohibit and prevent, which display the prepositional con-
struction. Fowler had not accepted the use of “from + gerund”.
As far as the table above is concerned, FLOB and Frown both include at least one
example of “from + gerund”. Despite the frequent discussion of the “from + gerund” con-
struction in existing studies, however, the type in which the gerund is directly dominated
by forbid is more common historically. The dataset of the OED provides 80 examples of
this type, whereas it gives only eight instances of the “from + gerund” type. Similarly, the
Helsinki Corpus provides six examples of the direct gerund type, but it yields no examples
of “from + gerund”.
28
In the historical data, forbid is spelled in various manners. Examples of orthographic
variants of this verb are all included under forbid in the present study. The practice of
counting all orthographic variants is followed in the analyses of the other verbs discussed
in the present research.
30 Chapter 2

He forbad that not any body should … use a silver drinking cup.
(1658 William Burton, A Commentary on Antoninus his Itinerary)

The following is essentially of the same kind, except that the conjunction that
is elliptical:

She silly Queene … Forbad the boy he should not passe those grounds.
(1599 William Shakespeare and others, The Passionate Pilgrime)

Examples of this kind are all considered under the category of that-clauses in
the present discussion (see Notes 3 and 10 in Chapter 1).
As mentioned above in 2.1., however, forbid has undergone a shift of
complements from that-clauses to to-infinitives in the course of its historical
development, and that-clauses have almost wholly been ousted by to-
infinitives or on occasions by gerunds with or without prepositions. Although
Herbst, Heath, Roe, and Götz (2004: 324–325) provide the following sentence
which illustrates the employment of that-clauses, examples like this are rare
today, as they themselves state. On the whole, the construction of forbid
followed by that-clauses is no longer available in ordinary circumstances in
Present-day English:

I suppose vanity forbids that he says it.

The employment of that-clauses is now restricted to the fossilised form God


forbid that …, where forbid is optative, in current English, as in:

Joseph answered, “God forbid that I should do such a thing!”


(New English Bible, Genesis 44:17)

The following example, where Lord is employed instead of God, is essentially


of the same kind:29

Lord forbid that she should look interested in him.


(Quinn Wilder, One Shining Summer, from BNC)

These are the only remnant usage of forbid followed by finite subordinate
clauses in contemporary English. Furthermore, the occurrence of this fossilised

29
God forbid that … and its variant forms such as Lord forbid that … and Heaven forbid that
… are all included here. This practice is followed throughout this study.
Verbs of implicit negation and to-infinitives 31

usage itself seems to be infrequent today. The BNC, which is enormous in size
as mentioned above, provides only four examples of God forbid and one
instance of Lord forbid.
In this manner, the complement shift from that-clauses to to-infinitives
was fairly complete in the case of forbid, whereas the second complement shift
or the rise of gerunds at the expense of to-infinitives has not wholly percolated
into the constructions of this verb in the history of English. As illustrated above
by the third and the fourth examples in the present section (see p. 28), there are
indeed instances of gerunds attested with forbid, but they are not yet frequent
in Present-day English. The following discussion will provide further details of
the historical development of forbid.

2.2.2. The decline of that-clauses after the verb forbid

As mentioned in the previous section, it was quite usual for the verb forbid to
dominate that-clauses in earlier English. The Middle English Dictionary indeed
includes a number of examples of “forbid + that-clauses” (s.v. forbeden). Later
on, however, the construction with that-clauses undergoes a dramatic decline in
the history of English, leaving only the fossilised form God forbid that …. As
far as the Bible in English is concerned, the example given below, which is
evidenced in the front matter of King James Bible (1611), is the latest example
of that-clauses dominated by the ordinary use of forbid (i.e. other than God
forbid that …):

We know that Sixtus Quintus expresly forbiddeth, that any varietie of readings of
their vulgar edition, should be put in the margine…
(King James Bible, Front matter)

Although the OED provides three examples of forbid followed by that-clauses


from the nineteenth century, I would presume that they are highly exceptional
for the date. Considering the difference in nature between the God forbid con-
struction and the other uses of forbid followed by that-clauses, it is necessary
to make a separate treatment of the former from the latter. I will, therefore,
exclude the examples of God forbid … from analysis for the moment in the
present section, although I will return to the discussion of the same form in
later sections.
When God forbid … is excluded from analysis, the picture of the com-
plement shift initiated by the decline of that-clauses is relatively clear and
32 Chapter 2

straightforward with this verb. See Figure 1, which displays the raw fre-
quencies of that-clauses, infinitives (mostly to-infinitives, cf. Note 32), and
gerunds dominated by forbid in different centuries in the OED quotations:30

80

60

40

20

0
9th c. 10th c. 11th c. 12th c. 13th c. 14th c. 15th c. 16th c. 17th c. 18th c. 19th c. 20th c.

that-clauses infinitives gerunds


31
Figure 1. Forbid and three patterns of complementation in the OED (raw frequencies)

Since the number of quotations per century is not constant in the OED (see 1.5.
above for details), it is essential to make a comparison and contrast within each
century only. Certainly, it is unwise to compare and contrast the raw fre-
quencies of different centuries under the same type of complements. More
specifically, one can focus upon the sixteenth century, for example, and say that
infinitives are the most frequent of the three patterns of complementation, but
one cannot compare the raw frequency of infinitives in the fourteenth century
with that of the fifteenth century and say that the use of infinitives declined
during this period.
Although relevant examples are rather restricted until the end of the
30
Fundamentally, gerunds were verbal nouns in Old English, which in the history of English
gradually strengthened their verbal features rather than their nominal features (cf. Tajima
1985, Fanego 2004b: 7, and Fischer and van der Wurff 2006: 178–179, among others). In
the case of early English especially, therefore, it is occasionally difficult to tell gerunds
from nouns. I have fairly generously included examples, only excluding clear cases of
nouns such as building, which has nothing to do with movement or action. This practice is
followed throughout the present research. See also 1.4. in the Introduction.
31
Examples of forbid are counted so long as they dominate that-clauses, infinitives, or
gerunds, even if the verb forbid is in non-finite forms such as participles. As for the dates
of texts from which examples are quoted, I have followed the datings in the OED. For
practical reasons, however, I have regarded “c1500” as “1500”, and so forth, in the clas-
sification of examples. Duplicated instances in the dictionary are counted only once in the
statistics here. These practices are followed throughout the present study.
Verbs of implicit negation and to-infinitives 33

Middle English period in the OED quotations, Figure 1 elucidates the overall
shifts of complements of the verb forbid fairly clearly. While the employment
of that-clauses was quite common with this verb during the Old and Middle
English periods—at least in comparison to the other forms—, the rise of to-
infinitives is attested from later Middle English onwards in this graph.32 By the
time of the early Modern English period, infinitives outnumber that-clauses,
perhaps representing the occurrence of the first complement shift.33 The rise of
gerunds is also witnessed during the Modern English period in Figure 1, but on
the whole its use is much less frequent than infinitives throughout the history
of English. In other words, the shift from to-infinitives to gerunds is not so
evident with the verb under consideration. Since I have already quoted some
historical examples of that-clauses above, I will provide in the following some
illustrations of to-infinitives and gerunds quoted from the OED:

He forbiddeth vs also to haue any by lusting.


(1583 Arthur Golding, tr., The Sermons of J. Calvin upon Deuteronomie)

Hee forbade all Glossators, and Commentators to expound it.


(1619 Sir Nathaniel Brent, tr., Sarpi’s Historie of the Council of Trent)

The Levitical law forbids cutting the hair, or rounding the head.
(1781 Samuel Peters, A General History of Connecticut)

He forbade both men and women from entering them.


(1841 Edward W. Lane, The Thousand and One Nights: A New Translation from
the Arabic)

The first two of the above examples of forbid illustrate the use of to-infinitives,
while the last two gerunds with or without the preposition from.
This result in the OED is largely supported by the Helsinki Corpus,
although relevant examples are relatively limited in it. See Figure 2 below,

32
While most examples of the infinitives in the graph are to-infinitives, there are a few
examples of bare infinitives included in Figure 1. For example: My seekly distresse For-
bad myn eres vsen hire office (c1412 Regement of Princes).
33
Figure 1 shows that there is even a twentieth-century example of that-clauses after forbid.
It is different in nature, however, as it is an example of It is forbidden that …, which I have
classified under the category of finite clauses. Although this classification may be con-
troversial, examples of this type are in any case not abundant in the corpora of the present
study. In other words, they do not affect much the entire picture of the historical develop-
ment of forbid as displayed in this graph.
34 Chapter 2

which gives the raw frequencies of the three complementation patterns in each
period of the Helsinki Corpus and which again excludes the examples of God
forbid …:

7
6
5
4
3
2
1
0
OE1 OE2 OE3 OE4 ME1 ME2 ME3 ME4 EMod1 EMod2 EMod3

that-clauses to-infinitives gerunds

Figure 2. Forbid and three patterns of complementation in the Helsinki Corpus (raw fre-
quencies)

As this graph exhibits, the period from OE1 (–850) to ME1 (1150–1250) pro-
vides fifteen relevant examples in the Helsinki Corpus, of which as many as
eleven illustrate the use of that-clauses. The situation of ME2 (1250–1350) is
not clear here as the period does not provide any relevant examples. Still, it is
perhaps safe to conclude from this graph that the period from ME3 (1350–
1420) is the key in terms of the complement shift from that-clauses to
to-infinitives, since one can observe a clear expansion of the latter form around
this time. The rise of gerunds is also witnessed to some extent in this graph, but
the dominant form of complementation of this verb during the Modern English
period is the to-infinitive throughout.
As discussed so far, the historical development of forbid displays an
unequivocal occurrence of the first complement shift or the rise of to-
infinitives, while it only shows some hesitant features of the second com-
plement shift or the shift from to-infinitives to gerunds. Sporadic occurrences
of forbid followed by the gerund from the later Middle English period onwards
may indicate the occurrences of the second complement shift with this verb,
although it may also be ascribable to the influence of other verbs of similar
meanings like prevent and prohibit, which display the use of gerunds. See also
Note 27 above on this issue.
Verbs of implicit negation and to-infinitives 35

2.2.3. God forbid

As hitherto discussed, the overall historical development of the forms of com-


plementation of forbid is transparent. However, the fossilised expression God
forbid … does not follow the ordinary path. Unlike the other uses of forbid, the
fixed form still preserves the finite clause introduced by the conjunction that in
Present-day English. The obvious question is whether it was fossilised at a very
early stage in the history of English, perhaps before the major shift of com-
plements took place with this verb. Most probably, it was indeed the case.
Quite clearly, God forbid in English corresponds to something like ‘let it not
be’ both in Greek and Latin, and has its own tradition as a biblical phrase.
Webster comments on this form of expression in the front matter of his
translation of the Bible and says:

It [i.e. God forbid] is several times used in the version,34 and without any author-
ity from the original languages, for the use of the name of God. The Greek phrase
thus rendered in the New Testament, signifies only ‘Let it not be’, or ‘I wish it
not to be’. I cannot think it expedient to suffer the phrase “God forbid”, to stand
in the text, for the reason assigned in the foregoing paragraph. And it is to be
regretted that a practice prevails of using it in common discourse. I have followed
Macknight in using for these words, By no means.35

As far as the Bible translations investigated in the present study are con-
cerned, the phrase under discussion goes back to the Wycliffite Bible, where it
is most typically used as an interjection and where it clearly corresponds to
Latin absit. See the following examples which illustrate this point:36

The lawe is synne? God forbede. (John Wycliffe, Romans 7:7)

cf. quid ergo dicemus lex peccatum est absit (Vulgate, Romans 7:7)

34
The front matter includes: “The copy used by the compositors was the quarto Bible, pre-
pared for the press by the late President Witherspoon, and published by the late Isaac
Collins, of New York”.
35
Webster uses The Lord forbid…, as in: “The Lord forbid that I should do this thing to my
master” (Noah Webster, I Samuel 24:6); and “The Lord forbid that I should stretch forth
my hand against the Lord’s anointed” (Noah Webster, I Samuel 26:11).
36
Latin examples in the present study are cited from Biblia sacra iuxta Vulgatam versionem
ed. by Robert Weber (Stuttgart, 1975 [2nd edition]).
36 Chapter 2

And if we sechen to be iustified in Crist, we oure silf ben foundun synful men,
whether Crist be mynystre of synne? God forbede.
(John Wycliffe, Galatians 2:17–18)

cf. quod si quaerentes iustificari in Christo inventi sumus et peccatores


numquid Christus peccati minister est absit
(Vulgate, Galatians 2:17)

In other translations of the Bible, God forbid is often followed by that-clauses,


although its use as an interjection is also common in them. Some illustrative
examples of this expression followed by subordinate clauses are:

God forbydd that thy servauntes shulde doo so (William Tyndale, Genesis 4)

God forbid that I should iudge you to be iust (Rheims Douai, Job: 27:5)

Heaven forbid we should ever abandon the law and its statutes
(New English Bible, I Maccabees 2:21)

Evidently, this phrase has been constantly inherited or copied by different ver-
sions of the English Bible, as the following examples illustrate:

But God forbyd that I shulde synne so vnto the LORDE …


(Miles Coverdale, I Samuel 12:D)

Moreouer God forbydde, that I shulde synne agaynst the Lorde …


(Great Bible, I Samuel 12:D)

Moreouer, God forbid that I shoulde sinne against the Lord …


(Bishops’ Bible, I Samuel 12:23)

Moreouer God forbid, that I should sinne against the Lord …


(Geneva Bible, I Samuel 12:23)

Moreouer, as for me, God forbid that I should sin against the Lord …
(King James Bible, I Samuel 12:23)

As for me, God forbid that I should sin against the Lord …
(New English Bible, I Samuel 12:23)
Verbs of implicit negation and to-infinitives 37

As for me, the Lord forbid that I should sin against him …
(Good News Bible, I Samuel 12:23)

The above is a fairly clear case where the same phrase is repeatedly used in
different versions of the Bible, showing how God forbid was a traditional
phrase in biblical translations.
A detailed analysis further reveals that there are some features which dis-
tinguish God forbid from the other uses of forbid from the beginning in the
history of English. The most striking is the absence after God forbid of ex-
pletive negation, which is often encountered in earlier English in the sub-
ordinate clauses dominated by verbs of implicit negation such as forbid, deny,
and doubt (see 1.1. in the Introduction). The following are some examples of
the ordinary use of forbid (rather than God forbid), where expletive negation is
attested:37

I will also forbyd the cloudes that they shall not rayne vpon it
(Great Bible, Isaiah 5:B)

The Lorde forbiddeth you (O ye remnaunt of Iuda) that ye shal not go into Egipte
(Thomas Matthew, Jeremiah 42:D)

The second of the above quotations exemplifies the ordinary use of forbid,
though its subject is the Lord, since the verb form is indicative. The negative
adverb not in these examples is not necessary from the semantic perspective
and in this sense it is expletive.
By contrast, God forbid that … provides examples like the following,
where expletive negation is not available:

And Iudas sayde: God forbyd, that we shulde fle from them
(Miles Coverdale, I Maccabees 9:B)

37
Outside the corpora investigated in the present study, there are some cases in Old English
where the ordinary use of forbid followed by that-clauses does not provide expletive
negation. Ishiguro (1994: 55–58) argues, however, that these examples tend to be found in
translated texts, whose Latin original does not include expletive negation. On the whole,
the decline of expletive negation takes place from later Middle English to early Modern
English. The OED’s latest example of expletive negation under the entry of forbid is from
Shakespeare. For further details on the phenomenon of expletive negation, see 1.1. in the
Introduction.
38 Chapter 2

But God forbid that I should glorie, sauing in the crosse of our Lord Iesvs Christ
(Rhaims Douai, Galatians 6:14)

That-clauses of these examples do not include negation.


I have collected a total of 123 examples of God forbid followed by sub-
ordinate clauses in the Bible in English on CD-ROM, none of which shows
expletive negation, whereas it is consistently attested with the other uses of
forbid up to Rheims Douai (Text 8), after which the phenomenon of expletive
negation itself declines. Then, eventually, that-clauses themselves disappear
except after God forbid in the history of English. This is graphically depicted
by Figure 3 (for God forbid) and Figure 4 (for the other uses of forbid) below,
both of which present the raw frequencies of the examples with or without
expletive negation:

15

12

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

with expletive negation without expletive negation

1. West Saxon Gospels (c990) 11. Daniel Mace (1729)


2. John Wycliffe (c1395) 12. Richard Challoner (1750–1752)
3. William Tyndale (1530, 1531, and 1534) 13. John Wesley (1755)
4. Miles Coverdale (1535) 14. John Worsley (1770)
5. Great Bible (1540) 15. Noah Webster (1833)
6. Thomas Matthew (1549) 16. Leicester Ambrose Sawyer (1858)
7. Bishops’ Bible (1568) 17. Twentieth Century New Testament (1904)
8. Rheims Douai (1582, 1609–1610) 18. New English Bible (1970)
9. Geneva Bible (1587) 19. Good News Bible (1976)
10. King James Bible (1611)

Figure 3. God forbid that … and expletive negation in the Bible in English (raw fre-
quencies)
Verbs of implicit negation and to-infinitives 39

15
12
9
6
3

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

with expletive negation without expletive negation

1. West Saxon Gospels (c990) 11. Daniel Mace (1729)


2. John Wycliffe (c1395) 12. Richard Challoner (1750–1752)
3. William Tyndale (1530, 1531, and 1534) 13. John Wesley (1755)
4. Miles Coverdale (1535) 14. John Worsley (1770)
5. Great Bible (1540) 15. Noah Webster (1833)
6. Thomas Matthew (1549) 16. Leicester Ambrose Sawyer (1858)
7. Bishops’ Bible (1568) 17. Twentieth Century New Testament (1904)
8. Rheims Douai (1582, 1609–1610) 18. New English Bible (1970)
9. Geneva Bible (1587) 19. Good News Bible (1976)
10. King James Bible (1611)

Figure 4. The ordinary use of forbid and expletive negation in the Bible in English (raw
frequencies)

Indeed, relevant examples are not numerous even in the Bible in English,
which is supposed to be most suitable for this kind of analysis. Figures 3 and 4,
however, demonstrate that there is a clear-cut division in the occurrence of exp-
letive negation between God forbid that … and the other uses of forbid follow-
ed by that-clauses until Rheims Douai (Text 8). The phenomenon is observed
in the latter type of forbid only and not in the former. Presumably, the absence
of expletive negation after God forbid is ascribable to its absence in the
original text which the translation is based upon. The following are examples
from the Vulgate that illustrate this point:

absit a te ut rem hanc facias et occidas iustum cum impio fiatque iusus sicut
impius (Genesis 18:25)

absit a nobis ut relinquamus Dominum et serviamus diis alienis (Joshua 24:16)


40 Chapter 2

Neither of these examples includes expletive negation in the subordinate clause.


This is most probably linked to the absence of expletive negation in God forbid
that ….
Another feature which distinguishes God forbid that … from the other uses
of forbid is that the former is again a fixed construction, where God forbid is
essentially followed directly by that-clauses without intervening elements. By
contrast, the intervention of the indirect object between the verb and that-
clauses is much more common in the other cases, as in:

I wil also forbyd ye cloudes, that they shal not rayne vpon it
(Miles Coverdale, Isaiah 5:A)

Wherefore I haue commaunded to forbyd those men that they shall not buylde vp
the citte… (Bishops’ Bible, I Esdras 2:28)

In these examples of forbid of the ordinary use, that-clauses are preceded by


the indirect object. This usage of that-clauses is, in fact, similar to the usage of
that-clauses which introduce an indirect speech. By contrast, the case in which
the indirect object occurs is very rare in occurrence after God forbid and may
be illustrated by the following possible example:

And Naboth said to Ahab, The Lord forbid it mee, that I should giue the
inheritance of my fathers vnto thee (King James Bible, I Kings 21:3)

The indirect object mee is indeed inserted in this example, but this is most
likely a separate construction since the that-clause here refers to the pronoun it.
In any case, this is a rather exceptional example. Apart from this, the Bible in
English does not provide any examples where God forbid is accompanied by
the indirect object.
In this manner, God forbid has been more or less fixed throughout the
history of English and has often been different from the other uses of forbid in
terms of the constructions it takes. And this has eventually led to the excep-
tional retention of that-clauses dominated by God forbid even after the overall
decline of that-clause dominated by the verb forbid.
In my view, however, there was in the past a crucial historical point where
God forbid approached the other uses of forbid and where it was about to
merge into the development of the latter. This is the point where the decline of
expletive negation took place in the history of English. As far as the Bible
translations are concerned, expletive negation is observed after forbid in the
Verbs of implicit negation and to-infinitives 41

ordinary usage up to Rheims Douai, after which the existence or absence of


expletive negation no longer differentiates God forbid from the other cases of
forbid because of the loss of this phenomenon. Around the same time, one
starts observing some mixed examples between the God forbid type and the
ordinary use of forbid, which indicates the possible confusion of the two. Con-
sider the following example from Shakespeare’s Twelfth Night, which il-
lustrates the mixture of the two types:38

Fortune forbid my outside haue not charm’d her!


(1601 William Shakespeare, Twelfth Night)

The textual note in the Arden Shakespeare to this line runs: “A verb of negation
(here ‘forbid’) is often followed by ‘not’” (see Lothian and Craik 1975: 41).
Supposing that the above is an illustrative variation of God forbid that …, the
use of not in the subordinate clause is not expected in theory, since the fixed
phrase is always devoid of expletive negation in English as discussed above.
The occurrence of negation in this example, therefore, displays how the mix-
ture of the two types of forbid was taking place during the early Modern
English period.
Incidentally, expletive negation starts to occur even in infinitival com-
plements from later Middle English onwards, although this usage is by no
means common.39 See the following illustrative example:

This lawe forbiddeth not onlely not to hurt, but to beware least any be hurt
(Geneva Bible, Exodus 21:34, side note)

In this particular case, though, the separation between the verb forbid and the
infinitive may have incited the occurrence of not.
Another feature shared by both types of forbid around this transitional
period is the development of non-assertive forms like any in their complements.
The decline of expletive negation after forbid was followed by the develop-
ment of non-assertive forms in its subordinate clauses. Non-assertive forms
like any, however, start to be attested in the subordinate clauses of God forbid

38
To avoid duplication, I have not so far presented the data of the OED, which also show the
occurrence of expletive negation only in the ordinary usage of forbid and not after God
forbid that …. The example from Shakespeare cited here is in fact the sole possible ex-
ception in the OED quotations.
39
As far as the data of the OED, the Bible in English, and the Helsinki Corpus are concerned,
to-infinitives usually do not include expletive negation.
42 Chapter 2

as well, even though this type of forbid did not show expletive negation from
the beginning. See the examples below. The first illustrates the ordinary use of
forbid, and the second the fixed form God forbid, both showing the use of
non-assertive any in their subordinate clauses:

Forbidding hereby that any gaine gotten of euil things should be applyed to the
seruice of God… (Geneva Bible, Deutronomy 23:18, Side note)40

God forbid, that we who have died to sin should live any longer therein
(Daniel Mace, Romans 6:2)

In the end, subordinate clauses dominated by God forbid and the other types of
forbid are both non-assertive in nature. Accordingly, the expanded use of non-
assertive forms is attested with both types of forbid from later Middle English
onwards, when the general development of non-assertive forms takes place.41
It is significant that the two types of forbid started to reveal some shared
features in early Modern English, since this means that the gap between them
was smaller and smaller in the history of English. Before the merger of the two
was completed, however, that-clauses dominated by the ordinary use of forbid
were increasingly replaced by to-infinitives and then by gerunds to some extent,
as discussed in 2.2.2. above. Due to the decline of that-clauses themselves in
the ordinary usage of forbid, the merger was not fulfilled in the end and the
fixed phrase God forbid that … was left behind again—now in a fossilized
form. If the complement shift from that-clauses to to-infinitives had taken
place slightly later with the ordinary use of forbid, the merger of the two types
could have occurred in time. In other words, due to the general decline of that-
clauses dominated by the ordinary use of forbid from later Middle English to
early Modern English, the fixed form God forbid lost the chance to be fully
assimilated into the ordinary use. It stands out again as a separate expression in
Present-day English, especially in respect of its retention of that-clauses. Now
it is a clearly marked form.

40
The development of any tends to be found more easily in front matters and side notes in
Bible translations. Presumably, translated texts themselves are inclined to be conservative
on this matter.
41
In Iyeiri (2002b), I have shown that the development of non-assertive any in negation was
earlier than argued to date in the literature and that it was evidenced already in later Middle
English. It is generally asserted that the development of non-assertive forms was even
earlier in negatively-coloured contexts (e.g. after verbs of implicit negation such as forbid,
deny, and doubt) than in authentic negation.
Verbs of implicit negation and to-infinitives 43

2.3. The historical development of refuse


2.3.1. Preliminary remarks

While a number of verbs of implicit negation have experienced the comp-


lement shift leading to the rise of gerunds in the recent history of English, the
verb refuse, which I intend to discuss in this section, is still accompanied by
to-infinitives today. In this sense, refuse is similar to forbid discussed in the
previous section, which also displays a clear preference for the infinitival
construction in Present-day English.
As far as the description of this verb in the OED (s.v. refuse) is concerned,
the employment of gerunds is not entirely excluded with refuse, but it is rare
throughout the history of English. Huddleston and Pullum (2002: 1227) list
refuse under the verbs which take to-infinitives in current English. Likewise,
Herbst, Heath, Roe, and Götz (2004: 666–667) mention the construction with
the to-infinitive only, and do not refer to the construction with the gerund about
this verb. By the same token, Egan (2008: 384) shows that the to-infinitive is
the only type of complementation attested with refuse in the BNC, which is
enormous in size. Thus, it is generally agreed in existing studies that refuse is
essentially followed by to-infinitives in today’s usage.
Since the tendency of this verb in Present-day English seems to be rela-
tively straightforward, I have investigated FLOB only (without exploring LOB)
to confirm the points mentioned in existing studies. Indeed, the 50 relevant
examples of refuse in the corpus all illustrate to-infinitives used as its com-
plement. Examples in FLOB include:

The man, who refused to be named, said later: … (FLOB, A)

The club whose members until this year refused to admit women, hosts 15 of the
top judges, including two from the House of Lords. (FLOB, G)

Both these examples illustrate the employment of to-infinitives after refuse.


As for further details of the infinitival constructions of refuse, Nilsen
(1968: 85) points out that it is followed directly by the to-infinitive without the
intervening indirect object. He lists refuse along with begin, cease, continue,
deserve, fail, forget, hope, promise, propose, try, and volunteer, all followed
directly by to-infinitives. The occurrence of the indirect object is possible with
refuse when it is followed by two (pro)nouns as illustrated by the example
below, but not when it is followed by the infinitive:
44 Chapter 2

Laycock had intended to use the Morningside asylum for teaching, but its
superintendent, Skae, refused him access; … (FLOB, H)

Here again, the usage of refuse is quite stable in contemporary English.

2.3.2. That-clauses and/or to-infinitives as the complement of refuse

The syntactic stability of the verb refuse is not confined to its usage in Present-
day English. The verb has, in fact, been stable in terms of the constructions it
shows throughout the history of English. In view of the fact that a number of
verbs of implicit negation have undergone some dramatic shifts of comple-
ments in the history of English, the chronological path followed by refuse is, in
a way, an exceptional one. From the very beginning when it was borrowed into
English, refuse was followed by to-infinitives and it still presents the same con-
struction in contemporary English. The following are some illustrative ex-
amples of this verb from earlier periods:

Bot otherwise, if thou refuse To love, thou miht so per cas Ben ydel.
(1390 John Gower, Confessio Amantis)

The duke of Brytayne began to estrange hym from the Kyng and refusyd to come
vnto his presence.
(1494 Robert Fabyan, The New Cronycles of Englande and of Fraunce)

I wil not refuse to shew you somwhat also of my feathered cattle.


(1577 Barnaby Googe, Heresbach’s Four Bookes of Husbandry)

All these examples of refuse are followed by to-infinitives in the same way as
in Present-day English.
Supposing that there exists a certain kind of correspondence between the
meaning of verbs and syntactic structures, it would not be surprising if the verb
under consideration were followed by that-clauses. As Los (2005: 309– 310)
demonstrates in a tabulated form, Old English forsacan, wiðcweðan, and wið-
sacan, all meaning ‘refuse’, were followed by that-clauses as well as by to-
infinitives. This syntactic structure, however, did not extend to the usage of
refuse, which was borrowed into English. According to Los (2005: 99), sub-
junctive that-clauses were an important source for the development of the
to-infinitive construction of intention verbs, to which refuse belongs. This
contention may be correct in most cases, but in the particular case under
Verbs of implicit negation and to-infinitives 45

consideration here, to-infinitives were the usual form of complementation from


early stages onwards.
Indeed, there may be some correspondence between meanings and syn-
tactic structures in language (see also 1.4. above), but it does not explain the
whole matter of English syntax. Incidentally, Los’s (2005: 73) category of
negative intention verbs aptly include avoid as well as refuse, while the former
is usually followed by gerunds and the latter by to-infinitives in contemporary
English. The contrast between the usages of avoid and refuse will be discussed
again later (see 3.3.).

2.3.3. To-infinitives and/or gerunds as the complement of refuse

Despite the virtual absence of the shift from that-clauses to to-infinitives with
this verb in earlier English, it is, at least theoretically, possible for refuse to go
through the second complement shift in its historical development whereby
to-infinitives are replaced by gerunds, since it provides an abundance of
examples of to-infinitives. In theory, they can function as sources for the
development of gerundial constructions in the late Modern English period. The
gerund is indeed attested with refuse in the historical dataset of the OED, as
illustrated by the following quotations:

One of the principal actors … refused going upon the stage.


(1753 L. M., tr., Du Bosc’s Accomplish’d Woman)

The Officer of the inferior Court can not refuse paying obedience to the Writ.
(1766 Sir James Burrows, Reports of Cases)

The following is another additional example where the gerund is employed in


the complement of refuse. Here, the meaning of refuse is close to ‘deny’:

The deponent refusing his having seen him.


(1752 Maccoll in The Scots Magazine, September)

As proved by these examples, gerundial constructions of refuse were not


entirely unavailable in the historical data, though not at all copious in the data-
set under analysis.
Visser (1963–1973: §1775) also provides examples of gerunds used as a
complement of refuse from 1697 onwards, pointing out at the same time that
46 Chapter 2

the first citation of the relevant construction in the OED goes back only to 1753.
By using the electronic search of the entire dataset of the quotations in the
OED, it is relatively easy now to find even earlier examples of the gerund,
although some illustrate slightly dubious cases as in the following sentence,
where refuse occurs side by side with another verb, which may have influenced
the type of complementation patterns. At least, it is not a genuine example of
the usage at issue. See the following:

He now refuseth and abhorreth the sacrificing of beastes, and al that furniture of
the Leuiticall Presthode, wherwith in the olde time he was delited.
(1561 Thomas Norton, tr., Calvin’s Institution of Christian Religion)

The example below from a Scottish text may also be included as a possible
case where refuse is followed by the gerund:

Four concubynes he … gerte refuse þe entremetynge Forthir till have with


Agrippine.
(c1375 Scottish Legends: Legends of the Saints in the Scottish Dialect of the
Fourteenth Century, Petrus)

While the above examples of gerunds are inclined to be nominal in nature,42


more verbal types of gerunds are also attested in the quotations of the OED.
See, for example:

It was no mean strain of his philosophy to refuse being secretary to Augustus.


(1685 Sir William Temple, Works)

Unlike nominal types, the gerund being in this example is not preceded by
qualifying constituents like the definite article the.
As expected, the occurrence of verbal gerunds increases in later periods
with the verb under consideration, since verbal gerunds themselves increase in
comparison to nominal ones in the course of the history of English, which is a
quite independent development of the gerund. The following is an additional

42
The fact that nominal gerunds are often attested here is in accordance with the contention
by Duffley (1999: 311) that gerundial complements fulfil “the role of direct object”, al-
though his argument is based upon different types of verbs in English. As the main body of
the discussion here reveals, however, nominal gerunds gradually recedes in the history of
English and verbal ones increase instead. See also Note 30 above.
Verbs of implicit negation and to-infinitives 47

example of refuse occurring with a verbal gerund, which I quote from the
eighteenth-century data of the OED:

He bought the Knives and Forks … and believed at that Time they had been Hall
marked, … and refused selling them when he found they were not Hall marked.
(1773 Rep. Comm. Assay Offices)

In this example, the gerund, i.e. selling, is verbal in nature and directly
dominates the object them.
Although examples of gerunds are available as illustrated by some of the
above quotations, they are not at all common with refuse when viewed in the
entire dataset of relevant examples. Figure 5 below elucidates this point by
giving the raw frequencies of infinitival and gerundial constructions occurring
with refuse in the OED quotations:

400

300

200

100

0
14th c. 15th c. 16th c. 17th c. 18th c. 19th c. 20th c.

to-infinitives gerunds

Figure 5. Refuse and its complements in the OED (raw frequencies)

As graphically presented here, constructions with gerunds have always been


marginal with refuse in the history of English. Figure 5 simply confirms how
predominant the use of the infinitive is with this verb.
In this concern, contentions in some existing studies are of some interest.
Fanego (1996b: 48–49) refers to a notable rise of gerunds with refuse after
1710. Rudanko (1998: 72–73), by contrast, claims that gerunds are always rare
during the period from 1680 to 1780 with this verb, although he admits that the
particular collocation cannot refuse is more inclined to be followed than other
uses of the same verb by the –ing form. Turning to the data in the present study,
Figure 5 seems to support the observation by Rudanko, who points out the
scarcity of the –ing form with refuse. I would surmise that an effect of
48 Chapter 2

acceleration would be necessary for a certain construction to replace another,


and that this was simply unavailable with refuse in respect of the historical
development of the gerund. The fact that intention verbs get along well with
to-infinitives in general may also be relevant to the stable use of to-infinitives
with refuse throughout the history of English, at least to some extent, although
at the same time, the relationship between meaning and structures is not
absolute as discussed above and in the Introduction (cf. 1.4.).
One additional point about the usage of to-infinitives after refuse is that the
“refuse + object + to-infinitive” construction occurs with this verb in the his-
torical data, though very marginally in any case. See the following illustration
of this construction:

This letter was … shown this day to Allan Stewart his son, who refuses it to be
his hand-writing.
(1753 The Trial of James Stewart, for the Murder of Colin Campbell of Glenure)

Supposing that this particular construction of the to-infinitive receded by the


time of Present-day English, this would mean that the usage of to-infinitives
after refuse is losing some of its syntactic possibilities, while at the same time it
may simply have been an exceptional construction even in the historical data.
The “refuse + object + to-infinitive” construction is not usually mentioned
under the entry of this verb in dictionaries of Present-day English. The Cam-
bridge Advanced Learner’s Dictionary (s.v. refuse), for instance, refers to the
“refuse + to-infinitive” construction only, and not to the “refuse + object +
to-infinitive” one. The same is true with Macmillan English Dictionary for
Advanced Learners, the Oxford Advanced Learner’s Dictionary, and Longman
Dictionary of Contemporary English (s.v. refuse). By the same token, Nilsen
(1968: 85) also remarks explicitly that refuse is usually followed by to-
infinitives without an intervening object. Thus, the “refuse + object + to-
infinitive” construction does not seem to be acknowledge in dictionaries or
usage guides in general.
This is in keeping with the analysis of the BNC as well. Since examples of
refuse are too numerous in the BNC, I have made a quick survey of the third
person singular form refuses only, which provides 739 examples including all
types, and could not find a single example of the “refuse + object + to-
infinitive” construction in them. Hence, the descriptions in dictionaries of con-
temporary English are corroborated by the data in the BNC.
Whether or not it was once an accepted construction, English has in any
case lost it in the course of the historical development of refuse. This would for
Verbs of implicit negation and to-infinitives 49

the most part be due to the decline of the sense ‘deny’ with this verb, since the
eighteenth-century example quoted above (p. 48) clearly shows that refuse is
used in the sense. In this particular respect, the meaning of refuse is related to
its syntactic development. In any case, relevant examples are so sparse in the
dataset of the present study that further discussions on this matter are im-
possible, at least for the moment.
3. Verbs of implicit negation and gerunds

3.1. Verbs of implicit negation and gerunds

The present chapter discusses two English verbs which have experienced a
significant rise of –ing in their historical development: forbear and avoid. They
are both listed by Visser (1963–1973: §1182) under the category of “verbs of
refusing, abstaining, failing, forgetting, omitting, etc. that combine with an
infinitive [in the history of English]”, and have both experienced the com-
plement shift which eventually led to the extensive use of gerunds. Forbear is
still followed by to-infinitives on occasions (see the OED, s.v. forbear), while
the usage with to-infinitives is unknown with avoid in usual cases in Present-
day English (see the Shorter Oxford English Dictionary on Historical Prin-
ciples, s.v. avoid).43 Visser (1963–1973) himself comments on what he calls
the category of “verbs of refusing, abstaining, failing, forgetting, omitting, etc.
that combine with an infinitive [in the history of English]” and says: “This
group is remarkable for the large number of verbs that nowadays are no longer
collocated with an infinitive” (§1182).
As the discussion in the present chapter reveals, the verb forbear is a fairly
clear example to illustrate the first and the second complement shifts in English
history: the decline of that-clauses usually followed by the rise of infinitives
first and the subsequent rise of gerunds. This is, however, not always the case
with avoid. The construction with gerunds was more or less established with it
at an early stage, and this complementation pattern is almost consistently
encountered throughout the history of English. Hence the rise of to-infinitives
is not clear with this verb, although the attestation of that-clauses and

43
Although the verb followed by to-infinitives is in usual cases unknown in contemporary
English, it is not entirely absent. As I discuss under the sections of avoid in the present
chapter, Vosberg (2006: 18) quotes some examples of the infinitival construction from
twentieth-century materials.
52 Chapter 3

to-infinitives in the past data of this verb certainly suggests that there was most
probably a stage where the historical shifts of complements took place.
This contrast may be, to some extent, related to the difference in the
origins of the two verbs under consideration. Forbear is Germanic and there-
fore has a much longer history than avoid, which was borrowed from French
and which was established in use only in later Middle English. It is a matter of
no surprise if the historical development of the former verb is characterised by
a fuller existence of various potential forms, which eventually leads to the exis-
tence of various complementation patterns even in contemporary English. Al-
though the occurrence of the verb itself is not very frequent in Present-day
English, forbear, when it occurs, is followed by to-infinitives as well as by
gerunds often with from, as mentioned above. Unlike the cases of prohibit,
prevent, and hinder, which are discussed in Chapter 4, however, the rise of the
construction with from is not an outstanding feature of later Modern English in
the case of forbear, although in current English, where the use of forbear is
extremely restricted in frequency, the gerund is inclined to be prefaced by from.
For this reason, the verb forbear is dealt with in the present chapter, while the
discussion of prohibit, prevent, and hinder is left to Chapter 4.

3.2. The historical development of forbear


3.2.1. Preliminary remarks

As mentioned above in the previous section, forbear seems to be a relatively


infrequent verb in Present-day English. The Cambridge Advanced Learner’s
Dictionary as well as Macmillan English Dictionary for Advanced Learners
notes that it is formal (s.v. forbear), suggesting that its employment is sty-
listically conditioned. Its scarcity in current English is also shown by the fact
that Herbst, Heath, Roe, and Götz (2004) do not deal with this verb in their
voluminous work on the complementation patterns of various English verbs.
Furthermore, neither LOB nor FLOB provides any examples of it in their
collection of 1,000,000 words. I have, therefore, decided to resort to the BNC
for perusal of the present state of affairs of forbear, which provides 43 relevant
examples followed by a complement of some kind.44 In view of the enormous

44
This is a verb investigated by Egan (2008) as well. Although he finds 50 examples of
forbear (rather than 43) in the BNC, the total is the number of the tokens of forbear itself,
i.e. all examples of forbear, perhaps including those without any complements. In any case,
the overall picture of this verb provided by his analysis is in accordance with the present
Verbs of implicit negation and gerunds 53

size of the BNC, this is not a high frequency at all, which again confirms how
limited its employment is in Present-day English. Although most of the ex-
amples of forbear are found in the written part of the corpus, there are three
instances in the spoken part—two in relatively formal meetings and one in a
training session.
Interestingly enough, many of the examples of forbear yielded by the BNC
are followed by to-infinitives, as in:

Edward grunted, but forbore to argue. (The World of Thrush Green¸ from BNC)

In the church Richard fell to prayer with such absorption that Lady Dalton
forbore to disturb him when she entered, but her son recognised him as a former
Oxford student.
(Marion Glasscoe, English Medieval Mystics: Games of Faith, from BNC)

From the knowledge I have acquired so far of Johnson, I can not imagine him
forbearing to make comment, nor can I imagine Boswell not seeking advice and
guidance. (Frank Delaney, A Walk to the Western Isles, from BNC)

Of the 43 relevant examples of forbear in the BNC, 30 illustrate the employ-


ment of infinitives as exemplified by the above quotations. In view of the fact
that most examples of to-infinitives of forbear are attested with verbs of saying,
e.g. mention, used in their complements, however, the infinitival construction
of forbear in current English may be a relatively fixed form of expression. To
illustrate this point, I will present some additional examples of forbear follow-
ed by to-infinitives:

He did not enquire after their progress and Nutty forbore to mention it.
(K. M. Peyton, Who, sir? Me, sir? , from BNC)

… she forbore to say very much about the unexpected guest.


(Mary Gervaise, The Distance Enchanted, from BNC)

research. Forbear is indeed an infrequent verb in contemporary English. Like the work by
Vosberg (2006), I obtained the study by Egan (2008) at a consolidating stage of the present
research project. While some of the verbs investigated in the present study are also dis-
cussed by him, the method of analysis employed by his research is different from mine. His
emphasis is placed upon the meaning of different complementation patterns in Present-day
English, whereas my focus is placed upon the diachronic shifts of complements which
various verbs experienced in the history of English.
54 Chapter 3

In other words, forbear, which is most probably in recession in contemporary


English, is retained mainly in some fixed environments or collocations. Of the
30 examples of to-infinitives following forbear in the BNC, four illustrate the
verb to mention, three to say, and two to express. Other verbs occurring with
forbear include: congratulate, indicate, point out, and give details. These are
all in the category of verbs of saying in a broad sense.
Turning to gerunds, examples are less common than infinitives in Present-
day English, which is rather unexpected in view of the fact that it is a verb
which clearly experienced the second complement shift or the rise of gerunds
in the history of English. As mentioned above, the BNC provides 43 relevant
examples of forbear, of which only thirteen illustrate the use of gerunds with or
without the preposition from. Examples of gerunds in contemporary English
include:

Or perhaps I should say: she forbore from actually striking.


(Harpers and Queen, from BNC)

On the part of the plaintiff, it has been urged that the cases cited for the defendant
were not cases where actions had already been brought, but only cases of
promises to forbear commencing proceedings.
(John C. Smith and Joseph A. C. Thomas, A Casebook on Contract, from BNC)

Eight of the gerundial examples in the BNC are accompanied by the pre-
position from and five are not. The gerund without from is not always noted in
contemporary dictionaries, although the construction is certainly existent in
Present-day English. See, for instance, the Cambridge Advanced Learner’s
Dictionary and Longman Dictionary of Contemporary English, both of which
refer to the constructions of to-infinitives and “from + gerund” in particular and
not to the use of gerunds without the preposition from (s.v. forbear).
What makes the current usage of forbear as described above particularly
interesting, especially for the purpose of the present study, is that the simple
gerund construction, which seems to be less common than to-infinitives and
gerunds with from, is the type that underwent a notable expansion in the history
of English, upon which the discussion of this chapter is centered. Unlike verbs
like forbid and refuse, which are usually found with to-infinitives in Present-
day English and which are treated above in the previous chapter, the historical
development of forbear displays a clear rise of gerunds in chronological terms.
Verbs of implicit negation and gerunds 55

3.2.2. The historical rise of to-infinitives and gerunds with forbear

Despite the occurrence in combination with to-infinitives in Present-day Eng-


lish, the verb forbear certainly went through the complement shift from to-
infinitives to gerunds in the history of English. As a matter of fact, it is one of
the verbs which experienced the first complement shift from that-clauses to
to-infinitives and the second shift from to-infinitives to gerunds in a fairly clear
form, although the dearth of its Middle English examples in the dataset under
analysis makes the first shift less obvious than the second one. Middle English
examples of forbear in the OED quotations are often accompanied by that-
clauses, as illustrated by the following quotation:

The king ne miZte tho uorbere, that he ne wep atte laste.


(1297 Robert of Gloucester, Metrical Chronicle)

As with other verbs of implicit negation, that-clauses dominated by forbear


usually include expletive negation. In the example above, it is represented by
the negative adverb ne in the that-clause. Also, the following is an additional
example of expletive negation found in the subordinate clause of forbear:

If eny man be … vndisposid vnscapabili, lete him abstene and forbere that he
come not into prelacie endewid.
(c1449 Reginald Pecock, The Repressor of over much Blaming of the Clergy)

As illustrated by this sentence, which includes expletive not in the that-clause


dominated by forbear, expletive negation is not always in the form of ne
despite the existence of some previous contentions to the effect (see 1.1. above).
Other negative forms are also encountered especially when the matrix clause is
in the affirmative (see also Warner 1982: 210 and Iyeiri 2001: 90).
These examples of that-clauses are, however, to be quickly superseded by
the construction with to-infinitives and then by gerunds in the historical de-
velopment of forbear. The examples cited below illustrate the employment of
the to-infinitive and the gerund respectively:

I forbeare to spend ouer-much time in these kinds of confutings.


(a1617 Samuel Hieron, Works)

I cannot forbear saying one word upon a thing they call a bank, which I hear is
projecting in this town. (1720 Jonathan Swift, Irish Manuf)
56 Chapter 3

To confirm this point, see the figure below, which displays the raw frequencies
of that-clauses, to-infinitives, and gerunds, dominated by the verb forbid in the
OED quotations:45

80

60

40

20

0
13th c. 14th c. 15th c. 16th c. 17th c. 18th c. 19th c. 20th c.

that-clauses to-infinitives gerunds

Figure 6. Forbear and three patterns of complementation in the OED (raw frequencies)

The graph here fairly clearly shows that two complement shifts are involved in
the historical development of forbear: (a) the rise of to-infinitives first, and (b)
the rise of gerunds at a later stage of its development. While the overall trend
depicted in Figure 6 differs significantly from the analysis by Vosberg (2006:
134), who demonstrates how to-infinitives were increasingly established during
the later Modern English period, it is in accordance with the contention by
Fanego (1996b), who discusses the historical development of forbear along
with other verbs. Unfortunately, she does not investigate that-clauses, but at
least about the occurrences of infinitives and gerunds, she claims that both
constructions are still in competition during the period 1640–1710 and that
gerunds are dominant after 1710.
Indeed, Figure 6 shows that the rise of gerunds is a characteristic feature of

45
Infinitives dominated by forbear are usually prefixed with to. The OED, however, gives the
following anomalous example of the bare infinitive occurring with but: I cannot forbear
but transcribe all of it hither (1658 William Burton, A Commentary on Antoninus his
Itinerary). Obviously, this is a structure which occurred under the influence of the cannot
but construction. This single exceptional example is included under the label “to-infini-
tives” in this graph, though in the strict sense it is not an example of to-infinitives, since the
essential aim of this graph is to show the contrasting frequencies among the three com-
plementation patterns of that-clauses, infinitives, and gerunds, following the verb forbear
in the history of English. See also Note 47.
Verbs of implicit negation and gerunds 57

forbear from the eighteenth century onwards in the OED quotations as well,
although the scarcity of relevant examples in the nineteenth and twentieth cen-
turies makes the most recent development of this verb slightly obscure. The
overall picture presented by the OED data is also in accordance with the
general patterns of complement shifts shared by a number of verbs of implicit
negation in English. As mentioned in detail in the Introduction (see 1.2.), the
shift following the decline of that-clauses and often leading to the rise of to-
infinitives usually takes place from later Middle English to early Modern
English and the second shift or the rise of gerunds at the cost of to-infinitives in
later Modern English.
Interestingly enough, the expanded use of gerunds in the later period of
Modern English is even more unequivocal when the data to be analysed are
limited to forbear occurring in negative sentences, as in:

No man could forbear weeping, his conveyance was so affecting.


(1703 James Kirkton, The Secret and True History of the Church of Scotland from
the Restoration to 1678)

Since I am engaged on this Subject, I cannot forbear mentioning a Story etc..


(1711 Joseph Addison, The Spectator, No. 215)

Figure 7 below exhibits the raw frequencies of different complement patterns


of the verb forbear used in negative contexts in the OED:

30
25
20
15
10
5
0
13th c. 14th c. 15th c. 16th c. 17th c. 18th c. 19th c. 20th c.

that-clauses to-infinitives gerunds

Figure 7. Forbear in negation and three patterns of complementation in the OED (raw
frequencies)
58 Chapter 3

The overall features in the rises and falls of different complementation patterns
are shared between Figures 6 and 7. In the latter figure, which displays the
situation of forbear in negation only, however, the expansion of gerunds in
later Modern English is even more prominent.46 Considering the fact that most
examples of forbear in negative sentences are of the following type, where the
same verb is accompanied by the auxiliary can (or could), it is presumably
reasonable to suppose that the use of gerunds along with forbear in negation
was increasingly fixed in the form “cannot (or could not) forbear + –ing” in its
historical development and that this is reflected in the relatively well-
established use of gerunds in negation in the later Modern English period in
Figure 7:

The thing appeared to her so very ridiculous, that … she could not forbear
bursting into a loud laughter.
(1745 Eliza Heywood, The Female Spectator, No. 21)

He could not forbear groaning inwardly.


(1747 Philip Doddridge, Life of Colonel James Gardiner)

Of the total of 60 examples of forbear in negation—all included in Figure


7 above—, as many as 50 include the auxiliary can (or could). Moreover, of the
50 examples of cannot (could not) forbear, as many as 33 are followed by
gerunds rather than that-clauses or to-infinitives. These frequencies reveal how
commonly the form “cannot/could not forbear + –ing” occurs in the entire
dataset under consideration. The same tendency is noted by Rudanko (1998:
61–62), whose nine examples of –ing include six examples of the form
“cannot/could not forbear + –ing”.
It is probable that “cannot forbear + –ing” was something like “cannot
help + –ing”, although the latter is much more frequent partly because help is a
much more frequently occurring verb than forbear in English.47 By contrast,
46
It is rather surprising, at least in contrast to the data displayed in this graph, that Vosberg
(2006: 141) reaches the conclusion that clauses including the negative adverb not tend to
opt for the use of to-infinitives throughout the history of English, despite the fact that his
analysis is based upon fourteen relevant examples only. Further investigations on this
matter may be necessary.
47
Fanego (1996b: 52) refers to the possibility that “cannot forbear + gerund” was in-
creasingly fixed and became idiomatic under the influence of “cannot help + gerund”.
Similarly, “cannot forbear but + infinitive” may have been under the influence of “cannot
but + infinitive”. See the following quotation, where but is followed by the to-infinitive, as
well as the example cited under Note 45 above: Cloria … could not forbeare but plainly to
Verbs of implicit negation and gerunds 59

forbear itself does not seem to be a frequently occurring verb at all. Its
attestation in the entirety of the OED dataset, including the cases where no
complements are involved, counts only about 550. For information’s sake, the
corresponding figure of the verb avoid or the second verb discussed in the
present chapter is around 2,300 and that of the verb prevent around 3,300.
Indeed, the use of forbear is thus restricted in the historical data, and it seems
to be even further restricted in use in Present-day English, as I discuss in the
following section.

3.2.3. The preposition from and forbear in current English

The foregoing discussion has been concerned solely with the differences
among the three complementation patterns that-clauses, to-infinitives, and
gerunds, dominated by the matrix verb forbear. In the case of gerunds, how-
ever, there is a further division to be made in respect of the constructions they
take. More specifically, some examples of gerunds are accompanied by the
preposition from, whereas others occur without it. Since some examples of
gerunds without from are already cited in the previous section, I will provide in
the following two examples with from:

He would have incurred more blame … if he had forborne from attempting to


recover them. (1841 Mountstuart Elphinstone, History of India)

With difficulty she forbore from repeating the cries of lamentation and alarm,
which were echoed around her. (1828 Sir Walter Scott, The Fair Maid of Perth)

In some cases, reflexive pronouns are inserted between the verb and the pre-
position from, as in:

I forbear myself from entering the lists.


(1865 Charles Merivale, The Conversion of the Roman Empire)

Despite the existence of gerunds with from as illustrated by these examples,


however, I have decided to deal with forbear in the present chapter rather than
in Chapter 4, which is supposed to investigate verbs that are in favour of the

tell him her thoughts (1653 Cloria and Narcissus). See also Vosberg (2006: 139–140),
whose exploration displays the process in which the proportion of “cannot forbear +
gerund” to the entire sample of forbear followed by gerunds rises in the history of English.
60 Chapter 3

construction with from in particular. This is due to the fact that the occurrence
of the preposition from is relatively scarce with forbear when viewed within
the framework of the history of English, at least in comparison to the cases of
verbs like prohibit, prevent, and hinder, all of which go through a notable rise
of the “from + –ing” construction in their historical development and which are,
therefore, dealt with in Chapter 4 of the present study.48 The entire dataset of
the OED quotations provides 74 examples of gerunds dominated by forbear
from the sixteenth century to the twentieth century, of which as many as 66
illustrate the employment of gerunds only and without any prepositions.49
Hence, the gerundial construction without from is dominant with forbear in
historical terms.
In view of the overall development of the complementation patterns of
forbear thus far described, it is rather surprising that the construction with the
preposition from (rather than the gerund only) is the form most typically men-
tioned in Present-day English dictionaries (e.g. Oxford Advanced Learner’s
Dictionary of Current English and Cambridge Advanced Learner’s Dictionary,
s.v. forbear), and that more than half of the examples of gerunds after the same
verb in the BNC are indeed accompanied by the preposition from (see 3.2.1.
above). It can also be a matter of surprise that the most frequently occurring
form of complementation in Present-day English seems, in fact, to be the
to-infinitive (rather than gerunds with or without from). This is at least the case,
as far as the examples of forbear in the BNC are concerned (see also 3.2.1.
above).50 Although it is not ideal to mix different types of corpora, in this
particular case the OED and the BNC, on usual occasions, I still find it
worthwhile to further this issue. The tendencies observed in the complementa-
tion patterns of forbear in the historical corpus of the OED do not link
smoothly to its tendencies observed in contemporary English.
48
This is essentially a matter of relative frequency and does not necessarily mean that the
“from + –ing” construction is almost absent with the verb forbear. It is simply more
restricted than in the cases of prohibit, prevent and hinder. In any case, one is obliged to
make a decision as to whether one should discuss the verb in Chapter 3 or Chapter 4, and
the present study simply opts for the former.
49
The remaining eight examples include six with the preposition from, one with for, and one
with of. As the OED (s.v. forbear) states, prepositions other than from are obsolete today.
The following are the examples of for and of: The Dictator … forbare somtime for making
any more lawes (1598 Richard Grenewey, The Annales of C. Tacitus: The Description of
Germanie); and De Beaufort, whom Strickland could not forbear of accusing of
unwarrantable caprice (1787 Ann Hilditch, Rosa de Montmorien: A Novel).
50
As mentioned above, Vosberg (2006: 134) shows the rise of to-infinitives with forbear in
later Modern English.
Verbs of implicit negation and gerunds 61

The points to be considered in this relation are perhaps twofold. Firstly, the
verb forbear is strikingly polysemous. The OED gives various meanings in-
cluding ‘bear’, ‘endure’, ‘submit to’, ‘avoid’, ‘shun’, and ‘refrain’. Interest-
ingly enough, some of the meanings in this list are in an essential contradiction
in some ways. Meanings like ‘bear’, ‘endure’, and ‘submit to’ imply that the
subject or the agent of forbear accepts the object matters of this verb, whereas
meanings like ‘avoid’, ‘shun’, and ‘refrain’ imply incidences of quite the op-
posite direction. Here, the subject of forbear does not accept the object matter.
I have so far dealt with both types of meaning of this verb equally, considering
that the central meaning of it is ‘to be patient’ about the object matter. When
the object matter has a negative implication for the subject or the agent of the
verb, ‘to be patient’ about it means ‘to bear’ it and ‘to endure’ it. Alternatively,
when the object matter possesses a positive implication for the subject of
forbear, ‘to be patient’ about it means ‘to refrain’ from it. In this manner, there
is a shared feature between the two contradictory domains of meanings listed
under the verb forbear.
It is not always easy to decide which of the two opposite types of mean-
ing is the most appropriate for a number of quotations in the OED dataset,
which are often devoid of sufficient contexts simply because of the nature of
the dictionary. The historical direction presented by this verb, however, seems
to be relatively clear. As the OED indicates under the entry of forbear, mean-
ings of accepting the object, e.g. ‘bear’, ‘endure’, and ‘submit to’, are obsolete
in current English. See also the Oxford Advanced Learner’s Dictionary of
Current English and the Cambridge Advanced Learner’s Dictionary, both of
which confirm this point. Consequently, the other meanings of being away
from the object are increasingly prominent with this verb. And the increasing
use of the preposition from is clearly consistent with this historical develop-
ment. The construction “forbear + from” not only follows the syntactic pattern
of refrain, which also implies being away from the object matter and which is
usually accompanied by from, but it shows in an evident manner that the
meaning indicated by this verb is unambiguous. In other words, the object of it
is clearly ‘avoided’, when the preposition from is involved. Accordingly, the
preposition from most commonly occurs with this verb, as mentioned in
contemporary dictionaries.
A careful look at the OED data indeed reveals that the increasing use of
from with forbear is observed in its historical development. See Figure 8,
62 Chapter 3

which exhibits the raw frequencies of gerunds only and gerunds with from after
the verb forbear in the OED quotations:51

40

30

20

10

0
16th c. 17th c. 18th c. 19th c. 20th c.

gerunds only from + gerunds

Figure 8. Forbear and gerunds with and without from in the OED (raw frequencies)

The dearth of relevant examples in the nineteenth and the twentieth centuries in
the OED dataset makes the point slightly unclear. Still, it is evident enough in
this graph that the relative frequencies of the construction with the preposition
from (i.e. the darker column on the right-hand side) were certainly enlarged in
the last two centuries.
Secondly, the extremely restricted occurrence itself of forbear from the
nineteenth century onwards needs to be considered in the interpretation of the
usage of this verb in Present-day English. Since the totals of quotations per
century are not at all constant in the OED in general, it is not recommended on
usual occasions to make a comparison of raw frequencies among different
centuries (see 1.5. above). It is immediately obvious from Figure 8, however,
that the employment of forbear itself is on the wane in the nineteenth century.
One can be fairly confident in stating this in the light of the fact that the
nineteenth century is known to provide the largest number of quotations in total
in the OED and that this would at least theoretically yield the largest number of
examples of forbear in this century,52 which is not the case in the above graph.

51
The examples of the prepositions for and of are included under “from + gerunds” in this
graph. Prepositions other than from are, however, extremely rare with forbear in any case.
See also Note 49 above.
52
I am aware that one needs to be very careful in stating this, since the dataset in the OED is
not a direct extract taken from existent written materials but a collection of consciously
quoted materials. Still, the argument here holds at least in theory.
Verbs of implicit negation and gerunds 63

Furthermore, the relative frequency of gerunds after forbear also increases


during the later Modern English period, which again leads to the theoretical
expectation that a larger number of gerunds with or without from are available
in the nineteenth century than in earlier centuries, and again this is not the case
in Figure 8. On the contrary, the raw frequencies of forbear followed by
gerunds decrease dramatically from the nineteenth century onwards, which
could most reasonably be attributed to the limited occurrences of the verb
forbear itself in recent years.
Considering the fact that the occurrence of forbear itself is thus restricted
in use, it is not a matter of surprise if its syntactic patterns are also restricted
and often found to be in fossilised or fixed forms, which happen to be to-
infinitives and gerunds with the preposition from. It is also relevant to mention
that the use of forbear in negation, which was a good source for the con-
struction of gerunds without from in the historical data, is not at all frequent in
the present-day data of the BNC, although it cannot be proved in the data at
hand whether the verb indeed displays increasing tendencies to be encountered
in affirmative contexts these days.
To confirm the second point, which is related to the quantitative decrease
in the use of forbear itself, I would like to quote in the following the table
showing the number of quotations in different centuries in the OED from Mair
(2004: 124):

Table 1. The number of quotations in different centuries in the OED


Period Number of quotations
–1000 19,769
1001–1100 2,324
1101–1200 11,582
1201–1300 46,205
1301–1400 97,150
1401–1500 96,411
1501–1600 253,528
1601–1700 383,208
1701–1800 273,676
1801–1900 763,987
1901–2000 481,376
Total 2,428,253

This table demonstrates how distinctively numerous the quotations from


nineteenth-century texts are in the entire dataset of the OED. It is by far the
64 Chapter 3

best-represented century in the OED quotations.53 While the total number of


quotations in this century is more than twice as large as the corresponding
numbers of the seventeenth and the eighteenth centuries, the examples of for-
bear with gerunds are much more limited in the nineteenth century as Figure 8
above demonstrates, revealing how much the use of the verb itself increasingly
receded in recent years.

3.3. The historical development of avoid


3.3.1. Preliminary remarks

Unlike the case of forbear discussed above, the present-day usage of the verb
avoid is relatively clear and straightforward. It usually takes the –ing form as
its complement. Swan (2005: 296) treats avoid as one of the verbs “that are
normally followed by –ing forms” in contemporary English. Likewise, both the
Oxford Advanced Learner’s Dictionary of Current English and the Cambridge
Advanced Learner’s Dictionary (s.v. avoid) clearly state that it is a verb which
takes gerunds as its complement, providing some illustrative examples. The
same applies to Longman Dictionary of Contemporary English and Macmillan
English Dictionary for Advanced Learners. In the same vein, Herbst, Heath,
Roe, and Götz (2004: 60–61) and Egan (2008: 327) show that avoid occurs
with gerundial constructions in contemporary English. Hence the use of the
–ing form with this verb is firmly established in Present-day English.
Indeed, all relevant quotations from the twentieth century in the OED (112
examples in all) are found with gerunds, as shown by the following illustra-
tions:

To avoid falling into the more obvious errors of overgeneralization about Latin
America as a whole.
(1970 Solon L. Barraclough in Irving L. Horowitz, Masses in Latin America)

The sweetness is fused with enough real feeling to avoid being sugary, except for
the rather yucky spoken introduction to “Meadows of Springtime”.
(1977 Oxford Times, 1 July 15)

Unlike forbear again, avoid is a frequently occurring verb, which provides a

53
Further detailed discussions on the irregularity of the number of quotations from different
centuries in the OED are given in the Introduction. See 1.5. above.
Verbs of implicit negation and gerunds 65

number of examples, at least in the OED dataset,54 and therefore the descrip-
tion of its behaviours in Present-day English as given above is more or less
reliable, although I admit that Vosberg (2006: 18) presents four exceptional
twentieth-century examples of avoid followed by to-infinitives.
As the OED states, however, the employment of to-infinitives is observed
with this verb in the historical data. Visser (1963–1973: §1182) also lists avoid
under “[v]erbs of refusing, abstaining, failing, forgetting, omitting, etc. that
combine with an infinitive” and provides a dagger to it to demonstrate that the
construction with the infinitive is obsolete in today’s English. He even gives
the following statement in his work: “This group is remarkable for the large
number of verbs that nowadays are no longer collocated with an infinitive”.
Certainly, avoid is one that was once collocated with infinitives but that is no
longer today. This is corroborated by the abundant data explored by Vosberg
(2006: 96–104), who depicts the relatively steady rise of the –ing form with
avoid in the history of English.
As for the timing of the historical expansion of the gerund with avoid,
Visser (1963–1973: §1775) suggests that the –ing form, which is the norm in
Present-day English, was only hesitantly accepted by some writers even in the
nineteenth century. He aptly refers to Goold Brown and says:

But in 1845, in his The Institutes of English Grammar (p. 163), Goold Brown still
questions if the use of the –ing form in “He studied to avoid expressing himself
too severely” is good English.

This statement is interesting despite the fact that Goold Brown is an American
grammarian, since it signals the possibility that the expanded use of –ing may
have occurred only quite recently with the verb at issue.
Vosberg (2006: 96–97), however, maintains that a significant rise of the
use of the –ing form with avoid took place in the course of the eighteenth
century, although his data are slightly irregular in the nineteenth century where

54
Avoid is indeed a frequently occurring verb, although most examples of it are evidenced in
non-finite forms like to avoid, which matter will be discussed later in the final section of
the present chapter. As mentioned in the main body of discussion, there are 112 examples
followed by a complement in the twentieth-century dataset of the OED. Furthermore, the
total of the same verb in the OED counts 848 in the twentieth century only, when the type
not followed by a complement is included. By contrast, there are only four examples of
forbear in the twentieth-century data of the OED, including all types, i.e. types with and
without complements. There is thus a notable contrast between avoid and forbear in terms
of their frequencies.
66 Chapter 3

a notable drop of the –ing form is found. It is, therefore, a matter of interest
where in the history of English the changeover from infinitives to gerunds took
place with this verb. The discussion in the following sections deals with this
question by analysing how avoid is used in the quotations data of the OED. It
will also explore how rapidly the newly-accepted form of the gerund expanded
with this verb in the history of English. Supposing that the gerundial con-
struction was still uncommon and therefore frowned upon in the nineteenth
century as Visser remarks, it would in one way or another be rather surprising
that the older construction of avoid with the to-infinitive has been almost
completely ousted by the use of the gerund in Present-day English. All these
issues will be considered in the following discussion.

3.3.2. The syntactic development of avoid in the history of English

It is evident from the above description that the Modern English period is most
likely the key period in respect of the development of the –ing construction
with avoid. More specifically, Visser’s (1963–1973: §1775) reference to Goold
Brown (1845) cited above hints at the relative importance of the later period of
Modern English in respect of this development. At the same time, however,
Visser (1963–1973: §1775) comments on the early period of Modern English
by referring to Söderlind (1958) and states:

On p. 18 of his syntax of Dryden’s Prose II Söderlind says that avoid is one of the
verbs that in Dryden’s prose take “an infinitive as only object”. Yet on p. 192 in
the same study he cites the following passages: 1693 Dryden, The Satyres (Wks.,
ed. Scott/S.) 84, “it is unpossible sometimes to avoid reading them”; ibid., Notes
190, “Demosthenes gave himself poison, to avoid being carried to Antipater”.
Earliest quotation with form in –ing in OED is dated 1722.

It is, therefore, most likely that the period shifting from early Modern English
to later Modern English is crucial in respect of the development of com-
plementation patterns of the verb avoid. This is also suggested by the data
provided by Vosberg (2006: 96–97).
To turn to the OED quotations data, one encounters both to-infinitives and
gerunds used as the complement of avoid in early Modern English. Illustrative
examples include:
Verbs of implicit negation and gerunds 67

Because he by that meanes would auoid to marry with Alice.


(1599 Richard Hakluyt, Voyages)

Avoyde filchinge and robbinge. (1597 John Payne, Royal Exchange)

In addition to these forms of complementation, examples of that-clauses are


also attested with avoid, as in:

To avoid that none … that had offended the laws, should be received into anie of
their dominions.
(1570–1587 Raphael Holinshed, The Scottish Chronicle)

In this manner, avoid is accompanied by that-clauses and to-infinitives as well


as gerunds in the historical data of the OED.
Further analyses of the shifts of frequencies of the three patterns of com-
plementation reveal that the establishment of the gerundial construction with
avoid was most probably earlier than suggested by Visser (1963–1973: §1775).
Namely, the overall rise of the –ing form in the OED quotations is similar to
the picture presented by Vosberg (2006: 96–97): it is more or less well-
established in the eighteenth century. See Figure 9, which shows the ratios of
that-clauses, to-infinitives, and gerunds to the totals of the three types of
complements in the OED dataset:

100

80

60

40

20

0
16th c. 17th c. 18th c. 19th c. 20th c.

that-clauses to-infinitives gerunds

Figure 9. Avoid and three patterns of complementation in the OED (%)

This graph needs to be handled with care, since there are only eleven possible
relevant examples in the sixteenth century, which makes the statistics there
68 Chapter 3

slightly unstable.55 Despite the scarcity of relevant examples in some centuries,


however, the overall historical tendencies of the three complementation pat-
terns are largely traceable in Figure 9. Certainly, that-clauses and to-infinitives
are once-common-but-now-declining constructions already during the early
Modern English period. Considering the fact, however, that many of the ex-
amples of gerunds in the sixteenth century are ambiguous as to whether they
are authentic examples of gerunds or not—they may simply be nouns ending
with –ing—,56 their proportion needs to be discounted to some extent. In other
words, infinitives were still a fairly common option in the sixteenth century
with this verb. Although this differs in some measure from the contention by
Vosberg (2006: 97), whose data from the period 1500–1700 reveal the
occurrence of the –ing form at the ratio of more than 90%, the situation attested
in the OED data is not really a matter of surprise in view of the fact that the
early Modern English period is often a period when the use of to-infinitives
rises at the cost of that-clauses with various verbs of implicit negation as
discussed in the present study.
The tendency, however, quickly shifts to the next stage in the case of avoid.
Figure 9 above clearly depicts the simultaneous expansion of the gerundial
construction, the use of which comes to be almost fully established by the
beginning of the eighteenth century. It is understandable that “avoid + –ing”
was a rising construction for Goold Brown, but however conservative a com-
ment he made, the overall direction of the historical development of avoid was
already fixed, which no one could possibly subvert. Indeed, it can be a matter
of surprise in a way, if the OED dataset fairly honestly represents the historical
development of avoid, that one can be so inaccurate as Goold Brown in
observing contemporary usages in language.57 Once again, however, it will be

55
Despite the dearth of relevant examples in the earliest period of this graph, I have chosen to
normalise the data by percentages, since the graph of raw frequencies, which is often used
in the present study, has turned out to be even more misleading because of the increasingly
frequent use of the verb avoid itself in the history of English, which will be discussed later
in the present section. The graph of raw frequencies would hide the relationship among the
three complementation patterns behind the enormous number of the examples of avoid in
later Modern English, which are almost always accompanied by gerunds.
56
Some examples of –ing are preceded by adjuncts, as in: To avoyde open ioynyng, … force
to force (a1548 Edward Hall, Chronicle: The Union of the Two Noble and Illustre Families
of Lancastre and Yorke). The elimination of these examples would certainly raise the pro-
portion of to-infinitives to some extent. In any case, the number of ambiguous examples
declines dramatically in the seventeenth century.
57
As I discuss in the following section, one cannot deny the possibility that the OED dataset
is slightly biased in terms of the citations involving the verb avoid.
Verbs of implicit negation and gerunds 69

relevant to note that he was a grammarian of American English, for which the
present study does not provide any data.
Finally, I would like to consolidate the above discussion and make a com-
parison and contrast between avoid and forbear, the two verbs discussed in the
present chapter. As Fanego (1996b) treats them together under the heading of
“verbs of avoiding and forbearing”, there are similarities shared by them in
their historical developments. In both cases, the semantic subject of the com-
plement is usually identical with the subject of the matrix clause. Both forbear
and avoid were followed by that-clauses and to-infinitives at an early stage,
which were relatively quickly replaced by gerunds in their chronological de-
velopment, although the increasing use of the preposition from with gerunds,
particularly in current English, is a marked feature of forbear alone.
For some reason, however, the establishment of the gerundial construction
takes place slightly earlier with avoid than with forbear. It is fully established
with avoid already in the eighteenth century, whereas the lingering of the older
construction with to-infinitives is observed with forbear in the same century.
As a matter of fact, it is remnant with forbear in the following century and
even in Present-day English. As far as the BNC is concerned, the com-
plementation pattern of the to-infinitive is the most common with forbear (see
3.2.1. for details).
Although it is not necessarily easy to find factors related to the distinction
between the historical developments of avoid and forbear, it is perhaps relevant,
at least to some extent, that forbear is a Germanic word inherited from Old
English. In view of the longer history it has, it is feasible that complements
other than gerunds were more fully established with it in earlier days than the
limited number of their examples in the OED dataset suggest. To eliminate
completely something fully established in language may not be an easy task.
By contrast, avoid is a newcomer in Middle English and is not Anglo-Saxon in
origin. It also tried to utilise that-clauses and to-infinitives in the same manner
as other verbs of implicit negation did, at an early stage of its development, but
once it fitted into the newly accelerating trend in the employment of gerunds,
its adaptation to it was a smooth one. This was simply because avoid was freer
than forbear from conventional usages.
In addition, avoid itself is a verb whose occurrences increased dramatically
in the course of the Modern English period. In this process, the expansion of
the use of the gerund or the uprising form was most likely accelerated by the
increasing number of the verb itself. On the other hand, the historical tendency
of forbear was in the opposite direction. As mentioned above, its employment
in Present-day English is much restricted in frequency. Unlike avoid, it is a
70 Chapter 3

verb whose occurrences decreased significantly during the Modern English


period. Although the employment of the gerundial construction also expanded
with it, the trend was perhaps unable to obtain the same kind of impetus as the
verb avoid experienced, since the occurrence of forbear itself was increasingly
limited. See the figure below, which displays the frequencies of forbear and
avoid of all types including those without complements per 10,000 quotations
in the OED. The graph shows the period from the thirteenth century onwards
only, when examples of avoid are available in the OED, although forbear goes
back to the Old English period:

20

16

12

0
13th c. 14th c. 15th c. 16th c. 17th c. 18th c. 19th c. 20th c.

forbear (all types) avoid (all types)

Figure 10. Forbear and avoid of all types in the OED (per 10,000 quotations)

Figure 10, in my view, needs to be handled with the highest level of caveat,
and therefore should be used only for the sake of reference. Although examples
in the OED are in most cases taken from authentic materials where natural
language is used, the compilation of them in the form of dictionary quotations
was an intended and manipulative process where editors chose examples for
their own purposes. I am uncertain, therefore, as to whether one can trust the
mere likelihood of occurrences of a particular word in the quotations (see also
Note 52), although it is probable that the problem is surmounted by the
enormous size of the dataset of the OED itself.58 Another problem is that it is

58
I would assume that analysing the relationship among different complementation patterns
of a particular word, i.e. forbid, refuse, forbear, avoid, prohibit, etc., as conducted in the
other places of the present study, is safe. If one extracts 100 examples of a certain verb
from a period when it is always followed by that-clauses, they all will be followed by
that-clauses as a natural consequence. Similarly, if one extracts 100 examples of a certain
verb from a period when the use of gerunds is dominant with it, the same tendency will be
obtained from the extracted data as well.
Verbs of implicit negation and gerunds 71

practically impossible to eliminate all duplicated examples from the entire


dataset of the OED, whereas my database of forbear and avoid counts
duplicated examples only once. In this way, their frequencies of the two verbs
at issue per 10,000 quotations are slightly, though not very much, discounted in
the above graph. Furthermore, I have used the figures in Table 1 to obtain
Figure 10. Supposing that they are based upon a more recent version of the
OED, the relative frequencies of both forbear and avoid in the above graph are
slightly discounted, since the total of the quotations in the twentieth century
can be larger than the CD-ROM version used in the present study.
In spite of all these difficulties, however, Figure 10 clearly demonstrates
how dramatically the use of avoid expanded in the course of the Modern Eng-
lish period. If the frequency per 10,000 quotations in the twentieth century
needs to be modified on the basis of the exact number of quotations in total, it
will be modified to a larger figure, only ascertaining even more strongly the
noteworthy expansion of the use of the same verb in Modern English. Al-
though the problem of calculating the likelihood of occurrences using
dictionary data remains, the attestation of 7,840 examples of avoid (all types
including ones without complements) in the BNC confirms how frequent the
same verb is in natural language contexts as well. The corresponding figure for
forbear is about 50 only.59 Unlike the case of forbear, therefore, the use of
avoid certainly expanded in terms of frequency in the course of the Modern
English period.60

3.3.3. The so-called horror aequi principle

The present section gives an additional account on the complementation pat-


terns of the verb avoid, and especially refers to the so-called horror aequi
principle, which Rohdenburg (2003) defines as in the following:

… the horror aequi principle involves the widespread (and presumably universal)
tendency to avoid the use of formally (near)identical and (near-)adjacent (non-
coordinate) grammatical elements or structures … (p. 236)

59
In obtaining this data, I have eliminated manually the examples of forbear with the mean-
ing of an ‘ancestor’.
60
I must confess, however, that further research is necessary to confirm this point. Vosberg’s
(2006: 97, 132–133) investigation does not necessarily display the steady rise of avoid, al-
though it supports the restricted use of forbear in recent years.
72 Chapter 3

As Rohdenburg (2003: 236, 244) himself mentions, this tendency or principle


has long and widely been noticed by various scholars including Karl Brugmann,
who according to Rohdenburg presumably coined the term horror aequi in
early dates of the twentieth century. For the purpose of the present study, it is
perhaps clearer to rephrase this principle in the following manner: the se-
quential occurrence of two non-finite forms of the same kind tends to be
avoided. In other words, to-infinitives are likely to be followed by –ing rather
than to-infinitives, and –ing forms are inclined to be followed by to-infinitives
instead of gerunds. This is simply to avoid the repetition of to-infinitives and
the repetition of the –ing form. For a further detailed description of horror
aequi, see Vosberg (2006: 40–51), who gives a number of examples to il-
lustrate this principle.
Although I am uncertain as to how strongly this principle affects the
choice of complementation patterns of English verbs in general, it is certain at
least that the sequence of –ing and –ing has attracted much attention of pre-
vious studies (e.g. Milsark 1972, Ross 1972, Pullum 1974, Bolinger 1979, and
Westney 1992). While some studies stress how unacceptable it is, others take a
more relaxed position on this matter. With his analysis of begin and start,
Gramley (1980: 166–167) shows that the principle under discussion is perhaps
a fairly strong rule.61 Westney (1992: 503), by contrast, concludes his paper by
saying: “… perhaps it is time to start considering stopping bothering so much
about multiple –ings”. Furthermore, Bolinger (1979), considering that the
matter is essentially a phonological one, remarks:62

Doubling [of the –ing] may be tolerated when it is less noticeable. This occurs when
something intervenes, whether lexical material or a prosodic (or grammatical?) break,
when one or both of the –ings is de-accented or not in a prominent position in the
sentence, and when one or both of the –ings is lexicalized as a non-verb (and
processed directly). It is also tolerated in some highly colloquial colligations. It seems

61
Řeřicha (1987: 130) also notes about begin and start that only the infinitive is selected in
examples like I’m beginning to understand, although he does not in particular refer to the
avoidance of the repetition of –ing. As a matter of fact, he seems to associate the selection
of this construction to the meaning of the infinitive, which according to him implies only
“entry into the initial phase of an activity”.
62
Bolinger’s analysis of the doubling of –ing is most detailed. He, for example, notices the
fine contrast between *He was keeping singing, which is avoided, and He was enjoying
singing songs, which is better (p. 42). He also points out that the same tendency as the
doubling of –ing is observed with other suffixes and gives the following example: “It was
rotten that Otten had gotten forgotten” (p. 56). Apparently, what is particularly disliked is
the “sequential” occurrences of the same forms or even sounds.
Verbs of implicit negation and gerunds 73

that when the urge to say something is powerful enough, the niceties of prosody can
be disregarded. (p. 55)

This suggests that the principle is not necessarily a strict rule.


Ross’s (1972: 68) study on the doubling of –ing includes similar arguments
to the above quotation, especially in relation to the intervention of elements
between the two –ings. He gives the following two sentences and claims that
the second which involves topicalisation is “rescued” in a way: (a) *I’m not
particularly keen on trying kissing this moray eel; and (b) Kissing this moray
eel I’m not particularly keen on trying.
In the case of avoid, the use of gerunds as its complement is fully es-
tablished by the time of the eighteenth century (see the above discussion), and
this means that the –ing form is practically the only possible complement
option for this verb even if it occurs in the –ing form. Accordingly, one
observes the existence of the following type, where both the verb avoid and its
complement are in the –ing form:

For the sake of avoiding unnecessarily wounding the timbers.


(1886 Encyclopaedia Britannica, s.v. Shipbuilding)

In this example, the horror aequi principle is violated in a way. About ten
examples of the combination of two –ing forms as illustrated by the above
example are found in the OED quotations, when the present participle, which
also ends with –ing, is included.63 The following is a quotation to illustrate the
case where the present participle is involved:

He was neatly avoiding spilling spaghetti sauce over a very snappy jacket.
(1968 Charles Drummond, Death and Leaping Ladies)

In this example, avoid, which occurs in the form of the present participle, is
followed by the gerund. As a consequence, the –ing form occurs in repetition.
This does not imply, however, the horror aequi principle is irrelevant to
the choice of complements of avoid. On the contrary, it seems most likely
relevant. Vosberg (2006: 97–99) quite convincingly demonstrates how much
delayed the establishment of the gerundial complement was in the environment
where the verb avoid occurred in the form of avoiding. He also displays the
process in which avoiding gradually accepted the complement of –ing in the
history of English and says: “je etablierter die –ing-Form im allgemeinen, desto
63
For the relevance of the present participle form –ing, see Note 62 above.
74 Chapter 3

höher ist sicherlich auch die Akzeptanz von –ing-Formen unter besonders
widrigen Bedingungen” (p. 98). “Unter besonders widrigen Bedingungen” here
means the context where avoid itself occurs in the –ing form. Thus, the –ing
complement was gradually accepted even in contexts where the verb avoid
itself occurred in the –ing form, as the use of the gerundial complement
became more and more established with this verb. This implies that the
doubling of the –ing form was avoided when there was a choice between the
infinitival and gerundial complements. In other words, the principle of horror
aequi affected the timing of the historical development of the –ing form with
the verb avoid, although it was not strong enough to hinder the employment of
–ing after avoiding once the gerundial form of complement was established
with this verb in the history of English.
To turn to the dataset of the OED, it is noticeable that a surprisingly large
number of examples of avoid itself occur in the to-infinitive form and thereby
welcome in a way the development of the –ing form. The following are some
instances to illustrate this:

To avoid ensuing jars Ile hamper vp the match, Ile … wed you here.
(c1590 Robert Greene, Friar Bacon)

The Curtizans make all the forepart of their gownes in like manner open, to avoid
wrinckling. (1617 Fynes Moryson, An Itinerary)

The Horses going … in a string and keeping the furrow, to avoid poching the
Land. (1677 Robert Plot, The Natural History of Oxfordshire)

The entire dataset of the OED quotations provides a total of 303 examples of
avoid followed by a complement of some sort, of which as many as 170
(56.1%) are attested in the to-infinitive form (i.e. to avoid). Furthermore, this
tendency seems to be accelerated as time passes, as the following table
exhibits:

Table 2. The proportions of to avoid to the totals of avoid with complements in the OED
to avoid Totals
16th century 7 (36.4%) 11
17th century 28 (45.5%) 33
18th century 48 (58.3%) 48
19th century 97 (51.5%) 99
20th century 112 (64.3%) 112
Totals 170 (56.1%) 303
Verbs of implicit negation and gerunds 75

Table 2 displays how frequently the verb avoid occurs in the infinitival form to
avoid in the form of percentages. As it shows, the occurrence of avoid is clearly
skewed in the way it supports the full development of the –ing form in its com-
plement positions, and this trend is increasingly marked in the history of
English.
It is also relevant to mention in this connection that exceptional examples
of to-infinitival complements are inclined to be attested with avoid used in the
–ing form. One of the two OED examples of to-infinitives in the nineteenth
century, for instance, is the following, in which avoid occurs in the form of
avoiding, again conforming to the horror aequi principle:

A talent … for fighting … and … a talent for avoiding to fight.


(1858 Thomas Carlyle, Frederick the Great)

The use of to-infinitives with avoid in the nineteenth century is already rather
exceptional in the light of the relatively early establishment of the gerundial
form with it as discussed above. It is, therefore, most likely that the choice of
this complement was encouraged by the –ing form of avoid and that this was to
avoid the sequential occurrence of the –ing form. Moreover, the four twen-
tieth-century examples of “avoid + to-infinitive” provided by Vosberg (2006:
18) also include three examples of “avoiding + to-infinitive”. The following is
one of them:

… entertainers who were moving from country to country and avoiding to pay tax
altogether. (1998 Times, from Vosberg)

Examples like these clearly satisfy the horror aequi principle, though they are
only marginally attested owing to the restricted occurrence of the infinitival
complement itself with this verb. All in all, it is safe to state that the horror
aequi principle is largely fulfilled by the complementation patterns of avoid,
though only as a consequence in some cases.
In passing, it is also worth noting that examples of refuse, which was
discussed above in the previous chapter and which is almost consistently
followed by to-infinitives during the entire period of Modern English, also il-
lustrate how the horror aequi principle works in the choice of complements. As
far as Vosberg’s (2006: 163–166) data are concerned, the eighteenth-century is
interesting to discuss in this relation, when there are some notable number of
examples of –ing following the verb refuse. He demonstrates that the ex-
ceptional examples of –ing are clearly inclined to be evidenced after refuse
76 Chapter 3

occurring in the to-infinitival form, i.e. “to refuse + –ing”, thus satisfying the
horror aequi principle.
As concerns the examples of refuse in the OED quotations in the present
research, the use of the to-infinitival complement is relatively well-established
with this verb throughout the history of English. Still, it is interesting that
relevant examples of refuse are often evidenced in the –ing form, i.e. gerunds
or present participles, satisfying again the horror aequi principle in a humble
manner, namely only as a result, as in:

The keelmen of Sunderland made a stick, refusing to work.


(1768 The Annual Register)

Gillespie had been deposed … for refusing to assist in the disputed settlement of
Inverkeithing. (1854 Hugh Miller, My Schools and Schoolmasters)

In both these examples, the matrix verb refuse itself occurs in the –ing form,
which is followed by the usual form of the to-infinitive. The occurrence of the
matrix verb in the –ing form itself cannot be the key factor for the choice of the
infinitival construction, since the to-infinitive is the dominant form of com-
plementation in any case. It still merits attention, however, that refuse often
occurs in the context which is perfect for the choice of to-infinitives, satisfying
the horror aequi principle as a consequence. This is similar to the case of avoid,
in which the same verb, whose usual form of complement is the –ing form, is
strongly inclined to occur in the form of the to-infinitive, satisfying the horror
aequi principle as a consequence.
A total of 1,027 examples of refuse followed by a complement in the OED
include 144 examples of refusing,64 of which as many as 143 are followed by
to-infinitives. As mentioned above, the effect of the horror aequi principle
should never be overrated. The use of to-infinitives is frequent with the verb
refuse any way, whichever form the matrix verb may take, and this perhaps
overrides other principles like horror aequi. Still, I find it impressive that both
refuse and avoid know, in their own behaviours, where they can most appro-
priately fit and how they can accomplish a good harmony of language. The
principle is at least satisfied by a number of the examples of these verbs
available in the OED, although this does not necessarily mean that it always

64
This means that about 15% of the relevant examples of refuse are in the –ing form. This is
not a small proportion, though it may sound small when compared with the fact that more
than half of the examples of avoid are found in the to-infinitive form. Cf. Table 2 above (p.
74).
Verbs of implicit negation and gerunds 77

has the functioning force in the choice of complement types by refuse and
avoid.65 As hitherto repeated, the principle is satisfied only as a result.
Before concluding this chapter, I would like to make an additional com-
ment on the horror aequi principle. Although the present section refers to these
two particular verbs, it is much more widely applicable to various verbs, and
not restricted to the cases of avoid and refuse. As a matter of fact, Vosberg
(2006: 275) maintains that the principle is applicable to most of the verbs that
he has investigated in his study. Since much attention is paid to this phe-
nomenon throughout the discussion by Vosberg (2006), however, I will discuss
this issue only sparingly in the present study. This does not mean that it is
insignificant in the understanding of the complementation patterns of verbs of
implicit negation.

65
As expected, previous studies tend to stress the importance of semantic factors in relation
to the difference of complementation patterns between avoid and refuse, which I under-
stand are certainly relevant. Vespoor (1999: 521–524), for example, considers in syn-
chronic terms that the choice of –ing by avoid and the choice of to-infinitives by refuse
may be explained within the framework of the distinction between “intention in action” and
“prior intention”. The relationship between the verb avoid and the event expressed by its
complement is simultaneous (“intention in action”), whereas “refuse is related to a future
event” (“prior intention”). It is important to note at the same time, however, that different
complementation patterns from the ones in current English are available with both avoid
and refuse in the historical data. In other words, the relationship between a particular verb
and a particular kind of complement is not necessarily fixed in chronological terms. See 1.4.
above for further discussions on this matter.
4. Verbs of implicit negation and gerunds
with prepositions

4.1. Verbs of implicit negation and gerunds with prepositions

The present chapter explores in succession the following four verbs: prohibit,
prevent, hinder, and refrain. Like the verbs analysed in the preceding chapter,
they have undergone the first complement shift from that-clauses to to-
infinitives and then the second shift from to-infinitives to gerunds in their
historical development, although the first shift is not always visible in a clear
form as far as the OED dataset is concerned, which is simply due to the paucity
of relevant examples from earlier dates in the quotations data.
Unlike the verbs investigated in the preceding chapter, however, they all
clearly display a further stage of the development of complementation, i.e. the
extensive employment of prepositions with the gerund as in “prohibit + object
+ from + –ing” and “hinder + from + –ing”. On the basis of the analysis of the
four verbs under consideration, the present study concludes that the con-
struction with a preposition followed by –ing is most probably a slightly later
development than the construction with the plain gerund. Although usage
guides and grammars often recommend the employment of prepositions with
these verbs or even consider it to be imperative in some cases, there is certainly
a stage in the history of English where simple gerunds were usual with them.
The construction with from grew, probably at the cost of the construction with
gerunds only in the historical process.
The discussion in the following sections deals with prohibit (4.2.), prevent
(4.3.), hinder (4.4.), and refrain (4.5.) in turn, highlighting their historical
developments leading to the expanded use of gerunds with prepositions as
observed in Present-day English.
80 Chapter 4

4.2. The historical development of prohibit


4.2.1. Preliminary remarks

Of the four verbs under consideration in this chapter, the first to be analysed is
prohibit. It represents, in my view, one of the most prototypical cases of the
historical development that various English verbs can possibly follow or may
have already followed to date. Prohibit experiences the complement shift from
that-clauses to to-infinitives (the first shift) and then the shift from to-
infinitives to gerunds with or without the preposition from (the second shift) in
its historical development. Consequently, the same verb in Present-day English
most typically presents the following two types of gerundial constructions, i.e.
one with from and one without:

The Assistant District Attorney tossed out the charges saying that although the
law prohibits Klansmen from appearing in public, the faces of the 24 were visible.
(1973 Black Panther, 1 September)

These band shallows are out of bounds for boats, because reservoir rules usually
prohibit them approaching within 50 metres of the bank, so anchoring-up near the
bank is “out”. (FLOB, F)

While the person to be prohibited from doing something is expressed in the


objective form in the second of the above examples, he or she can also be ex-
pressed in the genitive form in Present-day English, as exemplified by the
following sentence:

It may be described in general terms as a duty which requires a partner to act


openly, fairly and honestly in all partnership dealings and which prohibits his
obtaining any private benefit at the expense of the firm.
(Charles Bonney, Solicitors’ Partnerships: The Law in Practice, from BNC)

In relation to these gerundial constructions attested in contemporary Eng-


lish, Wood (1981: 213) remarks that the “prohibit + object + from + –ing” is
the correct one, and that prohibit directly followed by the gerund, i.e. the
gerundial construction without from, is to be avoided in use, although it is
heard on occasions. On the other hand, the Oxford Advanced Learner’s Dic-
tionary (s.v. prohibit) admits both types of gerundial constructions, i.e. the one
with from and the one without.
Moreover, there is another contrast which is more frequently commented
Verbs of implicit negation and gerunds with prepositions 81

upon in usage guides and handbooks, namely the relationship between the in-
finitival and gerundial constructions following prohibit. Buchfield (1998: 627)
refers to the “prohibit + to-infinitive” construction and says that it was dis-
approved of by Fowler (1926) and that it is “rare and probably obsolescent” in
contemporary English. Similarly, Greenbaum and Whitcut (1996: 375) do not
admit to-infinitives and state that “prohibit someone from doing it [an action],
but not someone to do it”, whereas Partridge (1994: 236) simply considers the
infinitival construction archaic. The OED (s.v. prohibit) and Visser (1963–
1973: §2078) hold the same position as Partridge and maintain that the
to-infinitive construction is archaic. Turning to the OED quotations, the most
recent example of “prohibit + to-infinitive” is the following, which goes back
to the nineteenth century:

Their rules prohibit them to work along with scab switchmen.


(1893 Columbus (Ohio) Disp., 27 September)

Unfortunately, Herbst, Heath, Roe, and Götz (2004) do not provide an entry for
prohibit in their voluminous work.
In analysing the historical data, it is also necessary to take into con-
sideration an additional construction, i.e. prohibit followed by that-clauses with
or without expletive negation. See the following examples which illustrate the
construction:

They did prohibit that no man shoulde ... sell openly ... wine of Candie or Spaine.
(1557 Sir Thomas North, Gueuara’s Diall of Princes)

He also prohibited that any thinge shuld be radde or spoken, reprocheable or


blasphemous to god. (1531 Thomas Elyot, The Boke Named the Gouernour)

The complement or the subordinate clause in the first of these examples in-
cludes expletive negation (i.e. no man shoulde …), whereas the one in the
second example does not.
As repeated several times in the above discussions (see 1.1. among others),
expletive negation of this kind was common or even compulsory in some cases
in Old and Middle English, but later experienced a gradual recession in the
history of English. The two sixteenth-century texts cited above present an
interesting contrast in this respect, one including expletive negation and the
other not. The second example, which is devoid of expletive negation, includes
82 Chapter 4

any instead (i.e. any thinge shuld …) and signals that the subordinate clause is
non-assertive in nature.
As the above survey shows, the “from + –ing” construction seems to be
more or less established with prohibit in Present-day English. At the same time,
however, it emerges from the accounts so far given that the same verb has
undergone an interesting history of trying out various different constructions in
its development. The following discussion is concerned with the types of con-
structions mentioned thus far—prohibit followed by that-clauses with or with-
out expletive negation, to-infinitives, and gerunds with or without the pre-
position from—, and sees how their relationships have changed in the history
of English.

4.2.2. The rise of to-infinitives and gerunds

The overall development of the complementation patterns of prohibit is rela-


tively straightforward. This is well-depicted by the figure below, which dis-
plays the raw frequencies of three different types of complementation
dominated by prohibit in the OED quotations:

40

30

20

10

0
16th c. 17th c. 18th c. 19th c. 20th c.

that-clauses to-infinitives gerunds

Figure 11. Prohibit and three patterns of complementation in the OED (raw frequencies)

Unlike verbs like forbid, whose occurrences go back to a much earlier period
of English, relevant examples of prohibit are available only from the sixteenth
century onwards as far as the OED quotations are concerned.66 Although
66
Although prohibition is attested in the fourteenth century, examples of the verb prohibit are
evidenced from the fifteenth century onwards in the OED. To find examples with some
Verbs of implicit negation and gerunds with prepositions 83

examples from relatively early periods are not numerous in this graph, one can
clearly observe in Figure 11 the lingering occurrences of that-clauses in the
sixteenth century. Simultaneously, the same century already signals the rise of
to-infinitives as a complement of prohibit. In other words, the complement
shift from that-clauses to to-infinitives or the first complement shift is already
taking place in the earliest data of prohibit, which is later on overtaken by the
second complement shift or the rise of gerunds. From the seventeenth century
onwards, the occurrence of gerunds is increasingly noticeable, and by the time
of the twentieth century, they virtually become the sole form of complements
dominated by prohibit. Thus, the two types of complement shifts are attested in
a fairly transparent form with this verb in the above graph. The only
unfortunate aspect, at least for the purpose of the present study, of the historical
development of prohibit is that relevant examples are unavailable in the
fifteenth-century data in the OED, since the employment of that-clauses may
have been more common during the Middle English period. If a larger number
of examples were available there, the first complement shift would be
represented in an even more perspicuous form.
Furthermore, the dearth of earlier relevant examples in the OED should
make the analysis of expletive negation difficult, at least in theory, since the
phenomenon itself recedes in a gradual manner from later Middle English to
early Modern English. Unexpectedly, however, expletive negation is well
preserved in the limited number of subordinate clauses available with prohibit
in the OED dataset. It provides eight examples of that-clauses dominated by
the verb under consideration,67 of which seven include negation that is se-
mantically unnecessary, illustrating the occurrence of expletive negation. The
only exception is the following, which was already cited under 4.2.1.:

He also prohibited that any thinge shuld be radde or spoken, reprocheable or


blasphemous to god. (1531 Thomas Elyot, The Boke Named the Gouernour)

The subordinate clause of prohibit in this example includes non-assertive any


instead of expletive negation, as mentioned above.

form of complements, one has to wait for another century, as far as the OED quotations are
concerned.
67
Later examples are rather dubious as to whether the that-clause under consideration is the
nominal one. It may be adverbial, in which case negation in it is necessary. It is impossible
to eliminate this ambiguity in the interpretation of expletive negation in general. See Note 6
in Chapter 1 for further discussion on this matter.
84 Chapter 4

An example of expletive negation was also cited above under 4.2.1., but I
will provide an additional one in the following to illustrate the point:

Helyas through the power of God, did prohibit that it should not rayne.
(1561 John Daus, tr., Bullinger’s Hundred Sermons vpon the Apocalips)

In this quotation, the negative adverb not in the subordinate clause introduced
by that is unnecessary, at least from the semantic perspective. In view of the
above two examples, the sixteenth century may have been the key in respect of
the disappearance of expletive negation with this verb.
As a matter of fact, the good retention of expletive negation in the ex-
amples of that-clauses dominated by prohibit suggests that to-infinitives quick-
ly penetrated into the domain of complementation before the weakening of
that-clauses was obvious. It is presumably for this reason that the dataset of the
OED does not provide any examples of but-clauses used after prohibit, which
are usually encountered at the time of the marginalisation of that-clauses es-
pecially when that-clauses try to linger until later dates. As mentioned above
(see 1.1. above), but-clauses were a useful device which mediated the loss of
expletive negation and the lingering of finite clauses. It preserves the tone of
negation on the one hand and it introduces a finite clause on the other hand as
the conjunction that does.
In the case of prohibit, however, the complement shift from that-clauses to
to-infinitives was quite successful or even prompt, and therefore, it was per-
haps not really necessary for the same verb to resort to the use of but-clauses,
an alternative construction to that-clauses with expletive negation, although
this does not necessarily imply the absolute absence of but-clauses outside the
OED data. During the early Modern English period, prohibit already provides
abundant examples of to-infinitives instead, as illustrated by:68

The reading of history prohibyteth reprouable persons to do mischeuous dedes.


(1523 John Bourchier Berners, Sir John Froissart, I, Preface)

A litle Balk of Sand cast up, the wich at low waters prohibitih the Se to cum about.
(1538 John Leland, The Itinerary)

68
Although expletive negation is essentially a phenomenon of finite clauses, it is occasionally
attested in infinitival constructions, as in: It cannot effectually prohibit the Heart not to
move, or the Blood not to circulate (a1667 Sir Matthew Hale, The Primitive Origination of
Mankind)
Verbs of implicit negation and gerunds with prepositions 85

The Queene by Proclamation prohibited any new dwelling houses to be built, ...
vpon paine of imprisonment, and losse of the stuffe brought for the building.
(1630 Robert Norton, tr., William Camden’s Annals: The True and Royal History
of Elizabeth, Queen of England)

The infinitival constructions as illustrated by these examples are then to be


replaced by gerunds in later dates, the discussion of which the following
section will concentrate upon.

4.2.3. Gerunds with or without the preposition from

As argued in the previous section, the rise of gerunds is a feature evidenced


slightly later than the rise of to-infinitives in the historical development of
prohibit. The present section investigates further details of the increasing use of
gerunds by the same verb in the historical context.
As the examples cited below illustrate, gerunds following prohibit may or
may not be accompanied by the preposition from:

The London Government Act of last year ... prohibited women serving as alder-
women or councillors on borough Councils. (1900 Daily News, 24 May)

The Assistant District Attorney tossed out the charges saying that although the
law prohibits Klansmen from appearing in public, the faces of the 24 were visible.
(1973 Black Panther, 1 September)

The gerund in the first of the above examples stands alone, whereas the gerund
in the second is preceded by the preposition from, illustrating the “prohibit +
object + from + –ing” construction. Although gerunds with or without
prepositions are often treated together in existing grammars, the expanded use
of prepositions—in this case from—seems to be a further developed stage of
gerundial constructions, as far as the chronological development of com-
plementation in English is concerned. The figure that follows, depicted on the
basis of the raw frequencies of the two types of gerunds in the OED quotations,
clarifies this point:
86 Chapter 4

30

20

10

0
16th c. 17th c. 18th c. 19th c. 20th c.

gerunds only from + gerunds

Figure 12. Prohibit and gerunds with or without from in the OED (raw frequencies)

As discussed in the preceding section, the expanded use of the gerund is attest-
ed with this verb from the seventeenth century onwards. At the early stage of
this development, gerunds most frequently stood alone without the preposition
from, and this is clearly shown in the above graph. As far as the OED quota-
tions are concerned, the occurrence of from suddenly becomes notably com-
mon in the eighteenth century, and exceeds the gerund without from in the
following century. This trend is continuous thereafter, and by the time of the
twentieth century, most examples of gerunds are accompanied by the pre-
position from with this verb. Thus, the rise of the gerundial construction with
from is noticeable during the late Modern English period with prohibit.
In the OED quotations, the use of gerunds without from in recent years are
often confined to the type exemplified by the following quotation, where the
gerund is a nominal one, typically preceded by the definite article the:

Canada prohibits the broadcasting of “spot ads”.


(1934 James Rorty, Our Master’s Voice Advertising)

Nominal gerunds of this kind may be almost equivalent in nature to simple


nouns occurring as the object of the verb (see also Note 42 in Chapter 2). When
these examples are excluded, relevant examples of prohibit are in usual case
accompanied by the preposition from in the twentieth century, although in
FLOB, examples like the following are still attested:69

69
FLOB provides only a few relevant examples of prohibit in all, none of which includes the
preposition from. This is a matter of surprise in a way, since, as mentioned above, gram-
Verbs of implicit negation and gerunds with prepositions 87

Peter Robinson points out that falconry is probably illegal under the 1979 Euro-
pean Directive on the Conservation of Wild Birds which prohibits keeping and
trading in wild birds and which gives no express authority for participation in
falconry. (FLOB, E)

This is a clear example where the verb prohibit is followed by the simple
gerund without from.
To strictly clarify the most recent usages of prohibit, further investigation
is inevitably necessary. As far as the historical development of prohibit is
concerned, however, the overall trend is relatively transparent in Figure 12,
which exhibits the situation of the OED quotations. The expanded use of ger-
unds was witnessed slightly later than the expanded use of to-infinitives, and at
the early stage of the development of this verb, gerunds often stood alone. The
employment of the preposition from was increasingly common with prohibit in
later Modern English, and in Present-day English the verb followed by “from +
–ing” is considered to be the standard form.

70
4.3. The historical development of prevent
4.3.1. Preliminary remarks

The second verb to be investigated in the present chapter is prevent, which


displays a similar path of development to the path followed by prohibit. About
this verb, Burchfield (1998: 622) states that “[t]here are three competing con-
structions when the verb prevent is followed directly or indirectly by a gerund
in –ing”, and gives the following list:71 (a) “prevent + possessive + –ing”
(OED 1841–) (e.g. Either the fire-guard or his mother prevents his reaching
the fire, H. T. Lane, 1928);72 (b) “prevent + object + –ing” (OED 1689–) (e.g.
Two women climb up the iron bars, which are meant to prevent people or
animals falling under the tram, J. Berger, 1972); and (c) “prevent + object +

mars and reference guides in Present-day English usually recommend the use of the form
“from + –ing” after prohibit.
70
An earlier version of the section on prevent appeared as Iyeiri (2008c). Peter Lang kindly
granted permission to use it in the present study.
71
In the interest of consistency throughout the sections under prevent, I have changed the
order of the three patterns at issue from the original list. Pattern (a), for instance, appears
last in Burchfield (1998: 622).
72
I have quoted only the first of the examples provided by Burchfield (1998: 622) under each
pattern.
88 Chapter 4

from + –ing” (OED 1711–) (e.g. Sleeman had tried to prevent a widow from
committing suttee, R. P. Jhabvala, 1975).73
Concerning the rise and fall of these three patterns, Burchfield refers to (a)
only, and comments that it “is beginning to fall into disuse” in Present-day
English. As for the remaining patterns, (b) is often commented upon in the
literature. Weiner and Delahunty (1994: 159), for example, regard (b) as in-
formal. Considering these accounts, it is most likely that “prevent + possessive
+ –ing” is on the decline on the one hand and that “prevent + object + –ing” is
increasingly acknowledged on the other hand, though it may still be regarded
as informal, in contemporary English. While the Oxford Advanced Learner’s
Dictionary admits both these constructions, Herbst, Heath, Roe, and Götz
(2004: 614) offer the construction “prevent + object + –ing”, but not the form
“prevent + possessive + –ing”. In the following sections, I will analyse the
quotations data of the OED and see how various patterns of complementation
observed in the past affect the present-day usage of prevent.
In this relation, it is particularly important to make clear that the present-
day usage of prevent in this study means the usage of the same verb in British
English only, since it is widely acknowledged that there is a significant dif-
ference between British and American English in terms of the usages of the
verb under consideration. This is striking in light of the fact that Görlach
(1987: 53–55), who comments on various conservative and progressive fea-
tures of American English in general, concludes that syntactic differences
between the two varieties are rather restricted. In American English, “prevent +
object + –ing” is extremely limited even in informal settings today, while it is
commonly attested in British English. Andersson (1985: 101) among others
notes that the construction without the preposition from tends to be found in
British English only. Furthermore, Mair (2002: 112–115) displays the statistical
data from the four corpora LOB, FLOB, Brown and Frown, and proves that
“prevent + object + –ing” is expanding in current British English and that its
use is extremely limited both in the 1960s and the 1990s in American English.
See also Mair (1999: 176–180), who provides additional data from The

73
Burchfield (1998: 622) gives incorrect dates for the first quotations of (b) and (c) in the
OED, which I have modified. The dates given by the OED are, however, not absolute in
any case, since they can easily be antedated. Visser (1963–1973, §2103), for instance,
gives an example of 1696 for the (a) type, while Visser (1963–1973, §2092) gives an
example of 1592 for (b). Furthermore, the electronic text search of the OED itself yields an
example of 1600 for the (a) type, an example of 1614 for (b), and an example of 1641 for
(c).
Verbs of implicit negation and gerunds with prepositions 89

Guardian (British English) and Miami Herald (American English) to


corroborate this point.
Frown, however, gives the following single instance of the “prevent +
object + –ing” form in American English:

No small group of people ever win a major war; but sometimes quite small
groups can prevent it being lost. (Frown, J)

Furthermore, examples like the following can relatively easily be found in


eighteenth-century American texts according to Mair (2002: 115), from whose
work the following examples are cited:74

… every Law we have hitherto made to prevent this Deluge of Wickedness


overwhelming us. (Franklin 1736–1757, 359: 1, from Mair)

So that, while the Enemy is the Heart of the Country, Cavils prevent any Thing
being done for its Relief. (Franklin 1757–1775, 507: 1, from Mair)

This is, however, a matter to be investigated in a separate piece of research on


different occasions. The essential focus of the present research is placed upon
the historical development of the patterns of complementation in British Eng-
lish (see 1.3. in the Introduction).

4.3.2. The establishment of gerundial constructions in the history of English

While Present-day English shows the vacillation among the three types of ge-
rundial constructions of prevent as mentioned above, there is a stage prior to
this in the history of English. As Figure 13 on the following page reveals, the
establishment of the use of gerunds after prevent is attested only in the course
of the seventeenth century as far as the quotations in the OED are concerned.
The graph shows the proportion of each type to the total of the three types of
constructions of this verb:75

74
It is also relevant to mention that The Columbia Guide to Standard American English
(Wilson 1993: 343) admits all of the three patterns under consideration. For a thorough
survey of previous studies on the American usage of prevent, see Tajima (1995: 3).
75
I admit that the sixteenth-century situation of this graph is rather unstable from the sta-
tistical perspective, since there are only ten relevant examples of prevent there. The general
trend of the syntactic development of prevent is, however, unequivocal in this graph.
90 Chapter 4

100

80

60

40

20

0
16th c. 17th c. 18th c. 19th c. 20th c.

that-clauses to-infinitives gerunds

Figure 13. Prevent and three patterns of complementation in the OED (%)

While this graph illuminates the constant increase of the use of gerunds by
prevent during the early Modern English period, it also shows that there are
still examples of prevent followed by that-clauses and to-infinitives during the
sixteenth and seventeenth centuries. See the following examples from the early
Modern English period, which illustrate the older constructions with that-
clauses and to-infinitives:

Doth it not stand her in hand to preuent that the number of catholiks do not
increase?
(1600 William Watson, A Decacordon of Ten Quodlibeticall Questions Concern-
ing Religion and State)

To preuent those Thracian theeves that they should not hide themselues within
their peakish holes … and ordinarie couert musets.
(1600 Philemon Holland, Livy’s Romane Historie)

To prevente the same [bone] to be … putrefacted and corrupted.


(1597 A. M. tr., Guillemeau’s Frenche Chirurgerye or All the Manualle Opera-
tions of Chirurgerye)

Whether that-clauses as in the first two of the above examples are a genuinely
nominal one is an open question. Although they may be a final clause with the

Relevant examples are abundant from the seventeenth century onwards, which corroborates
the establishment of the use of gerunds with this verb during the early Modern English
period.
Verbs of implicit negation and gerunds with prepositions 91

sense ‘so that’, however, this does not matter at least for the purpose of the
discussion in the present study. The important and relevant point here is that
examples of this kind are increasingly restricted in use in the history of the
English language, whatever the nature of that-clauses may be. The later de-
velopment of the verb prevent is clearly depicted in Figure 13. The employ-
ment of gerunds is increasingly predominant with this verb from the seven-
teenth century onwards.

4.3.3. Three types of constructions with gerunds in later Modern English

As discussed in the previous section, the established use of gerundial con-


structions is evidenced only from the seventeenth century, before which the
patterning of complementation of the verb prevent was still unstable, in that
other types like that-clauses and to-infinitives were available. The changeover
from that-clauses to to-infinitives or the first shift is an interesting aspect of the
shifts of complements in the history of English, but unfortunately, relevant ex-
amples are not at all copious with prevent in the OED quotations. Hence, the
focus in the following discussion will be placed upon the several hundred years
from the seventeenth century, when the second stage of the development of
prevent is represented and when relevant examples are much more abundant in
the OED dataset.
From the seventeenth century onwards, three different types of gerundial
constructions compete, and as mentioned in the introductory section of the verb
at issue (see 4.3.1.), this competition has not been completed yet even in
contemporary English, which yields the following three types of constructions:
(a) “prevent + possessive + –ing”, (b) “prevent + object + –ing”, and (c)
“prevent + object + from + –ing”. Some historical examples illustrating the
three types are cited below from the OED quotations:

He built a house without his Camp for all strangers …, whereby he prevented
their sneaking into his Camp. (1656 North’s Plutarch)

A free Confession … easily prevents a little Error growing to a great Evil.


(1718 John Fox, The Wanderer)

Nothing but an invincible Resolution … could have prevented me from falling


back to my Monosyllables. (1714 Joseph Addison, The Spectator, No. 556)
92 Chapter 4

The first of the above examples illustrates the (a) type, the second (b), and the
third (c).
The relationship among the three patterns under consideration is relatively
well discussed in existing studies, especially in respect of Present-day English.
As mentioned above, Burchfield (1998: 622) argues that “prevent + possessive
+ –ing” (i.e. (a)) is beginning to decline in today’s English. This is confirmed
by Heyvaert, Rogiers, and Vermeylen (2005: 84), who provide the table
below:76

Table 3. Prevent and three gerundial constructions in previous studies


prevent + possessive prevent + object prevent + object +
+ –ing + –ing from + –ing
Kirsten 18th century (25) 30% (8) 10% (49) 60%
Kirsten 19th century (48) 53% (5) 6% (37) 41%
van Ek 1950–1964 (2) 5% (14) 38% (21) 57%
COBUILD corpus (32) 6% (120) 23% (367) 71%
(from Heyvaert, Rogiers, and Vermeylen 2005: 84)

Table 3 is based upon the analysis by Heyvaert, Rogiers, and Vermeylen (2005)
of the COBUILD corpus (around 1990 and after) as well as the data provided
by Kirsten (1957) and those by van Ek (1966). Although the dataset in this
table is a collection of various materials which may be inconsistent from the
stylistic perspective, the table reveals in a most explicit manner the dramatic
decline of the “prevent + possessive + –ing” construction in the twentieth cen-
tury. It is perhaps safe to conclude from this table that this pattern has indeed
been on the recession in recent years as Burchfield claims.
One thing that is not so evident in this table is whether “prevent + object +
from + –ing” is expanding at the cost of “prevent + object + –ing”. As far as
Table 3 is concerned, the construction with the preposition from is most likely
on the increase in contemporary English, but at the same time, there are other
studies which make contentions to the opposite effect. Mair (2002: 112), for
instance, investigates the two constructions of “prevent + object + –ing” and

76
In the interest of consistency throughout this section, I have slightly modified the presenta-
tion of this table from Heyvaert, Rogiers, and Vermeylen (2005: 84). Needless to say, all
the figures including percentages remain the same.
Verbs of implicit negation and gerunds with prepositions 93

“prevent + object + from + –ing” in the LOB and FLOB corpora, and provides
the following results:77

Table 4. Prevent and three gerundial constructions in LOB and FLOB


prevent + object + –ing prevent + object + from + –ing
LOB (1961) 7 34
FLOB (1991/1992) 24 24
(from Mair 2002: 112)

Unlike Table 3 on the previous page, Table 4 hints at the possibility that the
pattern “prevent + object + –ing” (rather than “prevent + object + from + –ing”)
is on the increase in Present-day English. It was not as frequent as the
construction with the preposition from in 1961, while it is already as frequent
in 1991/1992. The question is where the difference between the data of
Heyvaert, Rogiers, and Vermeylen (2005) and the data of Mair (2002) arises.
Although the ultimate answer to this question can be given only after
further analyses of Present-day English are conducted, the historical data of the
OED yield an important suggestion in this concern. And this is what the present
study is interested in discussing. There is a perspicuous difference in terms of
the pace in the historical development of the three constructions at issue,
depending upon whether the “possessive” or the “object” before the gerund is a
personal pronoun or not. In the following, I will present the raw frequencies of
the three gerundial patterns found in the OED, Figure 14 showing the case of
personal pronouns and Figure 15 displaying the case of nouns.78 Examples of
“prevent + her + –ing” (nine examples in all) are all excluded from this
analysis, since it is practically impossible to tell whether her in this
construction is possessive or objective. See the two graphs on the following
page:

77
In the interest of consistency, I have slightly modified the presentation of Table 4 from
Mair (2002: 112). It is a matter of course, though that the frequencies under each pattern
have not been changed.
78
Pronouns other than personal pronouns are treated under the category of nouns, considering
the fact that they have lost the case distinctions which personal pronouns retain.
94 Chapter 4

140
120
100
80
60
40
20
0
17th c. 18th c. 19th c. 20th c.

prevent + possessive + -ing prevent + object + -ing prevent + object + from + -ing

Figure 14. Prevent with personal pronouns and the three gerundial constructions in the
OED (raw frequencies)

300
250
200
150
100
50
0
17th c. 18th c. 19th c. 20th c.

prevent + possessive + -ing prevent + object + -ing prevent + object + from + -ing

Figure 15. Prevent with nouns and the three gerundial constructions in the OED (raw
frequencies)

As repeatedly mentioned above, it is illogical to compare and contrast the raw


frequencies of different centuries under the same type of constructions, since
the number of quotations per century is not constant in the OED dataset. Al-
ternatively, attention has to be paid to the relationship among the three patterns
in each century, which makes it evident that the tendencies of personal pro-
nouns (Figure 14) and those of nouns (Figure 15) are markedly contrastive.
The most outstanding distinction between the two graphs is that “prevent +
possessive + –ing” as illustrated by the example below is relatively well
preserved in the case of personal pronouns, whereas the same construction
Verbs of implicit negation and gerunds with prepositions 95

recedes much earlier when nouns are involved instead of personal pronouns:

Their Leaves … may be tied in knots, which will prevent their spindling.
(1707 John Mortimer, The Whole Art of Husbandry)

As a matter of fact, “prevent + possessive + –ing” is the predominant form


until the nineteenth century in the case of personal pronouns, and even in the
twentieth century, it is still observed to a noticeable extent.
This is most likely a reflection of the general tendency of using possessive
pronouns rather than objective pronouns before gerunds, a tendency not re-
stricted to the verb prevent but observed across the board in the history of
English. Denison (1998: 269) maintains that “with personal pronouns a pref-
erence for the genitive [i.e. possessive] form was often vociferously expressed
in the prescriptive tradition” and refers to Dekeyser (1975: 180–181), who
gives the proportions ranging from 2.3% to 6.7% for nongenitive pronouns in
the nineteenth century. Furthermore, even today, Quirk, Greenbaum, Leech,
and Svartvik (1985: §15.12) state that “[i]n general, the genitive is preferred if
the item is a pronoun, the noun phrase has personal reference, and the style is
formal”, although, according to Visser (1963–1973: §1102), the use of
objective pronouns before the gerund is increasingly frequent “from about the
middle of the nineteenth century”. In addition, there are contemporary gram-
matical reference guides and handbooks which even state that the use of the
possessive is the norm (e.g. Wood 1981: 136).
Thus, the statistical data provided by Heyvaert, Rogiers, and Vermeylen
(2005) and Mair (2002) need further subdividing to present a clearer picture of
the development of gerundial constructions with prevent in the twentieth cen-
tury. More specifically, nouns and pronouns need to be treated separately.
Unfortunately, Mair (2002) analyses nouns and personal pronouns together.
Heyvaert, Rogiers, and Vermeylen (2005), on the other hand, restrict their
analysis to personal pronouns, but compare their data with those presented by
Kirsten (1957) and van Ek (1966), who investigate nouns and pronouns
inclusively.79 Thus, the entire situation of the gerundial constructions of pre-
vent needs to be reanalysed and reconsidered.
Once the division is made between the cases of personal pronouns and
nouns, some interesting points are revealed about the relationship between
“prevent + object + –ing” and the construction with the preposition from. It

79
Kirsten (1957) does not state this, but I have confirmed this point by double-checking some
of the works investigated by him.
96 Chapter 4

seems to be relatively stable in the case of nouns from the nineteenth century
onwards (see Figure 15), “prevent + object + –ing” occurring about half as
frequently as the construction with from. By contrast, a much more radical
change is taking place when the subject of the gerund is a personal pronoun
(see Figure 14), in that the relative frequency of “prevent + object + –ing” to
that of “prevent + object + from + –ing” seems to be rising.
The question arises as to whether the expanded use of “prevent + object +
–ing” in the case of personal pronouns is caused by the decline of the older
possessive construction alone or also at the cost of the decline of the con-
struction with from. This is not an easy question to answer, but Figure 16,
which indicates the changing proportions of the three patterns to the total, will
provide a hint:

100
80
60
40
20
0
17th c. 18th c. 19th c. 20th c.

prevent + possessive + -ing prevent + object + -ing


prevent + object + from + -ing

Figure 16. Prevent with personal pronouns and the three gerundial constructions in the
OED (%)

Figure 16 suggests that the rise of “prevent + object + –ing” was most plausibly
triggered by the drop of the possessive construction from the eighteenth cen-
tury to the nineteenth century. By contrast, the construction with the pre-
position from also seems to be on the constant increase from the seventeenth
century. In fact, it is on the increase even in the twentieth century. In other
words, the gap caused by the decline of the possessive construction is filled not
only by “prevent + object + –ing” but also by “prevent + object + from + –ing”.
Still in other words, the decline of the possessive construction is likely to be
the key factor in the expansion of “prevent + object + –ing” with personal
pronouns, at least as far as the OED quotations are concerned.
Incidentally, the proportion of the “object + –ing” construction with a
Verbs of implicit negation and gerunds with prepositions 97

personal pronoun is not as low as Dekeyser (1975: 180–181) states about the
general tendency of gerunds, in the nineteenth century, as far as the case of
prevent is concerned. As mentioned above, he gives the proportions ranging
from 2.3% to 6.7% to the objective case as opposed to the possessive one,
while the proportion of the “object + –ing” construction in the nineteenth
century in Figure 16 is about 8.6%, which means that the proportion will be
even larger when the construction with from is excluded from the data and only
“object + –ing” and “possessive + –ing” are considered.
The whole issue may be a matter of differences due to different verbs.
Alternatively, stylistic factors may be relevant, which is beyond analysis in the
present study since the OED dataset draws materials from various types of
texts with some emphasis upon literary ones. All one can confirm in this
relation is that the possessive type is more favoured in formal style than in
informal one, according to previous studies. See Quirk, Greenbaum, Leech,
and Svartvik (1985: §15.12) and Huddleston and Pullum (2002: 1192) for
further discussion on this matter.

4.3.4. Is “prevent + object + –ing” an elliptical form of “prevent + object +


from + –ing”?

While the discussion by Heyvaert, Rogiers, and Vermeylen (2005) is based


upon the assumption that the employment of the objective nouns and pronouns
in the form “prevent + object + –ing” is a way to indicate the subject of the
–ing form, there is also a widely-prevalent view that the pattern is a short form
of “prevent + object + from + –ing”. In other words, it is also considered to be a
form produced through the ellipsis of from in some previous studies. The OED
(s.v. prevent), for example, takes this position, as noted by Visser (1963–1973:
§2092) in his description of the verb prevent. Among more recent studies,
Dixon (1991: 194) supports the same contention.80
As I discuss later in this section, both factors are perhaps relevant, but for
the moment, I would like to take the side of Heyvaert, Rogiers, and Vermeylen
(2005), since their argument seems to be better justified in view of the
historical development of the relevant constructions. Figure 16 above clearly
80
Dixon (1991: 236–237) also pays attention to the meaning differences between the ge-
rundial construction with from and that without, stressing that the effect of the performer is
indirect in the case of the former and more direct in the latter. See also 1.4. above for some
detailed discussions on the possible existence of meaning differences in different forms of
complementation.
98 Chapter 4

reveals that the initial rise of “prevent + object + –ing” was triggered by the
decline of “prevent + possessive + –ing”. The graph also shows the constant
rise of both “prevent + object + –ing” and “prevent + object + from + –ing” in
the history of English. It is, therefore, more likely that “prevent + object +
–ing” develops at the cost of the decline of the possessive construction than at
the cost of the decline of the construction with from.
Furthermore, this view is consistent with what Rohdenburg (1996: 151)
calls the Complexity Principle, which runs as follows: “In the case of more or
less explicit grammatical options the more explicit one(s) will tend to be favor-
ed in cognitively more complex environments”. If “prevent + object + –ing” is
the short and reduced form of “prevent + object + from + –ing”, and if the
Complexity Principle holds for this case, the form with from, which is the more
explicit option, would be favoured by nouns, since nouns are more complex
than pronouns from a cognitive perspective, according to Rohdenburg (1996:
152–159). In other words, if ellipsis is the key factor in the development of
“prevent + object + –ing”, it would take place more rapidly with pronouns than
with nouns, since pronouns are by nature less complex than nouns. However,
the historical development of “prevent + object + –ing” displays the opposite
tendency. The construction develops much more quickly with nouns than with
personal pronouns (see Figures 14 and 15 above). Accordingly, the assumption
that “prevent + object + –ing” is the short form of the pattern with from is
refuted. In other words, one needs to consider another factor in relation to the
expanded use of “prevent + object + –ing”, i.e. the decline of the “possessive +
gerund”, which is in fact more relevant.
Incidentally, the Complexity Principle here is very similar to the Quantity
Principle proposed by Givón (1991: 87), which consists of three principles: (a)
“A larger chunk of information will be given a larger chunk of code”; (b) “Less
predictable information will be given more coding material”; and (c) “More
important information will be given more coding material”. As a consequence,
Givón’s principle would make the same prediction. Nouns are a larger chunk,
less predictable, and more important as a piece of information than pronouns,
and therefore nouns are more likely to occur in the form “prevent + object +
from + –ing” if it is in competition with “prevent + object + –ing”. Once again,
the assumption that “prevent + object + –ing” is the short form of the pattern
with from is refuted.
The hitherto-given discussion does not lead to the conclusion, however,
that the omission of from is wholly irrelevant to the expanded use of “prevent +
object + –ing”, especially in the context of contemporary English. In the end,
the almost complete obliteration of the construction “prevent + possessive +
Verbs of implicit negation and gerunds with prepositions 99

–ing” results in the situation in which only the two forms “prevent + object +
–ing” and “prevent + object + from + –ing” are available. In other words, the
expansion of one form is possible only at the cost of the other under such
circumstances. Given the similarity between the two forms, it is not a matter of
surprise if language users regard the former construction as a short and reduced
form of the latter, as a number of linguists have done. This widely-held view
can further accelerate the use of “prevent + object + –ing”, perhaps in col-
loquial contexts first and then in other contexts later, although I do not know at
this stage to what extent this has indeed taken place in practice. In any case, the
relationship between “prevent + object + –ing” and “prevent + object + from +
–ing” needs to be paid attention, once the employment of “prevent + possessive
+ –ing” has declined to a noticeable extent in the history of English.81

4.4. The historical development of hinder


4.4.1. Preliminary remarks

The verb hinder in Present-day English is hardly discussed in usage guides and
grammars. Herbst, Heath, Roe, and Götz (2004) do not spare an entry for this
verb, either. Needless to say, the Cambridge Advanced Learner’s Dictionary
and Macmillan English Dictionary for Advanced Learners provide an entry for
the same verb, but they do not mention how the verb is used in current English.
Despite the restrictedness of the treatment of this verb in existing studies,
however, it is generally known that it ordinarily occurs in Present-day English
in the “hinder + object + from + gerund” construction, as the Oxford Advanced
Learner’s Dictionary (s.v. hinder) shows. See the following illustrative ex-
amples quoted from FLOB and BNC:

…although specifically intended to hinder nobles from retaining royal justices as


clients or followers, … (FLOB, F)

Like the sugar tonics, the glycerophosphates have no long-term beneficial effects
on health, and indeed perhaps hinder the body from restoring its own efficient
breakdown of glucose to release energy. (New Scientist, from BNC)
81
See Vosberg (2006: 152–153), who investigates the verb prevent and who shows that the
construction without from occurs far more frequently when the object of the verb is the
pronoun it than in the other cases. It seems likely that there are some linguistic conditions
related to the choice between the two forms of complementation, which are certainly worth
investigating in future studies.
100 Chapter 4

Both these examples in Present-day English illustrate the form “hinder + object
+ from + gerund”.
Side by side with this construction, hinder in the BNC also shows the
simple gerund construction without from, as exemplified below, although its
existence is hardly mentioned in the literature:

… which prevent or hinder them making use of the educational facilities


generally provided. (Ideas in Action programmes, from BNC)

She was born with a blockage in her nose which hindered breathing, so an airway
is being kept open by a small tube. (Daily Telegraph, from BNC)

The first of the above examples includes the indirect object them, while the
second does not. In any case, examples of this kind, namely examples without
from, are not at all restricted as far as the BNC is concerned. They are almost as
frequent as the examples of “hinder + object + from + gerund” in it, although
the gerund used in “hinder + –ing” is inclined to be more nominal than verbal,
suggesting that the gerund here is almost like the nominal object of hinder.
Moreover, the BNC provides a single example followed by the to-infinitive,
which is rather exceptional:

… but they altogether refused, and seeing that they refused I did not hinder them
to go, for I will press no man.
(C. Gordon Booth, An Islay Notebook, from BNC)

Despite the existence of these different constructions, however, “hinder +


object + from + gerund” seems to be the dominant form in Present-day English.
The OED does not even refer to the form “hinder + –ing”, although it is
certainly encountered throughout the history of English.82 Alternatively, the
OED refers to the existence of the preposition in in place of from, but this
usage, though indeed attested in the history of English, is not at all common in
Present-day English. The entire dataset of the OED provides only three
examples of this type, all of which go back to the earlier half of the eighteenth
century. The latest one runs as follows:

That it may not hinder them in rising or coupeeing.

82
Apart from these, which are the main concern of the present study, Dixon (1991: 199)
notes the existence of “hinder + object + relative clause” as a favoured construction of this
verb.
Verbs of implicit negation and gerunds with prepositions 101

(1748 A True and Particular Relation of the Dreadful Earthquake Which Hap-
pen’d at Lima, the Captial of Peru)

Likewise, the OED also refers to the use of the prepositions of and for, stating
again at the same time that they are rare or obsolete. See the following il-
lustrative examples quoted from the OED:

Lest it should feble hys fleshe … and hyndre hys harlot of teming
(1532 Sir Thomas More, The Confutacyon of Tindales Answere)

Marcellus … determined to hinder Bomilcar for arriving at Saracose


(1600 Philemon Holland, Livy’s Romane Historie)

All in all, it is safe to state that prepositions other than from are rare with this
verb throughout the history of English.
Finally in this section, it is important to note that the frequency of this verb
itself is inclined to be low, at least in contemporary English. The sheer fre-
quency of hinder in LOB and FLOB would be surprising for unprepared
readers. They provide only four examples of hinder of all kinds each, ex-
cluding the examples of unhindered.83 The fact that LOB and FLOB present
about 100 examples of prevent of all kinds reveals how infrequent hinder is in
comparative terms in Present-day English. Supposing that some semantic field
is shared between hinder and prevent, one is inclined to surmise that the former
is increasingly replaced by the latter in current English, which would probably
explain the difference in frequency.
Hinder is, however, a Germanic word, which has a much longer history
than prevent, while prevent is a loan word and a newcomer in English. Further-
more, the establishment of the use of the preposition from with hinder is
perhaps earlier than that of some other verbs including prevent, as the fol-
lowing discussion explicates. In this manner, hinder may have functioned as a
leader for a series of English verbs taking the same syntactic pattern.
The present section explores various shifts of complements that the same
verb has experienced in the process of the establishment of the preposition
from followed by the gerund. As the OED (s.v. hinder) claims, prepositions
other than from occurred with this verb in the past. The complementation
utilising to-infinitives and that-clauses were also possible with hinder in the
past. The last quotation of the to-infinitive construction of hinder in the OED
83
“Of all kinds” here means that examples without any complements are included in these
figures.
102 Chapter 4

(s.v. hinder) is from 1741 and the last quotation of the that-clause construction
of hinder is from 1843, although they are not marked as obsolete yet. Hence,
the syntactic behaviours of hinder in the history of English is fairly com-
plicated, the analysis of which is the principal concern of the present section.

4.4.2. The rise of gerunds and the decline of the other constructions

In contrast to its almost consistent use of gerunds in Present-day English,


hinder was on occasions followed by that-clauses in historical data. The fol-
lowing example, which includes expletive negation, is a case in point:

The use of which bones, is to hinder that the valve do not easily totter.
(1668 Nicholas Culpepper and Abdaih Cole, Bartholinus’ Anatomy)

Accordingly, one could expect the occurrence of the decline of that-clauses fol-
lowed by the rise of to-infinitives (the first complement shift) and then by the
rise of gerunds (the second complement shift) in the history of hinder.
Indeed, the decline of that-clauses takes place, but the rise of to-infinitives
is not as explicit in the case of hinder as in the case of some other verbs like
prohibit. As Figure 17 indicates, to-infinitives, which were probably on the
increase during the early Modern English period, were quickly superseded by
gerunds. In this sense, the overall development of hinder is similar to that of
prevent, which also shifted quickly to the established use of gerunds in the
history of English.

120
100
80
60
40
20
0
15th c. 16th c. 17th c. 18th c. 19th c. 20th c.

that-clauses but-clauses to-infinitives gerunds

Figure 17. Hinder and four patterns of complementation in the OED (raw frequencies)
Verbs of implicit negation and gerunds with prepositions 103

The rise of to-infinitives may have occurred with hinder as well as with other
verbs of implicit negation hitherto discussed, but the sixteenth century already
observes a clearly expanding use of gerunds. This inclination is further ac-
celerated in the following centuries as the figure above shows. On the whole,
the expanded use of gerunds with hinder seems to have occurred about a cen-
tury earlier than with prohibit, which as mentioned above displays a notable
expansion of gerunds only in the eighteenth century. In this sense, the tendency
in the rise and fall of different complementation patterns of hinder is similar to
that of prevent, which also experiences a relatively early expansion of gerunds,
making the clear observation of the rise of to-infinitives difficult. In the fol-
lowing, I will cite a few examples of to-infinitives used with hinder to confirm
that the usage was indeed available in the past:

What hinders then To reach, and feed at once both Bodie and Mind?
(1667 John Milton, Paradise Lost)

The pressure of the quiescent body against the obstacle that hinders it to move.
(1753 Ephraim Chambers, Cyclopædia; or, an Universal Dictionary of Arts and
Sciences, Suppl. App. s.v. Force)

Considering the fact that Visser (1963–1973: §2075) cites an example of


the for to-infinitive used with this verb from c1396, the construction with in-
finitives goes back to an earlier period than Figure 17 suggests. On the whole,
however, it has never been so frequent as to become a dominant form of com-
plementation of the verb under consideration in the entire history of English.
After being half as frequent as gerunds in the seventeenth century, to-infinitives
went through a gradual recession as far as Figure 17 is concerned. Visser
(1963–1973: §2075) states that the construction is now obsolete. Likewise, the
OED says that it is obsolete or rare today.
With regard to the usage of gerunds with this verb, Visser (1963–1973:
§1775) aptly remarks that “hinder + –ing” is available despite the absence of it
in the OED, providing the following example from among relatively recent
literary works: To-day a slight rain hindered riding (1874 George Eliot, Daniel
Deronda). Indeed, the type of “hinder + –ing” is available at least in the
quotations of the OED up to the nineteenth century, and as shown above with
some quotations from the BNC, it is still available in contemporary English.
The following are some illustrations of the “hinder + –ing” type in support of
Visser (1963–1973: §1775):
104 Chapter 4

The will can hinder seeing, not immediately, but by the loco-motive power; by
closing the eyes.
(1640 Edward Reynolds, A Treatise of the Passions ad Faculties of the Soul)

Dante … Hated wickedness that hinders loving.


(1855 Robert Browning, One Word More)

While the above examples of hinder are followed by simple –ing, examples of
the following type, where the gerund is more nominal, are more frequently
attested in the OED quotations:

For if these Renegades had formed such a Conspiracy, what hindered their
accomplishing it?
(1751 Affecting Narrative of the Catastrophe of his Majesty’s Ship ‘Wager’)

Least the intention of to much Reading hinder the working of those vertuous
drugs.
(c1614 Sir John Davies, Letters to the Earl of Salisbury)

The gerund in these examples is preceded by a genitive personal pronoun or


the definite article, showing its nominal feature.
Simultaneously, the verb hinder gradually experiences the rise of the con-
struction “hinder + from + –ing” in its historical development, as in the case of
prohibit and prevent discussed above. In other words, “hinder + –ing” is most
probably the older and original form of the gerundial construction, and there-
fore more abundant in the historical data than in Present-day English. See
Figure 18, which displays the raw frequencies of gerunds with and without the
preposition from with hinder in the OED dataset:

80

60

40

20

0
15th c. 16th c. 17th c. 18th c. 19th c. 20th c.

gerunds only from + gerunds

Figure 18. Hinder and gerunds with or without from in the OED (raw frequencies)
Verbs of implicit negation and gerunds with prepositions 105

Despite the paucity of relevant examples in an early period of the development


of gerunds, it is most probably safe to conclude from this graph that the ge-
rundial usage starts off as the simple –ing form with hinder and that the pre-
position from was introduced in an increasing manner in the course of its
development. While the use of from becomes more frequent than its absence
only in the nineteenth century with prohibit (see Figure 12 on p. 86), the
notable expansion of the employment of from is attested in the seventeenth
century with hinder in Figure 18. In other words, the expanded use of from
occurs about two centuries earlier with hinder than with prohibit. Likewise, it
is about a century or two earlier than the expanded use of from with prevent
(see Figures 14 and 15 on p. 94). Therefore, hinder practically led other verbs
of the same pattern in the historical development of complementation.
One additional point to be made in this connection is that hinder, which is
Germanic and which goes back to the Old English period and is therefore older
than prohibit and prevent, displays some lingering features of older con-
structions like that-clauses and to-infinitives, although on the whole the same
verb shows comparatively earlier shifts of complements, as far as the quota-
tions of the OED are concerned. While constructions other than gerundial ones
are virtually obsolete with prohibit and prevent in the nineteenth century on-
wards (or perhaps from the eighteenth century), that-clauses and to-infinitives
are still encountered with hinder in the same period (see Figure 17 above).
Examples of these older types include:

He does hinder that it become … a part of it.


(1843 Thomas Carlyle, Past and Present)

What hinders me to make my single will The world’s whole law?


(1878 Bayard Taylor, Prince Deukalion)

Furthermore, hinder, as illustrated by the quotation below, provides some ex-


amples of but-clauses, which are considered to take the place of older that-
clauses with expletive negation in the history of English. This is another
feature not clearly evidenced with prohibit and prevent but with hinder, as far
as the OED is concerned. See the following example:

That hinders not but that they are generally less doubtful.
(1690 John Locke, An Essay Concerning Human Understanding)

As I argue in Chapter 5, the occurrence of but-clauses like this is a


106 Chapter 4

particularly characteristic feature of verbs of implicit negation with which the


complement shift from that-clauses to nonfinite forms like to-infinitives and
gerunds is not very prominent. The verbs treated in the next chapter—fear,
doubt, and deny—are of this type, with which there are abundant examples of
but-clauses available in the historical data. On the whole, the verbs treated in
the present chapter are, by contrast, of the opposite kind, in that they all
experienced the shift of complements quite quickly, almost suppressing the
first stage or the shift from that-clauses to to-infinitives and immediately
moving on to the second stage or the shift from to-infinitives to gerunds. This
was particularly the case with prohibit, for example, which was treated in 4.2.2.
above.
Theoretically, however, it is also possible for verbs explored in the present
chapter to provide examples of but-clauses, though not copiously at all in prac-
tice. Indeed, some sporadic examples of but-clauses are encountered with hin-
der, but as expected, they are limited only to the transitional period of the
seventeenth and eighteenth centuries (see Figure 17 above). The existence of
but-clauses is marginal in any case, however, and it is no longer witnessed in
the nineteenth century with hinder.

4.4.3. The contrast between nominal and pronominal objects

As hitherto discussed, the complementation patterns of hinder display a rela-


tively early development in chronological terms, at least in comparison to verbs
like prohibit and prevent, both presenting the same pattern of the preposition
from followed by the gerund in Present-day English. The rate of the changes,
however, differs depending upon whether the semantic subject of the com-
plement is a personal pronoun or not, as in the case of prevent discussed above.
See the two graphs on the following page, which depict the raw frequencies of
that-clauses, to-infinitives, and gerunds in the cases of personal pronouns
(Figure 19) and the other forms of semantic subjects, i.e. mostly nouns (Figure
20):84

84
As in the case of prevent discussed above, pronouns other than personal pronouns are in-
cluded in Figure 20.
Verbs of implicit negation and gerunds with prepositions 107

50
40
30

20
10
0
16th c. 17th c. 18th c. 19th c. 20th c.

that-clauses to-infinitives gerunds

Figure 19. Hinder with personal pronouns and three patterns of complementation in the
OED (raw frequencies)

50
40
30
20
10
0
16th c. 17th c. 18th c. 19th c. 20th c.

that-clauses to-infinitives gerunds

Figure 20. Hinder with nouns and three patterns of complementation in the OED (raw
frequencies)

The overall tendencies are similar between the cases of personal pronouns (see
Figure 19) and nouns (see Figure 20). In both cases, a significant rise of
gerunds is evidenced in the course of the seventeenth century, and their use is
predominant thereafter up to the present day.
At the same time, however, it is also possible to trace a slight difference
between the cases of personal pronouns and nouns in the above graphs. In
particular, a slight but notable rise of to-infinitives or the first complement shift
is more transparently observed with nouns in the seventeenth century than with
personal pronouns, with which the first stage of the shift of complements is
almost suppressed by the sudden rise of gerunds in later dates. In the later part
108 Chapter 4

of the Modern English period, however, that-clauses85 and to-infinitives are


inclined to linger more with personal pronouns than with nouns, as far as the
OED quotations are concerned. With the exception of these minor points,
however, the relatively early establishment of the use of gerunds is a shared
feature between Figures 19 and 20.
The real and noteworthy difference between the cases of personal pro-
nouns and nouns is found in the process of the introduction of the preposition
from. As the following two figures reveal, the development of the “hinder +
from + –ing” construction occurs much earlier with nouns than with personal
pronouns. Figure 21 displays the raw frequencies of gerunds with or without
from in the case of personal pronouns, and Figure 22 corresponding fre-
quencies in the case of nouns. Both graphs are based upon the dataset of the
OED quotations:

50

40

30

20

10

0
16th c. 17th c. 18th c. 19th c. 20th c.

gerunds only from + gerunds

Figure 21. Hinder with personal pronouns and gerunds with or without from in the OED
(raw frequencies)

85
Here I use the term “semantic subject” to cover the objective form preceding to-infinitives
and objective and possessive forms preceding gerunds. In the case of that-clauses, however,
the semantic subject means the real subject within them.
Verbs of implicit negation and gerunds with prepositions 109

50

40

30

20

10

0
16th c. 17th c. 18th c. 19th c. 20th c.

gerunds only from + gerunds

Figure 22. Hinder with nouns and gerunds with or without from in the OED (raw fre-
quencies)

It is immediately clear from Figures 21 and 22 that the shift from simple ger-
unds to gerunds with from is attested much earlier in the case of nouns than in
the case of personal pronouns. The contrast is evidenced already in the six-
teenth century, when all relevant examples of nouns display the employment of
from and when all examples of personal pronouns present gerunds without the
preposition. The tendency is even more pronounced from the seventeenth
century onwards when there are more abundant relevant examples available in
the OED dataset. When nouns are involved, the construction with from is
predominant throughout the Modern English period, whereas in the case of
personal pronouns, the construction without from competes with the con-
struction with from for some time up to the nineteenth century, when the latter
becomes the predominant form. In this manner, examples with nouns used as
the semantic subject of complements are more progressive than examples with
personal pronouns, as far as the shift from plain gerunds to gerunds preceded
by from is concerned.
The main factor behind this time lag is essentially the same as in the case
of prevent discussed above (cf. 4.3.), although at first glance the situation of
hinder may seem to be contradictory to the situation of prevent, with which the
relative proportion of the examples of “from + –ing” to the examples of “object
+ –ing” is larger when personal pronouns are involved. To clarify this point,
one needs to note that many of the “hinder + –ing” in the case of personal
pronouns are in fact the “hinder + possessive + –ing” construction. In the case
of nouns, by contrast, the form “hinder + gerund only” means, in most cases,
the construction “hinder + object + –ing” (rather than “hinder + possessive +
110 Chapter 4

–ing”). This is indicated in the following two graphs, which are in parallel with
Figures 14 and 15 (p. 94) under the section of prevent:

25

20

15

10

0
16th c. 17th c. 18th c. 19th c. 20th c.

hinder + possessive + -ing hinder + object + -ing hinder + object + from + -ing

Figure 23. Hinder with personal pronouns and the three gerundial constructions in the
OED (raw frequencies)

50

40

30

20

10

0
16th c. 17th c. 18th c. 19th c. 20th c.

hinder + possessive + -ing hinder + object + -ing hinder + object + from + -ing

Figure 24. Hinder with nouns and the three gerundial constructions in the OED (raw
frequencies)

As shown by these graphs, all relevant examples of personal pronouns are in


the possessive form up to the eighteenth century. Some illustrative examples
are:
Verbs of implicit negation and gerunds with prepositions 111

He ranne hastily to the shore to hinder their disembarking.


(1632 James Hayward, tr., Biondi’s Eromena; or Love and Revenge)

The Sand was at last so gravelly, that it hinder’d our boring any deeper.
(1713 William Derham, Physico-Theology; or a Demonstration of the Being and
Attributes of God from his Works of Creation)

Only from the nineteenth century onwards, the OED yields examples like the
following, where personal pronouns are in the objective case:

What would hinder him lasting out to ninety years or a hundred even?
(1881 Mrs. J. H. Riddell, Senior Partner)

By contrast, the possessive form is extremely rare in the case of nouns. The
only possible example in the OED quotations runs as follows:

It is useful to … hinder the Lympha’s being plentifully spewed out of the Glands.
(1710 Thomas Fuller, Pharmacopæia Extemporanea)

All the other relevant examples are in the objective form when the semantic
subject of the gerund is a noun.
As discussed under the sections of prevent, “hinder + object + –ing” is
unlikely to be an elliptical form of “hinder + object + from + –ing” in historical
terms. The chronological development of this verb is more likely to be in the
opposite direction, however the relationship between the two constructions
may be conceived by language users in the context of contemporary English.
Unlike the case of prevent, where “prevent + object + –ing” is encountered to a
notable extent, however, hinder displays a relatively quick shift to “hinder +
object + from + –ing”, leaving only scanty examples of “hinder + –ing” when
nouns are involved. This is how the competition between “hinder + object +
–ing” and “hinder + object + from + –ing” is difficult to grasp in the case of
nouns. This again confirms that the historical development of hinder is faster
than many other verbs of implicit negation which present similar patterns in the
shifting of complementation. With personal pronouns, by contrast, the older
form “hinder + –ing” is still observed to a notable extent, mainly in the form of
“hinder + possessive + –ing”.
Once again, the above discussion on prevent and hinder confirms that the
historical timings of the shifting of complementation patterns can differ de-
pending upon whether the semantic subject of the complement is a personal
112 Chapter 4

pronoun or not. The overall historical direction is consistent irrespective of this


factor, but the rate of change can differ to a noticeable extent. I would assume
that the same applies to prohibit, the first verb discussed in the present chapter,
although relevant examples of it are not so copious in the OED quotations as to
allow the same kind of analysis.

4.5. The historical development of refrain


4.5.1. Preliminary remarks

Before concluding the present chapter, I would like to make a short historical
survey of another verb which is also accompanied by “the preposition from +
–ing” in most frequent cases. This verb is refrain, which differs from the three
verbs discussed above in the present chapter, in that the semantic subject of its
complement clause is in usual cases identical with its own subject. In this sense,
refrain is similar to the verb forbear, which is treated above in Chapter 3. By
contrast, verbs like prohibit, prevent, and hinder (all discussed in the present
chapter) assume the existence of another person who is prohibited, prevented,
or hindered from performing something. The obvious question arises as to
whether this semantic distinction leads to a different path in the historical de-
velopment of refrain.
As far as the syntactic forms are concerned, refrain is similar to prohibit,
prevent, and hinder. Unlike the case of forbear, the use of “from + –ing” is
fully established with refrain in Present-day English. The following are some
illustrative examples cited from the BNC:

First, children might refrain from reporting or at least giving evidence if they
know that a substantial prison sentence for a member of their family is the prob-
able outcome. (Andrew Ashworth, Principles of Criminal Law, from BNC)

... though they should refrain from expressing in debate views contrary to that
advice ... (Responses to Crime, from BNC)

Likewise, LOB provides ten relevant examples of refrain, all of which are
accompanied by the preposition from followed by –ing. Some additional
examples of this construction are:

The Commissioners of Inland Revenue wisely refrained from asking how she
paid for the mink coats but demanded a receipt instead. (LOB, R)
Verbs of implicit negation and gerunds with prepositions 113

A little more honesty, even if one refrains from going into too many details,
would help many a child to make a proper adjustment to life as it grows up.
(LOB, D)

Hence, the present-day usage of the verb under discussion seems to be more or
less established. The Shorter Oxford English Dictionary on Historical Prin-
ciples (s.v. refrain), however, writes that there was once a construction with the
to-infinitive with this verb in English, which suggests that the history of refrain
is certainly worth investigating. Unfortunately, Herbst, Heath, Roe, and Götz
(2004) do not provide an entry for this verb.

4.5.2. The rise of to-infinitives and gerunds

Despite the differences in meaning, at least in Present-day English, the his-


torical development of refrain is strikingly similar to that of prohibit, except
that the timing of the shifts of complements seems to be always well ahead in
the case of refrain. It also presents the complementation patters of that-clauses
and to-infinitives in earlier periods of English, as illustrated by the examples
below, which are, however, clearly marked as “obsolete constructions” today in
the OED (s.v. refrain). They have by now been superseded by gerundial ones:

I shalle soo refrayne hym that he shalle no more dare demaunde suche thyng.
(1483 William Caxton, The Golden Legende)

Lawrence Ualla … could not refrayne to enveygh against the Popish clergie.
(1561 John Daus, tr., Bullinger’s Hundred Sermons upon the Apocalips)

The first of these examples quite unequivocally confirms that refrain once had
a meaning similar to the meanings of the three verbs discussed above in the
present chapter and that it could take an indirect object. This does not neces-
sarily apply to the second of the above examples, where the to-infinitive is
employed, but the existence of examples like the following in the historical
data is noteworthy in this relation:

Assone as … he knewe it was Le Surnome, he coude not refrayne him to kisse


him. (c1500 The Three Kings’ Sons)

This is an example where the to-infinitive is preceded by its semantic subject,


114 Chapter 4

which is the indirect object of the matrix verb refrain. Although it is a reflexive
one, the presence of this example confirms that the transitive use of the same
verb was certainly available in the past. This may possibly be an explanation
for the fact that refrain took a similar path as the path taken by prohibit in its
historical development.
Indeed, the first meaning in the list of the OED (s.v. refrain), which is
marked as obsolete in current English, is a transitive one: “To restrain, hold
back, check (a person or thing)”. It is also relevant to refer to the fact that
Garner (2003: 693) gives the following example, which is in his view a mis-
taken use of refrain for restrain: I had to refrain myself from snapping that I
wasn’t quite ready to date (Miriam Sagan, “How to Talk to a Widow”,
Albuquerque Journal, 2 March 1997). Although this may simply be a mistaken
use of refrain for restrain as Garner maintains, it is also possible that the usage
under consideration is an inherited one from the past, which is occasionally
found in Present-day English in a seemingly-accidental manner. As a matter of
fact, Garner (2003: 692) himself admits that the verbs refrain and restrain were
“once almost interchangeable”, suggesting that he was aware of the existence
of the transitive use of refrain in the history of English.
Simultaneously, the OED quotations include examples like the following
where refrain is followed by that-clauses and where the meaning of the same
verb is almost equal to its meaning in Present-day English:

I may of myne owne strength refraine that I do mine enemy no hurte.


(1549 Miles Coverdale, The Second Tome or Volume of the Paraphrase of Eras-
mus upon the Newe Testament)

The that-clause in the above example includes expletive negation, as in the


case of other verbs of implicit negation.
Irrespective of the meaning of refrain, however, that-clauses and to-
infinitives are to be taken the place of by gerunds fairly quickly in the course of
the historical development of this verb. Vosberg (2006: 158–159) demonstrates,
on the basis of his investigation of a large number of literary works, that ger-
unds are already dominant in the early Modern English period with the verb at
issue and that they become almost the sole form of complementation in the
nineteenth century, although his analysis is based upon to-infinitives and ger-
unds only and does not deal with that-clauses. More or less the same tendency
is obtainable from the quotations data in the OED as well. See Figure 25,
which depicts the raw frequencies of the three complementation patterns of
refrain in different centuries in the OED:
Verbs of implicit negation and gerunds with prepositions 115

30

20

10

0
15th c. 16th c. 17th c. 18th c. 19th c. 20th c.

that-clauses to-infinitives gerunds

Figure 25. Refrain and three patterns of complementation in the OED (raw frequencies)

Unfortunately, relevant examples of refrain are not abundant in the fifteenth


century in the OED quotations, while the rise of to-infinitives and the rise of
gerunds are clearly visible in the sixteenth century. In other words, the shift to
to-infinitives and the shift to gerunds take place almost simultaneously with
this verb. The decline of to-infinitives in later Modern English and the more
and more established use of gerunds in this graph, however, reveal that these
are two separate shifts involved in the historical development of refrain. The
graph clearly shows that gerunds are virtually the sole form of complementa-
tion with this verb by the time of the nineteenth century.
As hitherto discussed, the historical timing of the expansion of gerunds is
notably early with this verb. It is well ahead of the cases of prohibit and pre-
vent, and perhaps even earlier than in the case of hinder, which also shows the
dominance of gerunds as early as in the sixteenth century (see Figure 17 on p.
102). In theory, therefore, refrain could also have led the historical
development of prohibit and prevent, both of which followed a similar path in
later dates, but in practice, it was increasingly intransitive in use, and separated
itself form the other transitive verbs in this group. Consequently, the leading
position may have been given up by this verb in this process.

4.5.3. Gerunds with or without the preposition from

The later development of the gerundial constructions with refrain is also


similar to that of prohibit, prevent, and hinder. The same verb also undergoes
the rise of the construction “refrain + from + –ing” side by side with “refrain +
116 Chapter 4

–ing”. In the following quotations, the two types of gerundial constructions are
contrasted. The first two illustrate the use of plain gerunds and the last two
gerunds with from:

A spark can no more refrain running into love after a Bottle etc.
(1678 Thomas Otway, Friendship in Fashion)

We could not refrain smiling at one another.


(1725 Daniel Defoe, A New Voyage round the World)

Lysicles could hardly refrain from laughing.


(1732 George Berkeley, Alciphron, or the Minute Philosopher)

The Board to have … the option of refraining from making any award, and of
publishing an opinionative report on the dispute instead.
(1894 The Pall Mall Gazette, 24 December)

Figure 26 below, which exhibits the raw frequencies of the two forms of
gerunds in the OED quotations, shows that the simple “refrain + –ing” con-
struction was superseded by the form with the preposition from in the course of
the Modern English period:

30

20

10

0
15th c. 16th c. 17th c. 18th c. 19th c. 20th c.

gerunds only from + gerunds

Figure 26. Refrain and gerunds with or without from in the OED (raw frequencies)

Here again, the dominance of “refrain + from + –ing” was already reached with
this verb by the time of the sixteenth century, although the relationship between
the two forms of complementation is slightly inconsistent or irregular in this
graph, especially in respect of the seventeenth century. On the whole, however,
Verbs of implicit negation and gerunds with prepositions 117

the overall expansion of “refrain + from + –ing” is straightforwardly visible in


Figure 26. The competition between the two forms of gerunds lasts a few
hundred years during the early Modern English period, after which the
dominance of the form with the preposition from was attained. It is outstanding
that the nineteenth and the twentieth centuries do not provide any examples of
simple gerunds, i.e. gerunds without prepositions, in view of the fact that the
simple gerund constructions are still available today with prevent and hinder
discussed above. On this matter, further research is in need.
This result is again largely in keeping with the data provided by Vosberg
(2006: 158–162), where the gerundial construction with from is always more
frequent than the one without and where the construction with from reaches the
proportion larger than 95% in the nineteenth century. Vosberg shows that ex-
ceptional examples of to-infinitives in the nineteenth century are attested in the
context where refrain itself occurs in the form of gerunds. In other words, they
are explainable from the perspective of the horror aequi principle discussed
under 3.3.3. above.
5. Verbs of implicit negation and subordinate
clauses

5.1. Verbs of implicit negation and subordinate clauses

The verbs discussed in Chapters 2, 3, and 4 have undergone the decline of that-
clauses followed by the rise of to-infinitives (the first complement shift), and
some of them have also experienced the subsequent replacement of to-
infinitives by gerunds (the second complement shift) in their historical de-
velopment. In addition, some of the verbs have further gone through the rise of
the preposition from in their gerundial constructions. Consequently, they are
most typically followed by to-infinitives or gerunds either with or without
prepositions in Present-day English.
The verbs to be discussed in the present chapter are slightly different in
nature from the verbs treated in the above chapters. Chapter 5 investigates
three verbs often followed by that-clauses even in Present-day English: fear,
doubt and deny. They also experienced to some extent the recession of that-
clauses during the period from later Middle English to early Modern English,
as other verbs of implicit negation did. Interestingly enough, however, the
recession of that-clauses did not always lead to the established use of to-
infinitives or gerunds with fear, doubt, and deny, although some evidence of
the rise of to-infinitives and gerunds is attested in the history of these verbs as
the discussions in the following sections reveal. In the case of doubt and fear,
particularly in affirmative sentences, the decline of that-clauses seems to have
led to the rise of the parenthetical use of I doubt and I fear, by which the
original subordinate clauses were converted into the matrix ones. In other
words, the loss of subordinate clauses occurred with these verbs as well, but in
a different manner from the verbs hitherto dealt with.
While some features of the same type of development were also evidenced
with deny, the parenthetical use was not established with it in the end in its
historical development. With this verb, the ordinary use of that-clauses, which
once receded in some measure, came back in recent years as a major
120 Chapter 5

construction after the failure of the development of the parenthetical I deny. As


a matter of fact, it is one of the verbs which have not so far been successful in
establishing any of the major paths provided by the historical shifts of com-
plements discussed in the present research, although it displays some traces of
having tried to follow several possible paths in line with other verbs of implicit
negation.
The discussion in the following sections will provide further details on
these matters. 5.2. below is concerned with the historical development of fear,
5.3.with the historical development of doubt, and 5.4. with deny.

86
5.2. The historical development of fear
5.2.1. Preliminary remarks

The verb fear is often classified in existing research under the category of
“verbs of emotion, affection, or liking”. Visser (1963–1973: §1847), for ex-
ample, allocates this verb the category of “verbs of affection & disaffection,
fearing, desiring, wishing, etc.” Among more recent studies, Rudanko (1998:
38) classifies the verb among verbs which express negative volition, together
with blush, decline, forbear, hesitate, refuse, regret, scruple, and shudder.
Moreover, Huddleston and Pullum (2002: 170) provide the category of “verbs
of cognition, emotion, and attitude” along with the illustrative verbs of agree,
believe, forget, hope, intend, know, like, love, realise, regret, remember, sup-
pose, think, understand, want, wish, and wonder as well as fear, whereas Dixon
(2005: 160) simply gives the designation of “liking verbs” to the same type of
verbs.
Despite the fact that fear is often quoted as one of the representative verbs
of this category, which is variously named depending upon the linguist,87 its
syntactic patterns or behaviours have hardly been discussed in the literature,
especially from historical perspectives, while some other verbs have often at-
tracted scholarly attention in this respect to this day. Bladon (1968), for
instance, investigates the uses of infinitives and gerunds dominated by like,
love, hate, dislike and prefer, but does not refer to fear. De Smet’s (2004)

86
An earlier version of 5.2. appeared as Iyeiri (2009a). The Modern English Association
kindly granted permission to use it in the present study.
87
Egan (2008: 24–26) makes a summary of different appellations given by different linguists
to verbs of this semantic category, and he himself lists fear under the category of “verbs of
attitude”, using the same label as Verspoor (2000) employs.
Verbs of implicit negation and subordinate clauses 121

interest is also directed to the verb like and De Smet and Cuyckens’ (2005)
interest to the verbs like and love only.
Fear is, however, an interesting verb for the purpose of the present re-
search, in that its history offers a number of hints as to possible paths various
other verbs can take in their historical development. As in the case of the other
verbs so far investigated, the decline of that-clauses occurs with fear in its
early history, perhaps side by side with the decline of the subjunctive in Eng-
lish. The later development that it experienced, however, differs to some extent
from the cases of the other verbs so far analysed. As demonstrated in the
following sections, the most notable with this verb is the development of the
parenthetical use, usually in the form I fear, although the expansion of to-
infinitives and gerunds also takes place in the process of its historical develop-
ment.

5.2.2. Fear in Present-day English

Dictionaries also tend to spare only limited space for the usage of the verb fear
in Present-day English. Longman Dictionary of Contemporary English, for ex-
ample, refers to some different shades of meanings of this verb within about a
score of lines and points out the occurrences of that-clauses and to-infinitives
as its complements. The same applies to the Oxford Advanced Learner’s Dic-
tionary of Current English, which gives a limited number of citations for the
verb fear, mentioning the use of gerunds as well as that-clauses and to-
infinitives as its complements. This is also true with Macmillan English Dic-
tionary for Advanced Learners (s.v. fear, verb), which provides one example
each for the constructions with that-clauses and to-infinitives.
Indeed, the three syntactic patterns of that-clauses, to-infinitives, and ger-
unds are observed with fear in Present-day English, as illustrated by the fol-
lowing examples cited from LOB and FLOB:

“I fear that if I don’t I shall lose any regard you have for me, so I will say at once
that I do”. (FLOB, P)

They feared to see “their commerce and manufactures completely destroyed by


competition and the interruption in the supply of Austrian coal”. (LOB, J)

It was as if she feared being loved, wrapping a cloak of scorn for sentiment and
outward affection around herself as if it might protect her from those payments
that love required. (FLOB, N)
122 Chapter 5

The three patterns of complementation are, however, not equally frequent in


Present-day English. An analysis of LOB and FLOB demonstrates that the
employment of that-clauses is predominant with fear, and that the other con-
structions, i.e. to-infinitives and gerunds, are much more restricted in oc-
currence. LOB yields 24 examples of fear followed by a complement, of which
23 illustrate the use of that-clauses (including clauses introduced by the con-
junction that and clauses whose introductory that is elliptical).88
Likewise, FLOB, whose data are 30 years later than those of LOB, pre-
sents 27 relevant examples of fear, of which 24 reveal the use of that-clauses.
In other words, the use of that-clauses is pretty well established with the verb
fear in Present-day English, whereas other constructions like to-infinitives and
gerunds are marginal,89 though mentioned and acknowledged in dictionaries of
contemporary English. This is largely in accordance with the description of the
usage of fear in Herbst, Heath, Roe, and Götz (2004: 304–305).90
One possible reason why the treatment of the syntactic features of fear is
slight in Present-day English dictionaries and reference materials is that even
the construction fear that ..., which is the predominant form that the verb takes,
is formal and increasingly replaced by be afraid that … on informal occasions
(Swan 2005: 28). The infinitival construction “fear + to-infinitive” is also
considered to be formal and equivalent to be afraid to do … (see Longman
Dictionary of Contemporary English, s.v. fear). It is feasible that the overall
employment of fear itself is less and less common, although further research
into Present-day English is certainly necessary to prove this point.

88
See also Note 3 in Chapter 1 on the ellipsis of the conjunction that. The verb fear provides
a significant number of clauses where the conjunction that is elliptical. The issue will be
discussed later.
89
Egan (2008: 249–253) pays particular attention to the minor constructions with the to-
infinitive and the gerund, and discusses possible differences in meaning between “fear +
to-infinitive” and “fear + –ing”. He considers that the subject of the matrix verb, i.e. fear,
possesses control over the realisation of the situation of the complement introduced by the
to-infinitive, but this is not the case when the complement clause is in the form of the ger-
und. See also 1.4. above.
90
For the sake of comparison, I have also investigated the complementation patterns of fear
in American English by using Brown and Frown. Apparently, the occurrence of types other
than that-clauses is slightly more extensive in American English than in British English.
The Brown corpus yields fifteen relevant examples, which include four to-infinitives and
one gerund. Similarly, Frown, which is some 30 years later than Brown in compilation,
gives a total of 46 relevant examples, which include one to-infinitive example and six
gerundial ones.
Verbs of implicit negation and subordinate clauses 123

5.2.3. Historical overview of the verb fear

The OED provides 2,192 examples of fear in total excluding duplicated ones,
of which about half are followed by complements of one sort or another. The
discussion in the following is based upon fear with complements, and the
present section, in particular, treats the relationship among the three major
types of complementation: that-clauses, to-infinitives, and gerunds.
As illustrated by the following quotations, the three types of complements,
which are encountered in Present-day English, are attested in the historical data
of the OED as well:

He feared that if they had not their pardons in likewise, they would either make
business or they would avoid.
(1514 Lord Mountjoy, Strype’s Ecclesiastical Memorials)

He leaped out of the bathe vnbathed, because he feared the bathe shoulde haue
fallen.
(1570 John Foxe, Actes and Monuments of these Latter and Perillous Dayes)

Hee feared to run into any such inconvenience, as might cause his friends to
relapse from him.
(1633 Sir Thomas Stafford, Pacata Hibernia, Ireland Appeased and Reduced)

He, like me, hated and feared being carried in this ship, for all his whistling in the
dark. (1968 Jon Manchip White, Nightclimber)

The first and the second of the above examples illustrate the employment of
that-clauses (or its ellipsis), the third the employment of the to-infinitive, and
the last the employment of the gerund. Despite the limited occurrences of to-
infinitives and gerunds with fear in Present-day English, they are relatively
well attested in the historical data, although in the light of the frequent oc-
currence of that-clauses in general, they are still marginal. See Figure 27 on the
next page, which displays the raw frequencies of the three complementation
patterns under consideration in the OED quotations:
124 Chapter 5

250

200

150

100

50

0
13th c. 14th c. 15th c. 16th c. 17th c. 18th c. 19th c. 20th c.

that-clauses to-infinitives gerunds

Figure 27. Fear and three patterns of complementation in the OED (raw frequencies)

This graph highlights how the employment of that-clauses is stable with fear
throughout the history of English. The predominance of this construction is
quite obvious from the sixteenth century onwards. The second most frequent
pattern of complementation is the one which employs to-infinitives, but it is far
less commonly attested than that-clauses in the OED quotations.
Interestingly enough, however, a closer look at the OED dataset reveals
that there is a marked difference in terms of the complement patterns between
fear in affirmative contexts and fear in negative ones. The two figures below,
which display the raw frequencies of relevant examples in different centuries,
aptly depict the contrast. See Figure 28 for fear used in the affirmative and
Figure 29 on the following page for fear used in the negative:

250

200

150

100

50

0
13th c. 14th c. 15th c. 16th c. 17th c. 18th c. 19th c. 20th c.

that-clauses to-infinitives gerunds

Figure 28. Fear in affirmative sentences and three patterns of complementation in the
OED (raw frequencies)
Verbs of implicit negation and subordinate clauses 125

12
10
8
6
4
2
0
13th c. 14th c. 15th c. 16th c. 17th c. 18th c. 19th c. 20th c.

that-clauses to-infinitives gerunds

Figure 29. Fear in negation and three patterns of complementation in the OED (raw fre-
quencies)

First of all, these two graphs demonstrate that there is an obvious contrast
between fear in affirmative sentences and the same verb used in negation, in
terms of the frequencies of the relevant examples themselves. To be more pre-
cise, fear has a marked tendency to occur in affirmative sentences rather than
in negative ones. Figures 28 and 29 in total yield 925 examples, of which as
many as 870 (94%) are attested in affirmative sentences or in Figure 28. Fear
is almost next to a positive polarity item, at least as far as the quotations in the
OED are concerned.91
Secondly, the patterning of complementation differs significantly between
fear in affirmative sentences and the same verb used in negation in the above
graphs. With fear in affirmative sentences, the employment of that-clauses is
distinctively frequent from the sixteenth century to the twentieth century
throughout, whereas with fear in negation, the same type of complementation
occurs much more sparingly. On the whole, to-infinitives and gerunds are more
common than that-clauses with fear in negation, although the dearth of the
overall data of fear in negation in the OED quotations makes it difficult to
draw definite conclusions for some centuries.
In view of the fact that most examples of fear are encountered in affirma-
tive sentences, it is of no surprise that the general tendencies of fear as shown
in Figure 27 above reflect its usage in affirmative sentences (i.e. Figure 28).
Consequently, one is inclined to observe frequent occurrences of that-clauses

91
The relatively infrequent examples of fear in negation include a notable number of ex-
amples with need (e.g. need not fear). The OED quotations contain a total of 62 instances
of fear used in negation and with complements, of which eleven are found with need.
126 Chapter 5

with fear in Present-day English as discussed above. All in all, it is most


important to make a separate treatment of fear in affirmative sentences and the
same verb used in negation, in discussing the historical development of its
complementation patterns.
As concerns fear in negative environments, the historical development is
in many respects similar to that of the other verbs so far discussed, displaying
the familiar shifts of complements, i.e. the rise of to-infinitives and gerunds in
succession. Although not in the clearest form, it is certainly discernible in
Figure 29 that there is a notable attestation of to-infinitives with fear in nega-
tion during the early Modern English period, which is later on overtaken by
gerunds. Unfortunately, the graph does not depict the stage where the use of
that-clauses was dominant, which is due to the fact that the OED quotations do
not include relevant examples before the sixteenth century. Hence, the first
shift or the decline of that-clauses is rather opaque with fear, although the
extensive use of to-infinitives, which is in usual cases a resultant situation of
the first complement shift, is clearly evidenced in Figure 29. Likewise, one
perhaps needs to admit that the second shift or the rise of gerunds at the cost of
infinitives is slightly obscure in the above graph, although the relative propor-
tions of gerunds to infinitives clearly rises throughout the Modern English
period. All in all, however, it is reasonable to assume that fear used in negative
contexts follows the usual path of development, which involves the shift from
that-clauses to to-infinitives first and then the shift from to-infinitives to
gerunds.

5.2.4. The parenthetical use of fear in affirmative sentences

The present section now turns to the historical development of fear in af-
firmative sentences. Although to-infinitives and gerunds are also attested with
fear in affirmative sentences, the complement shifts as described above in rela-
tion to fear used in negation are not at all clear here throughout the history of
English. This is, I propose, due to the fact that the possible instability of that-
clauses from later Middle English to early Modern English, which was equally
observed with fear in affirmative sentences, led to a different path from the
expanded use of to-infinitives, in the case of fear in affirmative sentences.
More specifically, the different path followed by affirmative sentences is the
development of the parenthetical use of fear, as illustrated by the examples
cited below. This is one of the most characteristic features of the verbs
Verbs of implicit negation and subordinate clauses 127

discussed in the present chapter, particularly of fear (discussed in the present


section) and doubt (discussed in 5.3.):

He is altogether French and will seek to draw this King into France, where his life
I fear will be vendible.
(1581 Burghley in Sir Dudley Digges, The Compleat Ambassador)

Take my armour of quickly, ’twill make him swoune, I feare.


(1598 Ben Jonson, Every Man in his Humor)

Though such things pass on those that sermons hear, It will not do with play-
judgers, I fear. (1672 John Lacy, Dumb Lady)

These are unequivocal cases of the usage under discussion, where the status of
I fear is clearly parenthetical and where the original subordinate clause has
almost obtained matrix clause status. In other words, the original super-
ordinate-subordinate relationship is no longer available in these examples. I
fear now functions as a set and almost as a sentential adverb.
Studies discussing this issue are abundant, though not necessarily in the
context of complement shifts as in the present research. As for the definition of
the term “parentheticals”, see Brinton (2008: 7), who remarks: “Parentheticals
are defined by their lack of syntactic connection with the clause to which
there [sic] are attached”. Similarly, Kaltenböck (2005: 21) states: “PC [i.e.
Parenthetical clauses] … have not syntagmatic (i.e. paratactic, hypotactic) link
to their host clauses. They are related to their host by linear adjacency but are
not part of any larger syntactic unit, i.e. they do not form constituents”.92 They
both refer to the disintegration of the expression concerned from the syntax of
the entire sentence.
While parentheticals can be defined most clearly when I fear is dislocated
as illustrated by the examples above, the OED quotations include abundant

92
While “parentheticals” are most commonly used in the sense of so-called “comment
clauses” in existing studies (cf. Brinton 2008: 18), Espinal (1991) employs this term in a
wider sense. She deals with various types of parentheticals, which she calls “disjunct
constituents”, and refers to the sentential type like I fear at the top of her list. Other types in
her list include adverbial clauses (e.g. if that makes you feel any better), prepositional
phrases (e.g. on the contrary), etc. It is also relevant to note that other terms than “par-
entheticals” are used in previous studies. Thompson (2002), for example, employs the term
“epistemic/evidential/evaluative formulaic fragments” (p. 125 et passim) instead of “par-
entheticals”.
128 Chapter 5

ambiguous examples like the following, where the conjunction that is elliptical
or unexpressed and where the verb fear is not dislocated:

I feare our happinesse is at the height.


(1594 William Shakespeare, The Tragedy of Richard III)

I feare I may doe wrong to your sufficiencies in the reporting them.


(1599 Ben Jonson, Cynthias Revels)

In sentences of this kind, it is almost impossible to tell whether I fear is purely


parenthetical or not.93 If it retains its original meaning and function, the clause
where the conjunction that is unexpressed is still a subordinate one, whereas at
the same time it is also possible that the clause is almost like a matrix one and
that its subject is the topic of the entire sentence. In the latter case, I fear is
already parenthetical and the original subordinate clause has obtained matrix
clause status.
The whole issue is essentially a matter of cline. It is, therefore, perhaps
unwise to try to make a clear distinction as to whether a particular case of I fear
has indeed obtained parenthetical status. Kaltenböck (2005: 44) describes the
gradualness in the shift from matrix clauses to parentheticals by using the
concept of “gradience”. Indeed, the path taken by I fear is similar to what
Thompson and Mulac (1991b), whose work is often referred to in the context
of discussing grammaticalisation in previous studies, argue about the develop-
ment of the parenthetical use of I think. They provide the following three
sentences to illustrate the gradual nature of the shift:

93
Spoken English has some phonological clues to make this distinction, while written Eng-
lish does not. Quirk, Greenbaum, Leech, and Svartvik (1985: §15.53) point out that
parentheticals usually have a “separate tone unit” and “are generally marked prosodically
by increased speed and lowered volume”. Despite the comment by Espinal (1991: 734−
735) that prosodic features do not necessarily guarantee the parenthetical status under
analysis, they are certainly of considerable help, at least as far as Present-day English is
concerned. Referring to previous research, Kaltenböck (2005: 44) remarks: “Most studies
… seem to agree that a parenthetical use of initial clauses without that can un-
ambiguously be assumed in case where they are separated by a pause (or comma) from the
following clause”. As Espinal (1991) argues, however, the situation indeed differs
depending upon the language. She states: “some parametric variation exists between
languages which assign comma intonation to disjunct constituents (like English) and
languages which do not seem to have comma intonation in appositive clauses and disjunct
adverbials (like Japanese)” (p. 735).
Verbs of implicit negation and subordinate clauses 129

(a) I think that we’re definitely moving towards being more technological.
(b) I think 0 exercise is really beneficial, to anybody.
(c) It’s just your point of view you know what you like to do in your spare time I
think.
(Thompson and Mulac 1991b: 313)

According to them, the topic of the sentences increasingly moves from the
subject of the matrix clause to the subject of the complement clause as one
goes down this list. The category shift of I think—the shift from the matrix
clause to the parenthetical—is more advanced in (b) than in (a) and in (c) than
in (b). They consider that I think in (b) and (c) is already what they call an
“epistemic phrase”, i.e. parenthetical in use, and that it functions “roughly as an
epistemic adverb such as maybe with respect to the clause it is associated with”
(p. 313). This process is known to have taken place with various other verbs as
well in the history of English. Denison (1998: 259) adds, for example, I guess,
I suppose, I believe, you know, and I hear in his account of this phenomenon in
the period from 1776 to 1997. See also Brinton (2001) for an inventory of
verbs which display the same tendency in the history of English. For a more
classic research into verbs of this kind, see Urmson (1952), who also provides
a list of relevant verbs in English.
The argument by Thompson and Mulac, which Brinton (2008: 35) calls
“the matrix-clause hypothesis”,94 is certainly in keeping with our intuition. A
number of existing studies are also in line with their contention. Huddleston
and Pullum (2002) argue, for example, that I think can be “backgrounded to the
status of a modal qualification, informationally comparable to a parenthetical”
(p. 953). The direction of their argument is certainly the same as that of
Thompson and Mulac (1991b). Quirk, Greenbaum, Leech, and Svartvik (1985:
§15.54) are also in accordance with Thompson and Mulac (1991b) when they
comment on the “reversal in syntactic roles” in relation to the development of
parentheticals, while referring to various types of expressions including I think
and I fear. Once again, this is a matter of scale, or what Aijmer (1997) calls “a
scale of pragmaticalization where one of the end-points is a pragmatic element
or ‘modal particle’” (p. 1).
As in the case of I think described here, the category shift of I fear was
most likely a gradual one, and as a consequence, it is not always easy to judge
whether I fear in a particular example is parenthetical or not. All one can

94
Brinton (2008: 11−12) remarks that this is not the only possible approach. She in fact notes
the existence of another approach, which assumes the independent occurrence of the par-
entheticals and their later insertion into the “anchor” clause.
130 Chapter 5

investigate for the purpose of drawing inferences in this relation is: (i) whether
the conjunction that is elliptical (as in (b) above);95 and (ii) whether the dis-
location of fear takes place in sentences (as in (c) above). As for the former
question, see the table below, which exhibits the raw frequencies of the con-
junction that expressed or unexpressed with fear in affirmative sentences in the
OED quotations:

Table 5. Fear in affirmative sentences and clauses where the conjunction that is expressed
or unexpressed in the OED (raw frequencies)
centuries 13th 14th 15th 16th 17th 18th 19th 20th
that expressed 0 0 4 30 38 17 69 46
that unexpressed 0 0 2 84 150 89 143 44

It is evident in Table 5 that the conjunction that is most commonly unexpressed


from the sixteenth century onwards, although the competitive frequencies be-
tween “that expressed” and “that unexpressed” in the twentieth century are not
necessarily consistent with the overall historical tendency shown in this table
and therefore will need further investigation. For comparison’s sake, Table 6
shows the case of fear in negative environments, where the ellipsis of that is
much less common:

Table 6. Fear in negation and clauses where the conjunction that is expressed or un-
expressed in the OED (raw frequencies)
centuries 13th 14th 15th 16th 17th 18th 19th 20th
that expressed 0 0 0 1 4 0 0 3
that unexpressed 0 0 0 1 4 0 2 0

Unfortunately, the total occurrences of that-clauses (including those with “that


unexpressed”) themselves are limited in the case of fear used in negative
environments, making the interpretation of statistical figures here slightly
difficult. Still, this table most feasibly indicates that the ellipsis of the con-
junction that is less frequent in negative contexts than in the case of fear in
affirmative sentences. Or, one could, at least, confidently conclude that the
resources for the development of the parenthetical use of fear are much more
restricted in the case of negation in any case.
Furthermore, the dislocation of I fear (including other subject forms like
we fear) is a prominent feature of affirmative sentences only, at least as far as

95
As Brinton (2008: 12−14) claims, the deletion of that does not necessarily guarantee the
parenthetical status of I fear, but is certainly one aspect worth considering.
Verbs of implicit negation and subordinate clauses 131

the OED quotations are concerned. A total of 716 instances of that-clauses


(often the conjunction that is elliptical) with fear in affirmative sentences in-
cludes 123 examples of dislocation, whereas the fifteen examples of that-
clauses with fear in negative environments do not include a single instance of
this phenomenon. Here again, there is a notable contrast between fear used in
affirmative sentences and the same item used in negative environments. As
mentioned above, the dislocation can be observed in middle positions within
the sentence as well as in final position. I will repeat in the following some of
the examples cited above to clarify this point:

He is altogether French and will seek to draw this King into France, where his life
I fear will be vendible.
(1581 Burghley in Sir Dudley Digges, The Compleat Ambassador)

Take my armour of quickly, ’twill make him swoune, I feare.


(1598 Ben Jonson, Every Man in his Humor)

Though such things pass on those that sermons hear, It will not do with play-
judgers, I fear. (1672 John Lacy, Dumb Lady Prologue)

Furthermore, the proportions of the dislocation at issue with fear in


affirmative sentences reveal a diachronic increase, as the table below, which
gives the relevant proportions in the OED data, shows:

Table 7. Fear in affirmative sentences and the proportions of dislocation to the totals of
clauses with that unexpressed in the OED
centuries 13th 14th 15th 16th 17th 18th 19th 20th
dislocation 0 0 0 17.9% 15.3% 23.6% 36.4% 27.3%

The overall increase of the proportions of dislocation as shown in this table


reveals the process in which the parenthetical use of fear developed in the
history of English. The proportions of 17.9% and 15.3% in the sixteenth and
seventeenth centuries are not at all small, however. It is, therefore, reasonable,
to surmise that potential candidates for the parenthetical use of fear are already
available at an early stage of its development. This is most likely why to-
infinitives and gerunds are always marginal in respect of the complementation
patterns of fear in affirmative sentences. In other words, the path leading to the
parenthetical use has always been prepared for the same verb in the affirmative
throughout the history of English.
While discussing the development of the parenthetical use of I think, I
132 Chapter 5

guess, etc., Brinton (2008: 223–224) refers to Wierzbicka (2006: 207), who
maintains that the development of this use becomes prominent only in the
eighteenth century, and says that it is in fact much earlier. She considers that it
goes back at least to the Middle English period. See also Rissanen (1991) and
Brinton (1996: 211–263) for some details of the same argument.96 As for the
parenthetical use of I fear, it is difficult to tell its clear state of affairs during
the Middle English period, which is due to the overall lack of relevant ex-
amples in the period in the OED dataset. Table 7 reveals at least, however, the
existence of its use in the sixteenth century and its subsequent increase in later
centuries. The deletion of the conjunction that does not necessarily prove the
existence of the parenthetical feature of I fear, but the dislocation certainly
does, in my view.
Incidentally, “that unexpressed” with fear in affirmative sentences always
includes a large number of I fear and we fear, i.e. examples in the present tense
with first person subjects, showing that parentheticals tend to be fixed in form.
Some illustrative examples follow:97

I fear he hath other fish to fry. (1660 John Evelyn, Memoirs)

I fear me this traveller hath dined but lightly.98


(1828 Sir Walter Scott, The Fair Maid of Perth)

We fear the sons are no great improvement upon the sires.


(1878 Charles Haddon Spurgeon, The Treasury of David; Containing an Original
Exposition of the Book of Psalm)

96
Brinton discusses here the usages of I guesse, I trowe, I woot, and other similar expressions
in Middle English and argues that they are not mere metrical tags but present various
features of pragmatic markers already during the period.
97
Aijmer (1997: 7), referring to Urmson (1963), says that “[o]nly non-factive predicates in
the first person can, for example, be used parenthetically”. Also, the relationship between
the ellipsis of the conjunction that and various linguistic conditions where it occurs has
been extensively discussed in the literature, although previous studies are inclined to deal
with verbs like think, tell and know rather than fear and their focus tends to be placed upon
Present-day English. For further details, see Thompson and Mulac (1991a, 1991b) and
Tagliamonte and Smith (2005).
98
The reflexive use of fear as illustrated by this example is also counted here. The OED (s.v.
fear) states that it is archaic now in the phrase I fear me, although it was constant in earlier
days. Furthermore, Lange (2007: 142) comments on this expression and says that the verb
fear inherited the construction of dreden when it took the place of the latter verb in the
course of the early Modern English period.
Verbs of implicit negation and subordinate clauses 133

See also the table below, which demonstrates how frequent these examples are
in the OED quotations:

Table 8. I fear and we fear in affirmative sentences in the OED


centuries 13th 14th 15th 16th 17th 18th 19th 20th
I fear (I fear me) 0 0 2 70 105 64 109 20
and we fear (100%) (83.3%) (70.0%) (71.9%) (76.2%) (45.5%)
The totals of “that 0 0 2 84 150 89 143 44
unexpressed”

It is difficult to tell from this table alone whether the proportions of I fear and
we fear are rising or falling in historical terms. By excluding the data of the
fifteenth and twentieth centuries, both of which are rather irregular, one could
possibly conclude that the fixed forms I fear and we fear account for around
70% to 80% of “that unexpressed” in affirmative sentences throughout. The
fifteenth century provides too few relevant examples to allow any fair judge-
ments or interpretations, while the situation of the twentieth century may be
indicative of a re-established use of that-clauses in recent years, which will be
touched upon in a later section.
Finally, the proportions of I fear are particularly large when compared with
those of we fear. The totals of I fear and we fear in Table 8 reaches 370, of
which as many as 353 are examples of I fear. Thus, the subject of the matrix
clause of “that unexpressed” has a strong tendency to be the first person sin-
gular and its verb tends to be in the present tense.

5.2.5. Fear and clauses of other types

The discussion has hitherto been focused upon the three complementation
types of that-clauses, to-infinitives, and gerunds and their related forms, all fol-
lowing the verb fear. The gist of the contention has been that separate treat-
ments are necessary between fear used in negation and the same verb in af-
firmative sentences, since their historical developments differ to a significant
extent, at least as far as their complementation patterns are concerned. In the
case of fear in negative environments, the rise of to-infinitives and the rise of
gerunds, both of which are in accordance with the shifts of complements in
general in the history of English, have most probably occurred, although rele-
vant examples are not so copious as to confirm this point with strict certainty.
In the case of fear in affirmative sentences, by contrast, the historical de-
velopment triggered by the decline of that-clauses turns out to be different in
134 Chapter 5

nature. The different development is characterised by the gradual establishment


of the parenthetical use of fear, particularly in the form I fear, which converts
the original subordinate clause into a matrix clause. The conjunction that was
increasingly elliptical in this process.
It is noteworthy, however, that that-clauses are not the only type of clauses
dominated by fear. It is on occasions followed by clauses other than that-
clauses in the history of English. Examples of other types include:

I fear lest it should be discovered by … clear reducting to letters no better than


nonsense.
(1816 Charles Lamb, Final Memorials; Consisting Chiefly of his Letters not be-
fore Published)

I doe not feare but that these few Souldiers will be able to returne againe.
(1641 Thomas Edwards, Reasons against the Independent Government of Par-
ticular Congregations)

The first of the above examples illustrates the employment of lest,99 and the
second the use of but.100 Although other examples like unless-clauses and
whether-clauses are also encountered with fear in the history of English, their
frequencies are extremely low, as far as the OED quotations are concerned. The
present section accordingly deals with the occurrence of lest-clauses and but-
clauses only, both of which are, in fact, relatively commonly evidenced with
verbs of negative connotations in general and not restricted in use with the verb
fear.
Strictly speaking, clauses introduced by lest may not be a pure com-
plement of fear, since they may still retain some adverbial nature.101 They are
worth mentioning in the context of the present study, however, in two respects.
First of all, their examples are notably abundant—84 in total in the OED quo-
tations—, displaying how essential their existence is in the discussion of the

99
Unless otherwise stated, lest here includes examples of lest followed by the conjunction
that (i.e. lest that).
100
But here includes examples of but that. See also Note 7 in Chapter 1.
101
Whether or not this use of lest is adverbial, it differs functionwise from the usual type of
lest, which introduces a final clause. Kikusawa (2009) notes, for example, that the sub-
junctive is far less frequent when lest is employed as a “complement”, i.e. after negative
predicates like fear, than when it is used as an ordinary adverbial clause, where the
subjunctive outnumbers the use of modal auxiliaries.
Verbs of implicit negation and subordinate clauses 135

syntactic development of fear in the history of English.102 Secondly, they seem


to exhibit an almost complementary distribution to but-clauses, which are cert-
ainly related to the decline of that-clauses, in respect of their occurrences as a
complement of fear. The OED (s.v. fear) states that the employment of but-
clauses is found in negative sentences, and indeed four examples of but-clauses
in the data of the present research are all attested with fear used in negative
environments. By contrast, the data of the present investigation reveal that lest-
clauses are strongly inclined to take place with fear in affirmative sentences. Of
the 84 examples of lest-clauses attested in the OED quotations, as many as 82
are evidenced in the affirmative context. Obviously, the fact that lest-clauses
are much more frequent than but-clauses is attributable to the general character
of fear, whose occurrences are skewed towards affirmative sentences.
While discussions on lest-clauses in existing research are sparse, but-
clauses, which are often discussed in the literature, are considered to be a form
which took the place of that-clauses with expletive negation and therefore to
have commonly been attested during the early Modern English period when the
phenomenon of expletive negation receded in the history of English (see 1.1. in
the Introduction). As far as the relevant examples of the verb fear are con-
cerned, three of the four examples of but are indeed found in the seventeenth
century, but there is a single lingering instance of it in the nineteenth century in
the OED quotations. Similarly, lest-clauses as a complement of fear were
observable up to the nineteenth century. Later on, their use quickly declines,
and by the time of the twentieth century they are almost extinct. In the OED
quotations, there are no twentieth-century examples of clauses introduced by
but or lest with fear. Relevant examples are also unavailable in the twentieth-
century data of LOB and FLOB.103

5.2.6. Additional comments on that-clauses

In this final section of the verb fear, I would like to make some additional

102
The OED (s.v. lest) provides a separate definition for the type of lest that occurs after
verbs of fearing and that can be replaced by the conjunction that.
103
Interestingly enough, the Brown corpus, which includes American English of the 1960s,
presents a single instance of lest used after fear: He suggested offering half to Sir Edward,
fearing lest “he shall thinke it to good for us and procure it for himselfe, as he served us
the last time” (Brown, G). The participle form fearing lest … as illustrated by this
example seems to be rather idiomatic, considering the fact that it also occurs from time to
time in the historical dataset of the OED quotations.
136 Chapter 5

comments in relation to the state of the verb in recent years. Although the
overall tendencies of the historical development of fear are fairly clear as
explored above, its situation in the twentieth century is erratic in some respects.
The ellipsis of the conjunction that is, for example, a marked feature with fear
in affirmative sentences in general, but in the twentieth-century data of the
OED, clauses with “that expressed” make a good competition with those with
“that unexpressed”. See Table 5 (p. 130), which seems to display the revival of
the conjunction that in the twentieth century. Also, the dislocation of fear,
which is on the steady rise in the history of English, suddenly decreases in the
twentieth century, as far as its proportion is concerned. See Table 7 (p. 131) to
reconfirm this point. Moreover, the ratio of the fairly fixed forms of I fear and
we fear is also smaller in the twentieth century than in the other centuries in the
OED data. In other words, the parenthetical use of fear seems to recede to a
certain extent in the twentieth century in the data so far explored.
Taking into account the fact that the OED quotations in the twentieth cen-
tury are much more limited in number than in the nineteenth century,104 an
immediate conclusion as to the twentieth century situation should be avoided in
discussion here. Further investigation of this matter is certainly called for in
future research. It is also possible that twentieth-century materials in the OED
quotations happen to be more strongly edited than those in the other centuries,
and this may have often led to the introduction of the conjunction that in an
“expressed” form in the twentieth century. I would, therefore, like to stress the
importance of further research once again.
Simultaneously, however, one could possibly postulate that the use of
subordinate clauses introduced by that was re-established in some measure in
the twentieth century after a long period of various competitions among
different complementation patterns including not only that-clauses, to-infini-
tives, and gerunds but also clauses introduced by but and lest. The twentieth
century is certainly a period when things were settled and consolidated and
when the number of variant forms was reduced. In other words, the forms
which survived a long period of competitions in the history of English have
now regained and reconfirmed their own statuses in the twentieth century.105

104
In respect of the number of quotations in different centuries in the OED, see 1.5. and 3.2.3.
above.
105
The re-establishment of the use of that-clauses could be a general feature of contemporary
English and not confined to the verb under consideration here. Fanego (1990: 141–146)
discusses the frequency of the ellipsis of that in Shakespeare’s English, pointing out that
it is much more frequent than in Present-day English, where that is relatively often “ex-
pressed”.
Verbs of implicit negation and subordinate clauses 137

This can possibly be one of the reasons why that-clauses seem to have a
stronger profile with this verb in the twentieth century than in the nineteenth
century, when the rise of the parenthetical use of I fear and we fear reaches the
summit.

106
5.3. The historical development of doubt
5.3.1. Preliminary remarks

The second verb investigated in the present chapter is doubt. It is similar to


fear, in that it also developed a parenthetical use in the history of English, as
the following accounts describe.107 With this verb, too, the decline of that-
clauses took place from later Middle English to early Modern English, which
led to the increasing use of I doubt as an epistemic parenthetical at the sen-
tential level. As in the case of fear, the development of the parenthetical use is
particularly noticeable in affirmative sentences, and therefore the discussion in
the sections that follow is again separated between doubt in affirmative sen-
tences and doubt used in negation.
One difference between fear and doubt is that the latter is, in general, more
characteristically followed by subordinate clauses (rather than non-finite com-
plement forms), which include not only that-clauses but also whether-clauses
and if-clauses.108 Some notable expansion of to-infinitives and gerunds is at-
tested with fear as discussed under 5.2., especially in negative sentences, but
the employment of these non-finite constructions is much more restricted with
doubt. It is for this reason that the discussion below is centred more upon the
historical development of various types of finite clauses than upon non-finite
forms.
As for Present-day English, doubt is indeed followed typically by that-
clauses and whether-clauses, the choice of which depends upon whether it
occurs in affirmative sentences or in negative ones. As Dixon (1991: 216)
maintains, whether-complements are the usual form dominated by affirmative
doubt in contemporary English, although that-clauses are also observed in the
106
An earlier version of 5.3. appeared as Iyeiri (2009b). John Benjamins kindly granted
permission to use it in the present study.
107
Herbst, Heath, Roe, and Götz (2004: 248, 304–305) do not mention the parenthetical use
of doubt, whereas they deal with this usage clearly under the entry of fear. The par-
enthetical use of I doubt is, however, certainly attested both in the historical data and in
contemporary English, as I discuss in the present chapter.
108
See also Note 11 in Chapter 1.
138 Chapter 5

same syntactic context. In the case of negative doubt, by contrast, that-clauses


are the usual form of complementation, except in negative questions, where
clauses introduced by whether are observed. To illustrate this point, I have
cited two examples from the BNC in the following as representative ones in
respect of their syntactic structures:

At first sight, the reader might doubt whether even people without brain injury
usually possess significant lip-reading skills.
(1991 Raymond Tallis and Robinson Howard, The Pursuit of Mind, from BNC)

Now I do not doubt that Mr Neighbours had good organizational skills; he did, I
understand, mastermind a number of large occasions with conspicuous style.
(1989 Kazuo Ishiguro, The Remains of the Day, from BNC)

The first of the above examples illustrates the case where affirmative doubt is
followed by whether, while the second exemplifies the case where negative
doubt is followed by a that-clause. This distinction between affirmative and
negative sentences in contemporary English is mentioned in a number of usage
guides and dictionaries (see Burchfield 1998: 229, Greenbaum and Whitcut
1996: 223, among others).
Simultaneously, however, it is often claimed that that-clauses are on the
increase nowadays even in affirmative sentences. Burchfield (1998: 229) con-
siders that the use of that-clauses dominated by affirmative doubt shows a
progressive increase from “the last quarter of the 19c. or so”, and says that the
meaning of such sentences is always ‘think it unlikely’. Greenbaum and Whit-
cut (1996: 223) point out the same tendency in Present-day English, while
referring to American and British varieties: “In American use, and increasingly
in British, doubt that is also used in positive sentences, meaning ‘think it
unlikely’”. They illustrate this point by giving the example I doubt that he’ll
come, which means ‘I think he will not come’. Furthermore, they maintain: “In
questions, doubt that and doubt whether are both available, according to the
degree of likelihood” (p. 223). Thus, the meaning of the verb is most likely
relevant to the construction of doubt, which will be considered in the dis-
cussion below, although I am conscious at the same time that too much
emphasis should not be placed upon meanings as I mention throughout the
present research.
In the following sections, I will also refer to clauses introduced by but,
which, as in the case of fear, also appear as subordinate clauses of doubt,
particularly in the historical data. Examples of but-clauses are considered to
Verbs of implicit negation and subordinate clauses 139

characterise the period when that-clauses with expletive negation declined (see
1.1. in the Introduction), and indeed the OED does not provide any relevant
illustrations from the twentieth century. But-clauses remain with doubt, how-
ever, though limited in frequency, even today. The following are some ex-
amples of but-clauses cited from the BNC:

She did not doubt but that somewhere amongst the records and corporate memory
of the Society would be someone who knew Hereward Marr very well indeed.
(1991 Diane M. Greenwood, Unholy Ghosts, from BNC)

I can see no more reason to doubt but that these causes in a thousand generations
would produce a marked effect, and adapt the form of the fox to catching hares
instead of rabbits, than that greyhounds can be improved by selection and careful
breeding.
(Peter J. Bowler, The Fontana History of the Environmental Sciences, from BNC)

Burchfield (1998: 229) considers that examples of but used with doubt are
somewhat archaic in Present-day English but are evidenced when the main
clause is negative or interrogative. The above two sentences from the BNC are
indeed found in negative contexts. As for American English, which is not
entirely related to the discussion of the present study but which is interesting in
this relation, The American Heritage Book of English Usage (1996: 91) takes
the view that but-clauses dominated by doubt are attested in informal speech
today, although they were once used by “refined” writers. In the discussion that
follows, clauses introduced by lest will also be discussed in relation to but-
clauses.109

5.3.2. Doubt in affirmative sentences and various types of subordinate clauses

As hitherto discussed, subordinate clauses, when doubt is in the affirmative, are


usually introduced by the conjunction whether in Present-day English. As far as
the quotations in the OED are concerned, however, whether-clauses do not al-
ways seem to predominate in this syntactic environment in the history of Eng-
lish. See Table 9, which displays the raw frequencies of various types of
subordinate clauses dominated by affirmative doubt in the OED quotations:

109
Note that but-clauses and lest-clauses are discussed together in the section of fear as well
(cf. 5.2.5. above).
140 Chapter 5

Table 9. Doubt in affirmative sentences and different subordinate clauses in the OED (raw
frequencies)
centuries 13th 14th 15th 16th 17th 18th 19th 20th
clauses introduced by that 0 2 9 13 44 25 30 14
clauses introduced by but 1 0 0 4 1 0 2 0
clauses introduced by lest 0 1 1 7 1 0 0 0
clauses introduced by whether 0 2 0 12 19 19 72 20
clauses introduced by if 0 0 0 1 4 2 24 27
other clauses 0 0 3 9 5 1 4 0

As repeatedly mentioned in the present study, the number of quotations per


century is not constant in the OED, and therefore it is not wise to compare and
contrast the raw frequencies of different centuries under the same type of
clauses. Still, this table explicates some interesting points about the uses of
doubt in affirmative sentences: (a) indeed subordinate clauses introduced by
whether are attested commonly, especially in Modern English, but other types
are also available; (b) if-clauses are increasingly common in the nineteenth and
twentieth centuries; (c) that-clauses are moderately common from late Middle
English onwards, despite the above-mentioned claim in the literature that the
employment of that-clauses following affirmative doubt has been a recent
phenomenon; and (d) subordinate clauses introduced by conjunctions other
than whether, if, and that are also available, though not frequent, throughout
the history of English.
The existence of various types of subordinate clauses most probably
reflects various different levels of likelihood expressed by doubt. I have al-
ready referred to the contention by Greenbaum and Whitcut (1996: 223) that
the choice of whether or that after doubt in affirmative sentences depends upon
the degree of likelihood of the proposition of the subordinate clause. On the
whole, the use of that indicates that the content of the subordinate clause is un-
likely, whereas the employment of whether signals genuine uncertainty. Al-
though whether is common with doubt, however, whether … or not is not
allowed according to Huddleston and Pullum (2002: 983). They maintain that
the “open type”, to which they say whether … or not belongs, does not occur
as a subordinate clause dominated by doubt. They present the following sen-
tences, and argue that only (a), which is a closed type, is acceptable with doubt:

(a) I doubt whether he wrote it.


(b) *I doubt whether he wrote it or not.
(c) *I doubt who wrote it.
Verbs of implicit negation and subordinate clauses 141

I am uncertain, however, as to how strict this rule is, especially in the context
of the history of English, since examples of what they call the open type are
available, though not frequently, in the OED quotations. All the examples clas-
sified under the category of “other clauses” in Table 9 are of this type. For
instance:

He doubted how he was garnysshed of his meyneyall seruantis.


(1494 Robert Fabyan, The Newe Chronycles of Englande and of Fraunce)

Some doubted how far such volage expressions inferred treason, being but
lubricum linguæ
(a1722 Sir John Lauder Fountainhall, The Decisions of the Lords of Council and
Session)

When used in this construction, the meaning of doubt is most probably very
close to ‘wonder’. The Bible in English also provides about 30 examples of
this type in total. Thus, it is evident that various degrees of likelihood are ex-
pressed by doubt, not only in Present-day English but also throughout the
history of English. And the types of subordinate clauses dominated by doubt
seem to be the key, in a way, to indicate the degree of likelihood. In this sense,
the meaning is relevant to the choice of different types of subordinate clauses
dominated by doubt.

5.3.3. Doubt in affirmative sentences and subordinate clauses introduced by


whether and if

As the preceding sections have shown, whether-clauses are regarded as the


standard form of complementation after affirmative doubt in Present-day Eng-
lish. The present section reveals that this is something that has gradually been
established in the historical development of doubt. To ascertain the chrono-
logical expansion of whether-clauses, I have calculated the proportions of
whether-clauses (together with if-clauses) to the totals of all types of finite
subordinate clauses dominated by doubt in affirmative sentences, using the
data from the OED quotations. See the graph that follows:
142 Chapter 5

100

80

60

40

20

0
13th c. 14th c. 15th c. 16th c. 17th c. 18th c. 19th c. 20th c.

whether-clauses if-clauses

Figure 30. Doubt in affirmative sentences and the proportions of whether-clauses and
if-clauses to the totals of all types of subordinate clauses in the OED (%)

Some illustrative examples of clauses introduced by whether and if are given


below:

At least a few Conservatives ... doubt whether Mrs Thatcher was prudent to
appear to rattle sabres and remind electors that women national leaders ... have
sometimes looked more warlike than men. (1976 Times, 27 January)

I doubt if you would really machine-wash a child’s walking shoe.


(1970 Guardian, 28 May)

Burchfield’s (1998: 229) account that the use of that-clauses after affirmative
doubt is a feature particularly observed from the last quarter of the nineteenth
century suggests that the use of whether-clauses has always been the well-
established norm. This is not supported, however, by the data under con-
sideration. As a matter of fact, it is whether-clauses and also if-clauses (rather
than that-clauses) that display historical expansion in affirmative sentences of
doubt. In other words, doubt came more and more to refer to a fairly neutral
point of uncertainty instead of extreme unlikelihood, as far as its use in af-
firmative contexts is concerned. The exceptionally large ratio of whether-
clauses in the fourteenth century in the above graph is ascribed to the overall
infrequency of relevant examples (see Table 9 on p. 140), which makes the
statistics rather unreliable. Apart from this, the proportions of whether-clauses
and if-clauses dominated by doubt in affirmative sentences were on a steady
increase in the history of English and nearing 80% in the twentieth century.
The relationship between if-clauses and whether-clauses is also interesting
in Figure 30, where one can observe a notable expansion of the former in the
Verbs of implicit negation and subordinate clauses 143

nineteenth and twentieth centuries. In fact, the extended use of if-clauses is


probably the factor behind the slight decrease of the use of whether-clauses in
the twentieth century in this graph. This is most likely a stylistic matter as well
as a chronological one. As for Present-day English, Whitcut (1994: 95) con-
siders that whether is formal usage and that if is informal. Also suggestive is
the fact that the Bible in English on CD-ROM provides no examples of if-
clauses, while it yields thirteen examples of whether-clauses in Modern Eng-
lish translations. It is a plausible inference that the linguistic style of the Bible
did not welcome the employment of if-clauses.

5.3.4. Doubt in affirmative sentences and subordinate clauses introduced by


that during the critical period

As far as Table 9 (see p. 140) is concerned, that-clauses after doubt in af-


firmative sentences seem to be better-rooted in the history of English than
whether-clauses. That-clauses are available from the earliest period of the
history of doubt. Still, a closer look at the data suggests that the status of that-
clauses used with doubt in the affirmative was not always stable in the history
of English, which the present section intends to clarify.
First of all, doubt also saw the rise of to-infinitives in place of that-clauses
from late Middle English to early Modern English to some extent, in the same
manner as other verbs of implicit negation did. In other words, there was
certainly a sign of the complement shift from that-clauses to to-infinitives in
the historical development of doubt used in affirmative sentences as well. See
the following examples which illustrate the employment of to-infinitives by
doubt:

I doubte me to haue shortly a strong werre & to haue a doo with a strong partye.
(c1500 Jean D’Arras, tr., Melusine)110

110
This example illustrates the occurrence of a reflexive pronoun after doubt. This phe-
nomenon is attested up to the early Modern English period as far as the OED quotations
are concerned. The latest example I have found is: I doubt me whether the very sober-
nesse of such a one, like an unlicour’d Silenus, were not stark drunk (1642 John Milton,
An Apology against a Pamphlet Called A Modest Confutation of the Animadversions upon
the Remonstrant against Smectymnuus).
144 Chapter 5

They doubted to fall in their handes.


(1568 Richard Grafton, A Chronicle at Large and Meere History of the Affayres of
England)

The OED yields a total of 27 instances of to-infinitives dominated by doubt in


the affirmative, of which as many as 20 are attested in the sixteenth and seven-
teenth centuries. Despite the irregularities of the number of quotations in dif-
ferent centuries in the OED, it is reasonably safe to conclude that the rise of to-
infinitives as illustrated by the above two examples was particularly note-
worthy during the early Modern English period, since the three centuries after
the early Modern English period, which provide a far larger number of quota-
tions in the OED, do not include as many examples of to-infinitives. Hence, the
rise of to-infinitives during the early Modern English period did not lead to
their chronological increase of this new construction in the end. On the con-
trary, to-infinitives dominated by doubt were increasingly marginal and became
obsolete before long. As mentioned above, it is perhaps necessary for a certain
construction to be established in historical terms to have a fairly strong impetus,
which was simply unavailable with the to-infinitival construction dominated by
doubt in affirmative sentences. Examples of doubt in this construction often
conveyed the meaning ‘fear’ as illustrated by the above examples. In this sense,
the meaning is relevant to the understanding of syntactic structures.
Incidentally, there are also less than ten examples of gerunds used with
doubt in affirmative sentences in the OED quotations as well, suggesting that
the second complement shift or the rise of gerunds was also a possible path
with this verb, at least in theory. The use of gerunds, however, has always been
marginal with it throughout the history of English. Relevant examples include:

The governor is pleased to doubt our having such letters as we mentioned.


(1759 Benjamin Franklin, The Works of Benjamin Franklin)

Examples of the gerundial construction were far too restricted to be established


as a form of complementation of doubt in the history of English. It was also
devoid of impetus, which is necessary in the establishment of a newly arising
syntactic form.
In the case of other verbs of implicit negation (see above chapters), the
development of non-finite forms like to-infinitives and gerunds in a way solved
the problem of how to deal with the obliteration of expletive negation from late
Middle English to early Modern English, since that-clauses which contained
expletive negation themselves declined. Doubt, however, had to deal with the
Verbs of implicit negation and subordinate clauses 145

matter in another way, since the development of non-finite complement forms


was marginal with it, as hitherto discussed. On the other hand, the use of
that-clauses was continuous with doubt even during the Modern English period.
And as in the case of fear, the employment of but-clauses instead of that-
clauses plus expletive negation was one way to solve the problem. Indeed,
but-clauses are attested with doubt during the critical period from late Middle
English to early Modern English as Table 9 (see p. 140) shows. As in the case
of fear again, however, the use of but is a more prominent feature of doubt in
negative sentences than in affirmative ones.
Apart from negative sentences, interrogative sentences are also considered
to yield but-clauses (see Curme 1931: 241). As for the data under investigation,
there are a total of eight relevant examples of but-clauses in non-negative sen-
tences (cf. Table 9), three of which are indeed found in interrogative sentences
as the following illustration shows:

Who can doubt any longer, but that you pricke at relygion?
(1560 John Daus, tr., A Famous Cronicle of oure Time Called Sleidanes Com-
mentaries)

Considering the rhetorical feature of interrogation here, the entire tone of this
sentence is more inclined to negation than affirmation. In other words, in-
terrogation can often be similar in tone to negation, and thus it is reasonable for
interrogative sentences to share some linguistic features with negative sen-
tences.
The same applies to the three examples of but in non-negative sentences in
the Bible in English, all of which are found in the unstable period, when the
phenomenon of expletive negation was in the process of recession or had just
receded. Challoner’s translation of the Bible, from which the following ex-
ample is cited, goes back to 1750–1752:

Who can doubt but that you intend some great evil? (Challoner, Exodus, 10:10)

Certainly, the above example of but following doubt is found in an inter-


rogative sentence.
Thus, but-clauses used with doubt are largely restricted to negative and
interrogative sentences and mostly found during the transitional period from
late Middle English to early Modern English, although examples like the fol-
lowing are not entirely unavailable in the OED quotations:
146 Chapter 5

I doubt but the man I let the land to is just making a kirk and a mill of it.
(1887 Mrs. Alexander, Mona’s Choice)

Alternatively, what is characteristic of doubt in affirmative sentences is the


employment of lest-clauses instead of that-clauses plus expletive negation.
This is again observed during the critical period from late Middle English to
early Modern English (cf. Table 9 on p. 140). The following are some il-
lustrations of this usage:

He douted lest in his absence there shuld arise some chaunge or mutacyon in
Fraunce. (a1470 John Tiptoft, tr., Iulius Cears Commentaryes)

I doubt lest we are gone out of the waye.


(1583 Claudius Hollyband, Campo di Fior, or else the Flourie Field of Foure
Languages)

Examples of lest-clauses are as restricted as but-clauses in frequency, but the


existence of this type itself is noteworthy, in that it is virtually confined to af-
firmative sentences as in the case of fear discussed above. It is, therefore, most
feasible that but-clauses and lest-clauses are distributed more or less com-
plementarily. This is in keeping with the distribution of but-clauses and lest-
clauses which occur after the verb fear as discussed under the section 5.2.5.
The above examples of lest-clauses suggest that the meaning of doubt in this
case is close to ‘fear’.

5.3.5. Doubt in affirmative sentences and the development of the parenthetical


use

Finally, there is one major and important aspect in the development of doubt
followed by that-clauses in affirmative sentences, and this is again similar to
the historical development of fear. Namely, doubt in affirmative sentences
displays a gradual development of the parenthetical use as illustrated by the
following instance:

This, I doubt, will prove an Utopian conceit.


(1659 Brian Walton, The Considerator Considered)
Verbs of implicit negation and subordinate clauses 147

I doubt in this example functions almost as an epistemic adverb at the sen-


tential level, while the original subordinate clause has become the matrix
clause. Hence, the decline of the original that-clause has occurred.
As in the case of fear, the development of I doubt, etc. as an epistemic
parenthetical is a gradual one, and can be traced in the form of the increasing
loss or the ellipsis of the conjunction that as shown in Table 10, which gives
the raw frequencies of “that expressed” and “that unexpressed” following
doubt in the affirmative in the OED quotations:

Table 10. Doubt in affirmative sentences and clauses where the conjunction that is ex-
pressed or unexpressed in the OED (raw frequencies)
centuries 13th 14th 15th 16th 17th 18th 19th 20th
that expressed 0 2 8 8 5 1 11 11
that unexpressed 0 0 1 5 39 24 19 3

Some examples with or without the conjunction that are:

Her Highness dowteth that yt may breed discredyt to dyvers of great quarrell.
(1590 Thomas Heneage, Lett. Lit. Men)

We doubt he will execute it with seuerenesse.


(1597 I. T., Sermon Preached at Paules Crosse)

Despite the existence of some exceptional centuries, it is immediately clear


from the above table that the conjunction that was increasingly unexpressed in
later Modern English, at least up to the nineteenth century.
I have made a further detailed analysis of the examples classified under
“that unexpressed” and found that most examples of elliptical that occur with
first person subjects and in the present tense, satisfying some conditions to be
an epistemic parenthetical. For instance, the seventeenth century provides 39
instances of “that unexpressed”, of which 33 are found with first person
subjects and in the present tense. The tendency is even more prominent in the
eighteenth century, when all the 24 instances of “that unexpressed” display
doubt in the present tense with first person subjects.
Presumably, expansion of the elliptical use of that is certainly a step to-
wards the development of the epistemic use of I doubt, although in practice, the
majority of the examples of elliptical that in the OED are like the following,
where it is difficult to judge purely from the form whether the clause after
doubt is a matrix one or a subordinate one:
148 Chapter 5

I doubt you want a spurrer-on to exercise and to amusements.


(1728 Jonathan Swift, Letter to Pope, 16 July)

I doubt here may still retain its matrix clause status, while it may already have
been established as a set to function almost adverbially at the sentential level.
It is perhaps of no use to try to make this distinction.
To summarise the historical development of doubt in affirmative sentences,
it utilised various means to survive the critical period in the history of English
when the status of that-clauses was unstable. The employment of to-infinitives
and occasionally of gerunds was certainly one way, which is in line with the
path taken by various other verbs of implicit negation (and perhaps those verbs
which are not at all negative as well), but ultimately it was not established with
affirmative doubt in the history of English. Similarly, conjunctions like but and
lest were introduced from later Middle English to early Modern English instead
of that-clauses with expletive negation, but their examples are again restricted
to a small number of cases with doubt.
The most powerful means to solve the problem of the weakening of
that-clauses was, by contrast, the development of I doubt as an epistemic
parenthetical in the case of affirmative sentences. This extinguished the
necessity of the existence of complement clauses themselves. An additional
bonus was the obvious fact that the gradual development of the epistemic
parenthetical blurred the problem of the decline of expletive negation. In the
example below, where the conjunction that is unexpressed, negation may or
may not be expletive depending upon the status of I doubt:

I doubt their preaching is not alwaies true …


(1640 Christopher Harvey, The Synagogue, or, the Shadow of the Temple)

However, the development of I doubt as an epistemic parenthetical was not


the ultimate solution, since its use was or has been stylistically conditioned in
English. In the Bible in English, the epistemic use of I doubt is not observed at
all. This is most likely due to the bookish style of this genre, although the in-
fluence of the tradition of translating the Bible may also be relevant. As in the
case of fear, the use of that-clauses seems to have been re-established with
doubt in recent years in the end. Most grammars and standard usage guides of
Present-day English note the use of that-clauses after doubt in affirmative
sentences as well as in negative ones. Table 10 above also exhibits somewhat
more expanded use of the conjunction that in the twentieth century than in the
eighteenth and nineteenth centuries. Once the critical period was over, it was
Verbs of implicit negation and subordinate clauses 149

perhaps easier again for doubt to re-select that-clauses, just in the same way as
the verb fear saw the revival of that-clauses in recent years. This issue will be
discussed later under the section of negative doubt again, which also shows the
re-established use of that in recent centuries.

5.3.6. Doubt in negative sentences and various types of subordinate clauses

The discussion now turns to doubt used in negative environments, and the
present section, in particular, gives an overall survey of what kind of sub-
ordinate clauses negative doubt dominates in the history of English. See Table
11, which shows the relevant raw frequencies in the OED quotations:111

Table 11. Doubt in negation and different subordinate clauses in the OED (raw fre-
quencies)
centuries 13th 14th 15th 16th 17th 18th 19th 20th
clauses introduced by that 2 4 6 14 17 17 47 16
clauses introduced by but 0 1 0 37 36 23 8 0
clauses introduced by lest 0 0 0 0 0 0 0 0
clauses introduced by whether 0 0 0 1 0 0 0 1
clauses introduced by if 0 0 0 0 0 0 1 0
other clauses 0 0 0 0 2 2 2 1

Unlike the case of doubt in affirmative sentences, examples in this table are
largely confined to that-clauses and but-clauses. Clauses introduced by whether
and if are also evidenced as illustrated in the following quotations, but there are
only a few relevant instances in all in the entirety of the OED data:

We maye not Iewishlye doubte whether it be done.


(1558 Thomas Watson, Holsome and Catholyke Doctryne Concerninge the Seven
Sacramentes)

I could not help doubting if everything was done on the square, as they say.
(1866 George Macdonald, Annals of a Quiet Neighbourhood)

As in the case of doubt in affirmative sentences, the category of “other clauses”


in Table 11 includes examples like the following:

111
In an earlier version of my discussion on doubt, there was an error in this table, which has
been corrected in the present study.
150 Chapter 5

He … cannot doubt how little Credit the Quotations deserve, where the Originals
are wanting. (1690 John Locke, An Essay Concerning Human Understanding)

As mentioned above again, the “open type” of complements as illustrated by


this sentence is not allowed in handbooks or grammars of Present-day English,
although its examples are attested in the historical data of the OED quotations.
The Bible in English presents a few instances of this kind as well.
The remaining point to make is that lest-clauses are absent in the case of
negative doubt. Instead, there are fairly numerous examples of but-clauses used
with doubt in negation. Judging from Table 11, the clauses to be discussed here
are that-clauses and those introduced by but, which the following sections con-
centrate upon.

5.3.7. Doubt in negative sentences followed by that-clauses and but-clauses

As some other verbs of implicit negation and doubt in affirmative sentences did
during the few hundred years from later Middle English, when the status of
that-clauses was unstable due to the decline of the subjunctive, doubt in nega-
tion also developed some uses of to-infinitives and gerunds in this transitional
period. While gerunds are too infrequent to reveal unequivocal tendencies in
the data under analysis, to-infinitives dominated by doubt in negation are
indeed inclined to be evidenced particularly in early Modern English. There are
41 instances of to-infinitives following negative doubt in the OED quotations,
of which 33 are encountered in the sixteenth and seventeenth centuries. See the
following examples:

We need not doubt to take away and freely to coerce that improficuous matter of
hair.
(1650 John Bulwer, Anthropometamorphosis; Man Transformed, or the Artifi-
cial Changeling)

The meaning of doubt in this example is something close to ‘hesitate’ as a nat-


ural consequence of the fact that the semantic subject of the to-infinitive is the
same as the subject of the main clause.
Unlike other verbs of implicit negation, however, doubt did not see the
ultimate establishment of non-finite forms of complements even in negative
environments. This is a noteworthy contrast with fear discussed in the present
chapter, which developed an epistemic parenthetical use in the affirmative, but
Verbs of implicit negation and subordinate clauses 151

which developed some extended use of to-infinitives and gerunds in negative


environments in its historical development, eliminating the problem of the
decline of expletive negation. Consequently, it was necessary for doubt in
negation to find another way to deal with the problem caused by the decline of
expletive negation. One obvious way was again to utilise but-clauses in place
of that-clauses with expletive negation. This development is particularly promi-
nent after negative doubt, and as Table 11 shows, particularly so during the six-
teenth and seventeenth centuries, i.e. the transitional period related to the de-
cline of that-clauses. It was a useful means to keep the tone of negation, which
was once involved in that-clauses.
As discussed by Warner (1982: 222–223) and López-Couso and Méndez-
Naya (1998) among others, the increasing replacement of that-clauses with
expletive negation by but-clauses is a general phenomenon to be noted with
various verbs from later Middle English. And indeed, many verbs so far dis-
cussed show the expanded use of but-clauses during this period. As far as my
reading experience is concerned, however, the employment of but-clauses is
particularly noticeable with doubt used in negative environments. The fact that
doubt did not rely upon the use of to-infinitives or gerunds in the end, though it
did for sometime in the history of English, may have instigated its frequent
employment of but-clauses. The following are some illustrative examples in
which but-clauses are employed after doubt in negation:

He did not at all doubt but that they would find matter enough to shop the
evidence himself before the next jail-delivery.
(1771 Tobias Smollett, The Expedition of Humphry Clinker, 11 June)

I do not doubt but you want constant every-day debaters.


(1804 George Rose, The Diaries and Correspondence)

In fact, but is occasionally found even before to-infinitives, as the following


example illustrates:

Hereafter I doubt not but to give you satisfaction that I am not worthy of this
wrong. (1601 Lord Mountjoy, Moryson Itinerary)

The use of but in this example is unnecessary, but it may have occurred under
the influence of but-clauses, which were relatively commonly attested with
negative doubt after the decline of expletive negation.
Examples of but-clauses are, however, limited largely to the transitional
152 Chapter 5

period of the development of doubt. As Table 11 demonstrates, their occur-


rences are inclined to be common particularly during the few hundred years
from the sixteenth century. Once expletive negation was successfully deleted
and the critical period was over, the same verb saw the re-establishment of
that-clauses, as shown in Table 11. In the nineteenth and twentieth centuries,
the employment of that-clauses as in the following example is again regarded
as standard after negative doubt:

We do not doubt that the ombudsman system would work well here.
(1970 Morris and Hawkins, Honest Politician’s Guide to Crime Control)

The re-establishment of that-clauses with doubt in negative environments in


recent years is in accordance with the same development with doubt in affirma-
tive sentences. It is also in accordance with the development of fear discussed
above in the present chapter. While examples of but are still witnessed with
doubt in Present-day English, its use is considered to be archaic or even in-
formal today, as mentioned in the introductory section of this verb (see 5.3.1.
above).

5.3.8. Doubt in negative sentences and the parenthetical use

Finally, I would like to discuss whether one can observe any features of the de-
velopment of negative doubt into an epistemic parenthetical in the history of
English. In the case of affirmative sentences, doubt in the first person and in
the present tense—most typically I doubt—was increasingly epistemic and
used as a parenthetical at the sentential level. Although this is usually con-
sidered to be a typical feature of verbs in affirmative sentences, doubt in nega-
tive environments also shows on occasions the deletion of the conjunction that
and the dislocation of I doubt not or I don’t doubt. This signals a possible
existence of this path of development with doubt in negation as well. Examples
include:

Father Smith prayed for our scow crew, I doubt not.


(1775 John Adams, Familiar Letters of J. A. and his Wife, Abigail Adams, during
the Revolution)

He’s the making of a very nice horse, I don’t doubt.


(1858 Anthony Trollope, Doctor Thorne)
Verbs of implicit negation and subordinate clauses 153

In these examples, the original subordinate clause functions almost as a matrix


clause and I doubt not and I don’t doubt occur in sentence-final position almost
parenthetically.
On the whole, however, the contrast between “that expressed” and “that
unexpressed” is not as striking in negative environments as in affirmative sen-
tences. See Table 12 below, which displays the raw frequencies of the relevant
examples in the OED dataset:

Table 12. Doubt in negation and clauses where the conjunction that is expressed or un-
expressed in the OED (raw frequencies)
centuries 13th 14th 15th 16th 17th 18th 19th 20th
that expressed 2 4 2 3 6 6 40 13
that unexpressed 0 0 4 11 11 11 7 3

This table also shows that the expansion of “that unexpressed” is observed
during the few hundred years after the turn of the sixteenth century, but the
overall contrast between the presence and the absence of the conjunction that is
not very striking in Table 12. Or at least, the contrast is not so striking as in the
case of doubt in affirmative sentences.
In this manner, epistemic uses are inclined to develop much more readily
in affirmative contexts than in negative ones. I would, however, still argue that
there are additional reasons to be noted in relation to this contrast between af-
firmative doubt and negative doubt from the perspective of the history of Eng-
lish. First of all, doubt used in the form of negation was, in my view, too busy
with another type of competition during the critical period, that is, the com-
petition between I doubt not and I do not doubt. It was perhaps difficult for
fluctuating forms to become an epistemic parenthetical and to be used almost
as a sentential adverb, since the process of the development of the parenthetical
use was also a process of becoming more and more fixed or frozen in form.
Unfortunately, the early Modern English period was an important period in
terms of the establishment of the auxiliary do as well. The auxiliary do per-
colates into the English verbal system, resulting in its fairly dramatic expansion
by around 1700 (Ellegård 1953: 162), although inconsistencies in respect of the
occurrences of do are still observed in the eighteenth century (see Tieken 1987,
1989, and Iyeiri 2004). Interestingly enough, doubt is one of the verbs which
were relatively slow in adopting the use of do (Ellegård 1953: 199, Visser
1963–1973: §1534, and Iyeiri 2004: 227–229). The relatively frequent oc-
curring of I doubt not even in the eighteenth century indicates that its use was
in a way idiomatic, if not epistemic, but it was in the end superseded by I do
154 Chapter 5

not doubt. Then, there was also a choice between I do not doubt and I don’t
doubt. The condition like this is certainly disadvantageous when I doubt not or
I do not doubt tries to be more parenthetical in use.
The second reason related to the hesitancy of I doubt not or I do not doubt
to become parenthetical, I surmise, is that no doubt, which arises from doubt
used as a noun, was increasingly established as a fixed phrase in the history of
English. Since no doubt often functioned as a sentential adverb, it was not nec-
essary for the verb doubt to function at the sentential level. See Figure 31 be-
low, which depicts the rise of the proportions of the combination of no and
doubt to the entire sample of doubt used as a noun in the OED quotations.112
The figure also shows the proportions of no doubt without the conjunction that,
which is in many cases, though not always, adverbial in function:

100

80

60

40

20

0
13th c. 14th c. 15th c. 16th c. 17th c. 18th c. 19th c. 20th c.

no doubt (all) no doubt (without 'that')

Figure 31. The proportions of no doubt to the entire sample of doubt used as a noun in the
OED (%)

The following examples illustrate the adverbial use of no doubt, which func-
tions at the sentential level:

No doubt the name and reputation thereof would have bin a spurre to these erec-
tions, as nurses for babes to suck in, till they might repaire thither sc. to the Uni-
versity to be wained.
(1571 Edmund Campion, A History of Ireland)

112
The graph simply displays the proportions of the combination of no and doubt, and there-
fore includes examples where no doubt is not parenthetical at all. It simply displays how
the connection between no and doubt was strengthened in the historical development of
this noun.
Verbs of implicit negation and subordinate clauses 155

The gentleman no doubt will fall to his jewlips.


(a1625 John Fletcher, The Humorous Lieutenant)

Dr. Shaw no doubt tapped the matter to the people.


(1768 Horace Walpole, Historic Doubts on the Life and Reign of King Richard
the Third)

Figure 31 displays the steady rise of no doubt, which is in fact due to the steady
rise of no doubt without the conjunction that. It is perhaps reasonable to sur-
mise from this graph that it was unnecessary for the verb doubt in negation to
find its way to function at the sentential level, since the same function was
often and perhaps more easily fulfilled by the noun doubt.
Finally, a brief comment can also be made upon the existence of doubtless,
as in:

They doubtles will resolue them that it is true which is here expressed.
(1590 Edward Webbe, The Rare and most Wonderfull Things Which E. Webbe
hath Seene in his Travailes)

The Lord of all things … doubtles will command the people to make good his
promises of maintenance more honorably unask’d, unrak’d for.
(1659 John Milton, Considerations Touching the Likeliest Means to Remove Hire-
lings out of the Church)

Doubtless many will gladly take up a work so vital to the welfare of the whole
community.
(1893 Arthur Cawston, A Comprehensive Scheme for Street Improvements in Lon-
don)

Doubtless it is easily confused with the pseudo-morality of sanctions.


(1943 Mind)

Though limited in frequency, doubtless when it occurs often functions at the


sentential level as the above examples illustrate, and it is inclined to occur in
medial and final positions. In other words, the use of doubtless is also an al-
ternative option. This is perhaps an additional factor that made it unnecessary
for I doubt not or I do not doubt to function at the sentential level, although as a
factor it was probably much weaker than no doubt, which was much more
frequent as far as the quotations in the OED are concerned. Further research is
necessary, however, to confirm this point, since the OED is a compiled text and
156 Chapter 5

not a direct extract from naturally-used language, and therefore not suitable for
frequency analyses of this kind.113
As discussed so far, the historical development of doubt is similar to that of
fear, in that in both cases the expanded use of the parenthetical use is typically
encountered in affirmative sentences. It is therefore important to make a sepa-
rate treatment between the verbs used in the affirmative and the verbs used in
the negative. Unlike the case of the verb fear, however, the development of to-
infinitives and gerunds is not really a feature of doubt even in negative contexts.
Negative doubt utilised but-clauses more extensively than fear, and survived
the period when that-clauses were unstable, and later on returned to the re-
established use of that-clauses. Furthermore, lest-clauses tend to be com-
plementary to but-clauses in distribution, in that they are attested with doubt in
affirmative sentences. This is a shared feature between fear and doubt.

5.3.9. The verb question

Before turning to the next verb deny, the present section makes an additional
account, namely an account on the verb question, which is occasionally used in
the meaning ‘doubt’ as the OED states (s.v. question, 4a, b). The usage of ques-
tion is in some ways similar to that of doubt, although the verbal use of ques-
tion itself is very restricted in frequency and does not provide numerous ex-
amples in the OED quotations data.
To be more specific about the similarity between the two verbs, the use of
whether-clauses and if-clauses are much more frequent when question is used
in affirmative sentences than in negative ones, which is comparable to the con-
trast between affirmative and negative sentences dominated by doubt. Af-
firmative question provides 67 relevant examples in total (from the fourteenth
century to the twentieth century), of which as many as 45 are followed by
whether-clauses or if-clauses. See the following examples, which illustrate the
use of whether and if:

I question whether the strongest north-wester would dissipate it.


(1751 Benjamin Franklin, Works)

I question if it be wise in running a railroad to water anything but the engine.


(1870 Maximilian Schele de Vere, Americanisms; the English of the New World)

113
See also Notes 52 and 58 in Chapter 3 on this issue.
Verbs of implicit negation and subordinate clauses 157

By contrast, question in negative environments provides seventeen examples in


all, of which only three are followed by whether-clauses and if-clauses, as in:

Nor did they question whether … false the Prophet were, that brought th’Embas-
sage.
(1621 Francis Quarles, Hadassa, or the History of Queene Esther, with Medita
thereupon)

No man can question whether wounds and sickness are not really painful.
(1758 Samuel Johnson, The Idler, No. 4)

We should not question if that he should live. (1594 First Pt. Contention)

Alternatively, the type of subordinate clauses most typically found with ques-
tion in negative environments is the one introduced by but. Some illustrative
examples are:

I question not but the Quakers … would play the part of the Puppet or
Punchinello in the Antelude of the Pageant.
(1668 Henry More, Divine Dialogue)

He did not question but the native Irish would join him.
(1878 Richard Simpson, The School of Shakspere)

As in the case of doubt, examples of but-clauses are attested with the verb
question in negative environments most characteristically during the early
Modern English period or the critical period in respect of the decline of that-
clauses in general. In the OED quotations, the seventeenth century, for example,
yields seven relevant examples, of which four are of this type. Apparently, the
use of but-clauses becomes less and less frequent with question after this peak
in the history of English. Regarding Present-day English, Herbst, Heath, Roe,
and Götz (2004: 641–642) do not list the construction with but under the entry
of the verb question, although this does not guarantee of course the non-
existence of but-clauses today. The OED does not regard this usage as obsolete
(s.v. question).
Also, there are a few examples of non-finite complement forms, i.e. to-
infinitives and gerunds, following the verb question in the seventeenth and
eighteenth centuries, which is again a shared feature between question and
doubt. As in the case of doubt, these constructions were never established with
158 Chapter 5

question, either, in the history of English. Some illustrative examples of infini-


tives and gerunds follow:

I don’t question but to action him out on’t.


(1733 Henry Fielding, Don Quixote in England)

As for the Acid Saline Principle, I suppose no person who hath tasted the Spirit of
Vitriol, … will question its abounding in that subject.
(1674 Philosophical Transactions of the Royal Society)

The first of the above is an example of the to-infinitive accompanied by but. As


mentioned above, the same phenomenon is evidenced with doubt as well.114
Despite the existence of these similarities, however, the verb question dif-
fers from doubt in some other respects. First of all, the verb question is much
more frequently followed by complements of the “open type” than doubt, as
exemplified by the following sentences:

I … was so bould as to question what they were, and of their businesse.


(1592 Robert Greene, A Quip for an Upstart Courtier)

They never questioned what crime he had done.


(1651 Thomas Hobbes, Leviathan, or the Matter, Forme, and Power of a Com-
monwealth, Ecclesiasticall and Civill)

This is particularly noticeable with question used in affirmative contexts, and


that in some earlier data, although the second of the above examples illustrates
its occurrence in negative environments. The fifteenth and sixteenth centuries
provide nine examples of the verb question used in affirmative sentences in the
OED, all of which are followed by an open type of complements. In this sense,
question is not so much a verb of negative implication as doubt, since it is more
open as to the judgement of whether the content matter of the complement is
negative or not.
Finally, the employment of that-clauses is very infrequent with the verb
question, and as a consequence, the development of the parenthetical use of I

114
The OED quotations provide only four examples of to-infinitives used with the verb ques-
tion, two of which include but in the form “question + but + to-infinitive” as illustrated by
this example. The other two are devoid of but, but both include an object before the to-
infinitive and display the “question + object + to-infinitive” construction. As for the ger-
und, the example cited here is the sole relevant one found in the OED.
Verbs of implicit negation and subordinate clauses 159

question is not really observed in its historical development, at least as far as


the OED dataset is concerned. In other words, sources for the development of
the parenthetical use of I question were not sufficiently numerous from the
beginning. This is in fact a marked difference between question and doubt, in
view of the fact that the rise of the parenthetical use of I doubt is one of the
major features of the historical development of doubt.

5.4. The historical development of deny


5.4.1. Preliminary remarks

The last verb to be investigated under the category of “verbs of implicit nega-
tion and subordinate clauses” is deny. It is most commonly followed by that-
clauses in Present-day English as the examples below, which are quoted from
the BNC, illustrate:

The frontman for the rave band has denied that the song, Ebeneezer Goode, is
another name for ecstasy and that the lyrics glamorise the drug.
(The Daily Mirror, from BNC)

What he is mainly denying when he denies that there are final causes in nature is
that the existence of each individual sort of thing is to be explained by its serving
some cause beyond it, in particular some kind of human interest.
(Timothy L. S. Sprigge, The Rational Foundations of Ethics, from BNC)

The construction with that-clauses as shown above is marked as “very fre-


quent” by Herbst, Heath, Roe, and Götz (2004: 220–221). It is perhaps safe to
conclude that this is the most common type attested with deny in contemporary
English.
Apart from this type of complements, gerunds are also found with this verb
as the following quotations exemplify. Although Herbst, Heath, Roe, and Götz
(2004: 220–221) provide the marking of “frequent” for this construction, it
seems to be much more restricted in frequency with deny in contemporary Eng-
lish, at least in comparison to that-clauses. The following are examples il-
lustrating the use of gerunds after deny, both cited from the BNC:

Mahfouz, who is in his forties, denied assaulting the girls.


(The Daily Mirror, from BNC)
160 Chapter 5

The influence of the tabloid press was particularly strong on those voters who
denied being party “supporters”, even when they had a party preference.
(William L. Miller, Media and Voters, from BNC)

In contrast to these two types of constructions, the employment of to-


infinitives seems to be marginal with deny in Present-day English. It is ad-
mitted in dictionaries if the verb is followed by the object in intervening posi-
tion (cf. OED, s.v. deny), but relevant examples are unavailable in the twentieth
century data of the OED. The following examples of this type, cited from the
OED, go back to the nineteenth century:

I beg leave to deny this to be law.


(1818 William Cruise, A Digest of the Laws of England)

Many authors deny the conjunction to be an aspect, because the stars do not be-
hold each other, but their influence is on the Earth, which they behold with a con-
junct aspect. (1819 James Wilson, A Complete Dictionary of Astrology)

There are nine examples of this kind in the nineteenth-century data of the OED,
which account for about 10% of the examples of deny followed by the com-
plements of that-clauses, infinitives, and gerunds in the century. This was,
therefore, not an uncommon construction with the same verb, at least up to the
nineteenth century, but it does not seem to be frequent any longer in Present-
day English. Egan (2008: 341) does not record this construction in his explora-
tion of the BNC.
Furthermore, examples like the one cited below where deny is directly
followed by a simple to-infinitive (i.e. without the intervening object) are now
considered to be dialectal (cf. OED, Shorter Oxford English Dictionary on His-
torical Principles, s.v. deny):115

They acknowledged him to be the true God, whome before they denyed to know.
(1611 Bible, Wisdom)

Again, this construction is not recorded by Egan (2008: 341), who analyses the
data of the BNC. To turn to the OED quotations, examples of to-infinitives

115
The Shorter Oxford English Dictionary remarks that this usage is obsolete except in
Scotland. The example here is cited from the OED, while all the other examples from the
Bible in English in the present study are cited from the Bible in English on CD-ROM. See
also 2.1. in Chapter 2.
Verbs of implicit negation and subordinate clauses 161

without the intervening object are available only up to the eighteenth century,
as in:

The faithful hound … Takes what the glutted child denies to eat.
(1718 Matthew Prior, Solomon)

Deny here is used in the meaning ‘to refuse’, which is obsolete today as I
discuss under 5.4.2. below.
All in all, the types of complements usually encountered with the verb
deny are that-clauses and gerunds in Present-day English. The Cambridge Ad-
vanced Learner’s Dictionary (s.v. deny) provides the constructions with that
and gerunds only. The same is also true of Macmillan English Dictionary for
Advanced Learners (s.v. deny).
Despite the usual existence of gerunds as well as that-clauses used as com-
plements of deny, I have decided to discuss the historical development of this
verb in the present chapter, since that-clauses are far more frequent in use than
gerunds at least as far as Present-day English is concerned. As Table 13 demon-
strates, both in LOB and FLOB the verb deny presents a marked tendency to be
followed by that-clauses instead of gerunds. The table presents the raw fre-
quencies of the relevant examples in the two corpora:

Table 13. Deny followed by that-clauses and gerunds in LOB and FLOB (raw frequen-
cies)
that-clauses gerunds
LOB 15 3
FLOB 14 1

As a matter of fact, this table seems to indicate an increasing tendency for deny
to employ that-clauses in the latter half of the twentieth century, since the
relative proportion of gerunds to that-clauses is even more restricted in FLOB,
which is about 30 years later than LOB, although relevant examples are not
abundant enough to immediately corroborate this hypothesis. In any case, the
frequent attestation of that-clauses in Present-day English hints at the pos-
sibility that deny escaped from the major complement shifts leading to a sig-
nificant rise or the establishment of to-infinitives and gerunds in its historical
development, which have affected various verbs of implicit negation in English
as thus far explored in the preceding chapters. The discussions in the following
sections are in the main concerned with this issue.
162 Chapter 5

5.4.2. The complement shift from that-clauses to to-infinitives

As the section above explicates, the verb deny usually provides that-clauses or
gerunds in Present-day English. To turn to the historical data of the OED, how-
ever, to-infinitives are not at all rare. As a matter of fact, examples of to-
infinitives are more copiously found than those of gerunds in the historical
context of English. The dataset under investigation provides a total of 427 in-
stances of deny followed by a complement of some sort, of which as many as
108 illustrate the use of to-infinitives. Some illustrative examples where to-
infinitives are employed follow:

Whiche thynges if he deny to dooe, then the confederates certifie hym, that thei
shall neuer cease till he be brought to reason.
(a1548 Edward Hall, Chronicle, Henry VIII)

If they denie to come, Swinge me them soundly forth vnto their husbands.
(1596 William Shakespeare, The Taming of the Shrew)

Quite clearly, deny in these examples are used in the meaning ‘to refuse’.
It is relevant in this connection that Rudanko (1998: 90) provides the fol-
lowing example from Defoe, where deny is followed by the to-infinitive, and
says that the meaning of the verb is ‘to refuse’:

From the River they travelled towards the Forest, but when they came to
Walthamstow the People of that Town denied to admit them, as was the Case
every where (Daniel Defoe, A Journal of the Plague Year, from Rudanko)

Rudanko refers to the OED (s.v. deny), and comments that this meaning is
obsolete today. It is, therefore, a reasonable conjecture that the decline of this
meaning and the decline of this construction are mutually related. The former
may have affected the latter, or vice versa.
As far as the syntactic development of the verb deny is concerned, the
early Modern English period sees a sudden increase in the use of to-infinitives
as Figure 32 shows. Of the 108 examples of to-infinitives in the dataset of the
OED, more than 70% are attested in the sixteenth and seventeenth centuries.
See the following graph which elucidates this point:
Verbs of implicit negation and subordinate clauses 163

100

80

60

40

20

0
14th c. 15th c. 16th c. 17th c. 18th c. 19th c. 20th c.

that-clauses but-clauses to-infinitives gerunds

Figure 32. Deny and four patterns of complementation in the OED (raw frequencies)

Like most other graphs in the present study, Figure 32 displays the raw fre-
quencies of different patterns available with this verb, and therefore it is
necessary that comparison and contrast should be made only within each cen-
tury. Still, the expanded use of the to-infinitive by deny during the early Mod-
ern English period is quite obviously witnessed in this graph. To-infinitives are
relatively frequent up to the eighteenth century and experience a significant
decline thereafter in this graph.
Despite the inference in the previous section, which is based upon the
contemporary data of deny, the frequencies of that-clauses, but-clauses, to-
infinitives, and gerunds as displayed in Figure 32 suggest that the first
complement shift or the decline of that-clauses most plausibly took place with
this verb as well. Although the Middle English period suffers from the dearth
of relevant examples in the OED quotations, one can still observe in the above
graph a relatively well-established use of to-infinitives around the transitional
period from Middle English to the beginning of the early Modern English
period. One could surmise from this state of affairs that this is due to the oc-
currence of the first complement shift, whereby the decline of that-clauses
often led to the rise of to-infinitives.
Unlike various other verbs of implicit negation, however, the shift was not
entirely successful with the verb deny. In usual cases, the first complement
shift, when it occurs, is so penetrating that the use of that-clauses quickly
recedes, leaving abundant examples of alternative constructions, which are in
many cases to-infinitives. In the case of deny, however, that-clauses do not
recede to the extent of near extinction in the end. On the contrary, they simply
kept lingering during the whole period of early Modern English, which most
likely allowed their return during the later Modern English period. From the
164 Chapter 5

nineteenth century onwards, the use of that-clauses is firmly established, and as


shown in Figure 32, it may be becoming even firmer these days. Although the
returning of that-clauses is a feature observed with some other verbs as
well—e.g. fear (see 5.2.6.) and doubt (see 5.3.7.)—, the same phenomenon is
so outstanding with the verb deny that it merits particular attention. On the
other hand, to-infinitives are increasingly marginal with deny in later Modern
English.116 As a result, that-clauses are the dominant form of complementation
of this verb in current English, and gerunds are only on occasions encountered.
One possible factor related to the historical development of deny as
described above is the fact that the range of its meanings of this verb was so
wide as to cover past and present events as well as future possibilities. When
the future was referred to or future implications were available, to-infinitives
were most presumably useful, as in fact shown by the historical data. On the
other hand, it is likely that that-clauses (and perhaps but-clauses) were of
particular convenience when the matters referred to were in the preterite, as
illustrated by the following:

Others … summarily deny, that ever this Kirk had any approved discipline.
(1621 The First and Second Book of Discipline of the Church of Scotland)

Dare you deny that these were my words? If you do you are a falsifier.
(1675 Richard Baxter, Catholick Theologie)

Examples where the complement clause is in the preterite tense are numerous
in the dataset under analysis. As a matter of fact, about one third of the finite
complement clauses are of this kind in the OED quotations. This is one of the
essential differences between deny and some other verbs of prohibition like
forbid, which are fundamentally related to future events or future possibilities
and which underwent the first complement shift quite smoothly and suc-
cessfully.
For the purpose of indicating past events, examples like the following are
also available in the OED quotations, where to-infinitives and gerunds occur in
the perfect form to present the same kind of tone or effect:

116
As mentioned in the main body of discussion, the decline of to-infinitives is in part linked
to the decline of the meaning ‘refuse’, since it is often the meaning of deny when it is
followed by to-infinitives in earlier data. At the same time, however, it was not the only
meaning of this verb followed by to-infinitives as I discuss later in this section.
Verbs of implicit negation and subordinate clauses 165

He is unkynde whiche denieth to haue receyued any benefite that in dede he hath
receyued. (1531 Sir Thomas Elyot, The Boke Named the Gouernour)

Stone at first peremptorily denied having seen that book.


(1752 Horace Walpole, Letters)

After all, however, examples like these are not at all common throughout the
history of English. I have found only six relevant examples of to-infinitives and
three of gerunds of this kind in the entire dataset of the OED quotations.
Obviously, the expected disadvantage in retaining finite clauses in early
Modern English is that the verb at issue had eventually to deal with the matter
of the decline of expletive negation, which was attested with deny as well in
earlier English, as in:

He denyet … þat noqwere he knew þat commly be keppet.


(c1400 Destruction of Troy)

As repeatedly mentioned throughout the present study, negation in the com-


plement clause as in this example is not necessary from the semantic perspec-
tive, but often occurs with verbs of negative connotation in Old and Middle
English. Van der Wurff (1998: 297) concludes that expletive negation of this
kind undergoes a significant decline after the late Middle English period per-
haps due to the decline of ne, which is often the form of expletive negation. As
discussed above, however, negative items other than ne are also encountered in
complement clauses of verbs of implicit negation, and indeed, the above ex-
ample illustrates the use of nowhere in the complement clause, which also
disappeared in a gradual manner in the history of English. The employment by
various verbs, of non-finite forms like to-infinitives or gerunds in place of
that-clauses in early Modern English was timely in this respect, since it could
automatically sort out the problem of whether to delete or retain expletive
negation in that-clauses. To-infinitives accompanied by expletive negation also
occur as the example below illustrates, but in effect they are rare. Expletive
negation is essentially a phenomenon encountered in finite subordinate clauses,
and that in earlier English.

Taxing the poore king of treason, who denied to the death not to know of any such
matter.
(1624 Captain John Smith, The Generall Historie of Virginia, New-England and
the Summer Isles)
166 Chapter 5

The verb deny, however, retains that-clauses to a significant extent as dis-


cussed above, and this is most presumably why the employment of but-clauses
became rather extensive with this verb during the transitional period from late
Middle English to early Modern English, when the obliteration of expletive
negation was in progress. Figure 32 above (p. 163) shows that the employment
of but-clauses became suddenly common in the sixteenth century, and it
became even more prominent in the seventeenth century with this verb. As
repeated in various places of the present study, it is widely acknowledged that
that-clauses with expletive negation were often replaced by but-clauses in the
history of English. Examples of but-clauses dominated by deny include:

First of all I suppose no man will deny, but that Paule doth denounce men to be
Justified by fayth. (1581 James Bell, Walter Haddon against Osorius)

And … wee cannot deny, but that there was some apparency.
(1604 Edward Grimston, tr., Acosta’s Naturall and Morall Historie of the East
and West Indies)

But-clauses are evidenced with a number of verbs of implicit negation, at least


to a certain extent, during the early Modern English period, but as discussed
above, they are particularly frequent with verbs which retain the use of that-
clauses in the end in the history of English. The verb doubt investigated above
(see 5.3.) is one of such verbs, and deny is certainly another. Here again, the
occurrence of but-clauses is largely confined to negative sentences, but is also
evidenced in interrogative sentences which expect a negative response.
Finally, examples of lest-clauses are unavailable with deny in spite of the
fact that about three fourths of the relevant examples of deny are found in
affirmative sentences. This is perhaps due to the meaning of deny, which is dif-
ferent from fear and doubt and which does not really get along well with the
purposive meaning of lest. See the above discussion for the argument that it
often refers to past events, although reference to future possibilities is not ex-
cluded with this verb.

5.4.3. The parenthetical use and the ellipsis of the conjunction that

Deny is similar to fear and doubt, both discussed in the present chapter, in that
it retained the use of finite clauses to a significant extent in the Modern English
period. On the other hand, the difference between deny and the verbs fear and
doubt is that the former sees some expansion of to-infinitives at a very early
Verbs of implicit negation and subordinate clauses 167

period of Modern English, while the latter group of verbs, especially in af-
firmative sentences, quickly turned to the development of epistemic paren-
theticals as illustrated below, instead of experiencing a notable development of
to-infinitives, when faced with the weakening of that-clauses, which is most
probably due to the decline of the subjunctive:

For farder Credit off your Worde, you will stande (I feare) for banckeroute.
(1566 Thomas Stapleton, A Return of Untruthes vpon M. Jewelles)

This, I doubt, will prove an Utopian conceit.


(1659 Brian Walton, The Considerator Considered)

In these examples, I fear and I doubt are used as a set, which functions almost
like an epistemic adverb to qualify the entirety of the sentence. In other words,
I fear and I doubt in them have shifted from the matrix verb to an epistemic
parenthetical in status, and the original subordinate clause now functions as the
matrix one.
As discussed above under the sections of fear and doubt (see 5.2. and 5.3.
above), it is not always easy to tell about each example whether the original
subordinate clause has indeed become the matrix clause, letting I fear and I
doubt function adverbially at the sentential level, since a large number of ex-
amples are of the following kind, where I fear and I doubt are not dislocated
but where the conjunction that is simply unexpressed:

I feare these stubborn lines lack power to moue.


(1588 William Shakespeare, Loues Labour’s Lost)

In this example, I feare may still retain the original matrix clause status, while
at the same time, it may be in the process of becoming an epistemic paren-
thetical which functions at the level of the entire sentence. It is simply im-
possible to make a clear distinction of this kind.
As discussed above in relation to fear and doubt, however, one could
perhaps trace the development of the parenthetical use by observing the in-
creasing deletion of the conjunction that, which is supposed to have taken
place in the course of the Modern English period. Like fear and doubt, the verb
deny indeed shows a slight increase of the ellipsis of the conjunction that in the
past. In other words, the development of the epistemic use was a possible path
for it as well. After all, however, examples with the conjunction that un-
expressed are by no means frequent throughout the history of this verb, as far
as the OED dataset of quotations is concerned. See Figure 33 below, which
168 Chapter 5

displays the raw frequencies of “that expressed” and “that unexpressed” after
the verb deny:

90
75
60
45
30
15
0
14th c. 15th c. 16th c. 17th c. 18th c. 19th c. 20th c.

that expressed that unexpressed

Figure 33. Deny and clauses where the conjunction that is expressed or unexpressed in the
OED (raw frequencies)

This graph indeed indicates that the same development was probably underway
during the early Modern English period with deny as well. A temporary
increase of the elliptical use of the conjunction that is indeed evidenced with
deny in the seventeenth century. However, in contrast to the cases of fear and
doubt, with both of which “that unexpressed” becomes far more frequent than
“that expressed” in an increasing manner, the rise of the elliptical use of the
conjunction that in the seventeenth century in this graph is only temporary and
the process does not accelerate into further stages in the case of deny. In other
words, the parenthetical use was not established in the historical development
of this verb in the end. This is a significant difference between deny and the
group of fear and doubt in the history of English.
As a matter of fact, the conjunction that is often retained with deny even
when the verb is dislocated from the original position, as illustrated by the fol-
lowing quotations:

That I have in my Army some of the Romish Communion, I do not deny.


(1642–1643 Earl of Newcastle, Declar. in John Rushworth, Historical Collec-
tions of Private Passage of State, Weighty Matters in Law, Remarkable Proceed-
ings in Five Parliaments)
Verbs of implicit negation and subordinate clauses 169

That this substantiates the charge of cruelty against us I altogether deny.


(1815 William Kirby and William Spence, An Introduction to Entomology)

That I am ungrateful I wholly deny.


(1875 Benjamin Jowett, The Dialogues of Plato)

The retention of the conjunction that in these examples confirms the status of
deny as the matrix verb. In other words, the path to become an epistemic
parenthetical was not so readily available to deny as it was to fear and doubt,
although I also admit that the environments where deny occurs in the above
examples are not really favourable for the development of epistemic paren-
theticals for other reasons in any case. Deny in the first of the above examples
occurs in negation, which is not favourable for the development of the
parenthetical use. Moreover, the same verb in the second and the third ex-
amples occurs with intervening adverbials, which is also unfavourable for the
development under consideration. I deny in these examples is far from being
fixed in any way. All in all, the slight rise of the elliptical use of that in the
seventeenth century alone is not convincing enough to confirm the establish-
ment of the path. It is, therefore, safe to argue that the path to the development
of the parenthetical use was not necessarily prepared for this verb.
To make this point even clearer, one could investigate how infrequently the
conjunction that is unexpressed when deny appears in the affirmative sentence,
since the parenthetical use is often a typical feature of the affirmative context.
Figure 34, which is based upon the OED dataset, graphically presents the raw
frequencies of the conjunction that expressed and unexpressed in the case of
affirmative deny only:

60

45

30

15

0
14th c. 15th c. 16th c. 17th c. 18th c. 19th c. 20th c.

that expressed that unexpressed

Figure 34. Deny in affirmative sentences and clauses where the conjunction that is ex-
pressed or unexpressed in the OED (raw frequencies)
170 Chapter 5

Here again, there is a slight rise of the elliptical use of that in the seventeenth
century, but on the whole, the conjunction that is well preserved throughout the
history of English with this verb. In other words, the overall tendency of Figure
34 does not differ much from that of Figure 33 above, which shows the overall
tendency of the verb deny including its examples used in negative environ-
ments. This fact itself indicates that deny did not really follow the path leading
to the parenthetical use, since the development of it is often observable in the
form of an increasing and significant gap or separation between the item con-
cerned in the affirmative and the same item in the negative, as discussed under
the sections of fear and doubt above. In the case of fear and deny, it was even
necessary to conduct a separate analysis between affirmative and negative
sentences. This is simply not the case with deny.
Additionally, one could also note that the conditions were not satisfied for
deny to become an epistemic parenthetical from the period of its early occur-
rences. See the following graph, which displays how infrequently deny occurs
with first person subjects in the entire data. Here I present normalised figures
by calculating the proportions of first person subjects to the whole set of rele-
vant examples:

100

80

60

40

20

0
14th c. 15th c. 16th c. 17th c. 18th c. 19th c. 20th c.

1st person

Figure 35. Deny with first person subjects in the OED (%)

As this figure demonstrates, the proportions of first person subjects count only
about 20% of the entire relevant data of deny throughout the history of Eng-
lish.117 This is a marked contrast with the situations of fear and doubt, which
most frequently occur in the forms of I fear and I doubt, namely with the first

117
Since relevant examples are sparse in the fourteenth and fifteenth centuries in the OED
quotations, the statistics for the period are rather unstable.
Verbs of implicit negation and subordinate clauses 171

person, and which have relatively smoothly developed the parenthetical use in
their historical development. As a matter of fact, the proportion of first person
subjects seems to decline as time passes in the case of deny in Figure 35, and
this fact itself also testifies that the development of the parenthetical use is not
a feature of the history of this verb. Remember the cases of fear and doubt,
where the subject of the verbs was increasingly in the first person in the history
of English.
This tendency stays the same even if one limits the data to deny used in
affirmative sentences. See Figure 36, which depicts the corresponding propor-
tions of deny, i.e. the same verb used with first person subjects, in affirmative
sentences only:

100

80

60

40

20

0
14th c. 15th c. 16th c. 17th c. 18th c. 19th c. 20th c.

1st person

Figure 36. Deny with first person subjects in affirmative sentences in the OED (%)

Despite the fact that the examples of deny in the affirmative only are invest-
tigated, Figure 36 again shows how infrequent the occurrence of first person
subjects is and how it undergoes an even gradual decline in the course of the
later Modern English period. As a matter of fact, the subject of deny has a
strong tendency to be in the second or third persons, as in:

You deny not that God knoweth from eternity whether the condition of each Event
will it self be or not. (1675 Richard Baxter, Catholic Theologie)

The Hermeticall Philosophers deny that there is a quintessence, because there are
not fower elements. (1605 Thomas Timme, Quersitianus)

In this manner, the circumstances where deny occurred in the historical


context were not entirely favourable from the beginning for the development of
172 Chapter 5

the parenthetical use. Presumably, the fact that the parenthetical use did not
really develop with the same verb left the unfavourable circumstances as they
were or perhaps made them even less favourable in later history, which is most
characteristically represented in the decline of the proportions of first person
subjects in recent years with this verb.
As hitherto discussed, the first complement shift or the weakening of finite
that-clauses affected the complementation patterns of deny to some extent from
later Middle English to early Modern English, but the shift itself was not at all
successful. Although the proportion of to-infinitives to that-clauses grew to
some extent in the early Modern English period, this tendency was not strong
enough to obviate that-clauses in the end. The deletion or the ellipsis of the
conjunction that is also observed with deny in some measure during the early
Modern English period—especially in the seventeenth century—, but this trend
was not strong enough to lead to the development of the parenthetical use of
deny. Unlike the cases of fear and doubt, linguistic circumstances of deny were
not really favourable for the development of the parenthetical use. After all, the
use of that-clauses returns in the case of this verb—this time in a firmer form—,
once the effect of the first complement shift was less and less obvious. The
whole story is most probably related to the meaning of deny, which was
unfavourable for the establishment of other complement types than that-clauses.
This is a major difference between deny and other verbs of implicit negation
investigated in the present study.

5.4.4. The rise of gerunds with deny

The discussion of the verb at issue has so far been centred upon the decline and
the later re-established use of that-clauses in the entire history of the English
language and the temporary rise of to-infinitives during the early Modern Eng-
lish period. As mentioned in the preliminary remarks of this verb (cf. 5.4.1.),
however, Present-day English offers gerunds as well as that-clauses with deny,
although the use of the former is much more restricted in frequency than the
latter. The present section explores the rise of gerunds in the historical de-
velopment of deny.
The usual source for the development of the gerund is the second
complement shift, which is characterised by the decline of to-infinitives and
which is often observed slightly later than the first shift or the decline of that-
clauses, perhaps sometime in later Modern English. Although the first
complement shift was not very successful in the case of deny as argued above,
Verbs of implicit negation and subordinate clauses 173

it is true that it produced a notable amount of to-infinitives with this verb, too,
during the early Modern English period. At least, the sixteenth and seventeenth
centuries evidence almost as many examples of to-infinitives as that-clauses,
both following the verb deny (see Figure 32 on p. 163). It is, therefore, not a
matter of surprise if they form the basis for the later development of gerundial
constructions in the history of English. Indeed, examples of gerunds dominated
by deny are available in contemporary English, as mentioned under 5.4.1.
above and as the following examples illustrate:

The Lt.-Governor had denied having stated that anti-social elements or goondas
enjoyed the patronage of politicians. (1969 Times of India, 30 July)

He pleaded Guilty to political crimes and diversionist activity, but denied col-
laborating with the Gestapo during the war. (1955 Times, 1 July)

On the other hand, the complement pattern with to-infinitives experiences a


gradual decrease from the eighteenth century to the twentieth century, and dic-
tionaries of contemporary English tend not to refer any longer to the to-
infinitival usage under the entry of deny. Accordingly, it is perhaps reasonable
to postulate that the complement shift from to-infinitives to gerunds occurred
with this verb, at least in a humble measure, although this is not at all out-
standing as far as Figure 32 (p. 163) is concerned. It is in a way hidden behind
the abundance of that-clauses occurring in sequence with the verb deny
throughout the history of English. Another unfortunate aspect of the second
complement shift with this verb is that it is based upon the unsuccessful shift to
to-infinitives (the first shift). This again makes the rise of gerunds rather
difficult to observe.
One interesting point perhaps related to the second complement shift is the
fact that usages of to-infinitives were increasingly less versatile in this process.
As mentioned in the preliminary remarks of this verb, simple to-infinitives as
illustrated by the example below do not follow deny in Present-day English any
longer, whereas examples of this type were available in the past:

Whiche thynges if he deny to dooe, then the confederates certifie hym, that thei
shall neuer cease till he be brought to reason.
(a1548 Edward Hall, Chronicle, Henry VIII)

In this example, deny is directly followed by to do without the intervention of


the semantic subject of the to-infinitive. Deny to do here means ‘to refuse to
do’. This usage is considered to be obsolete except in dialects in Present-day
174 Chapter 5

English, and simultaneously the meaning of ‘refuse’ is also obsolete. By con-


trast, the “deny + object + to-infinitive” construction as in the quotation below
is acknowledged in the OED, although examples of this type are limited to the
period up to the nineteenth century, as far as its citations are concerned:

He positively denies himself to be either the inventor or revivor of it.


(1841 Samuel Fenton, The Excellent Properties of Salted Brandy as a Most Ef-
ficacious Medicine and Sedative for Internal as Well as External Diseases, In-
flammation, and Local Injuries)

I would find it intriguing that the decline of the simple infinitival con-
struction (i.e. “deny + to-infinitive”) seems to be related to, and consequently
coincides with the possible complement shift from to-infinitives to gerunds in
the history of English. See Figure 37 below, which reveals the raw frequencies
of the to-infinitive with or without the intervening object after deny in the OED
quotations:

25

20

15

10

0
14th c. 15th c. 16th c. 17th c. 18th c. 19th c. 20th c.

simple infinitives object + infinitives

Figure 37. Deny and two types of infinitival constructions in the OED (raw frequencies)

As far as this graph is concerned, the verb deny directly followed by the to-
infinitive was not rare at all up to the early Modern English period, whereas the
usage becomes increasingly marginal in the course of the later Modern English
period and eventually disappears, at least in standard English. I would surmise
that this is one aspect of the shift of complements triggered by the weakening
of to-infinitives, since it shows the shrinking of its usage in a way. In other
words, the complement shift from to-infinitives to gerunds, though not
prominent against the marked expansion of that-clauses, indeed took place
with the verb under investigation. It took place, first in the form of the loss of a
Verbs of implicit negation and subordinate clauses 175

particular construction with to-infinitives, and later in the form of the real
decline of to-infinitives of all types. In this process, the meaning ‘to refuse’
was also lost with this verb. The loss of the meaning is certainly linked to the
decline of a particular construction, though it is not easy to tell which caused
which.
Furthermore, it is interesting that the decline of the simple infinitive con-
struction without the intervening object (i.e. “deny + to-infinitive”) as depicted
in Figure 37 coincides with the rise of simple gerund constructions or the
gerund without its expressed semantic subject in historical terms. See Figure 38,
which displays the raw frequencies of gerunds with or without their semantic
subjects after the verb deny in the OED quotations:

7
6
5
4
3
2
1
0
14th c. 15th c. 16th c. 17th c. 18th c. 19th c. 20th c.

simple gerunds semantic subject + gerunds

Figure 38. Deny and two types of gerundial constructions in the OED (raw frequencies)

Although relevant examples are not altogether abundant in the OED quotations
data, Figure 38 hints at the possibility that the shrinking of the infinitival usage
was accompanied by the expansion of the gerundial usage with deny. The
complement shift from to-infinitives to gerunds is, if it ever occurred, evi-
denced with deny in this humble form, although the increase of gerunds them-
selves is not particularly prominent with this verb in the entire history of Eng-
lish in any case, which is due to the outstanding return of that-clauses in recent
years. Both infinitives and gerunds were increasingly marginal with deny in the
history of English.
6. Summary and conclusions

6.1. Verbs of implicit negation and different types of complements in the


history of English
6.1.1. Verbs of implicit negation and to-infinitives

The foregoing chapters have discussed various complementation patterns en-


countered with eleven verbs of implicit negation in English, paying special at-
tention to their historical developments. The verbs investigated in Chapter 2 are
forbid and refuse, both of which are in usual cases followed by to-infinitives in
Present-day English. The discussion there can be summarised in the following
manner.
While the use of that-clauses is no longer allowed after the verb forbid in
current English except in the fixed form God forbid that …, the use of that-
clauses is commonly attested with the same verb in the historical data. As a
matter of fact, that-clauses were the most dominant form of complementation
of forbid up to the late Middle English period, when there was a significant rise
of to-infinitives side by side with the decline of that-clauses. Obviously, there
occurred a complement shift from that-clauses to to-infinitives from later Mid-
dle English to early Modern English with the verb forbid, and the employment
of to-infinitives is continuous with it even today, although Present-day English
observes some sporadic uses of gerunds and gerunds with the preposition from,
perhaps under the influence of some other verbs of implicit negation which are
constantly followed by gerundial constructions (e.g. prohibit and prevent). In
general, however, gerunds are marginal and by no means established with for-
bid in contemporary English.
Along with the matter of the complement shift, Chapter 2 gives a full ac-
count of the fossilised expression God forbid that …, which alone preserves
that-clauses in contemporary English. The chapter argued that there was per-
haps an interesting gap of usage between God forbid and the other uses of
forbid from the beginning in the history of English. The most striking feature
178 Chapter 6

of God forbid, which in origin corresponds to Latin absit, is the absence of


expletive negation in its subordinate clauses. This is in clear contrast with the
other uses of forbid, which in usual cases dominated subordinate clauses with
expletive negation in earlier English. In addition, forbid in ordinary usage was
more flexible than God forbid in terms of the constructions it took from the
earliest stages of its development. God forbid was usually followed directly by
that-clauses, without any intervening indirect objects, whereas the other uses of
forbid were, often though not always, followed by an indirect object plus that-
clauses.
There were, however, increasing similarities between God forbid and the
other uses of forbid from the later Middle English period onwards. Due to the
decline of expletive negation in general, which was more and more prominent
from later Middle English onwards, that-clauses, whether they were subor-
dinate to the ordinary use of forbid or to the fixed form God forbid, no longer
included expletive negation. At the same time, non-assertive forms like any
were more and more frequently evidenced from later Middle English onwards
within the verbal complements dominated by adversative predicates including
verbs of implicit negation. And this was the case not only with forbid of or-
dinary uses but also with God forbid that …. Thus, God forbid approached the
other uses of forbid around this time, in that the distinctions between them
were increasingly slight.
Unfortunately, however, that-clauses after forbid of ordinary uses were in-
creasingly replaced by to-infinitives and occasionally by gerunds before the
merger of the two types of forbid was fulfilled in English. In other words, the
complement shift from that-clauses to non-finite forms occurred only with
forbid in ordinary usage in the end, and God forbid lost the chance or the
timing to merge into this historical development. Consequently, Present-day
English still exhibits the existence of that-clauses in the construction God
forbid that …. Certainly, the fact that it was inherited by later versions of the
Bible as a fixed form of translation also contributed to the survival of God
forbid plus that-clauses in Present-day English. If the two different types of
forbid had approached each other in nature before the occurrence of the com-
plement shift, however, their merger might have occurred in the history of Eng-
lish. Then the construction God forbid that … might not be as fixed as in
Present-day English, although this is no more than simple conjecture.
To turn to the second verb refuse, the historical shift of complements was
not observed to any noticeable extent in the dataset investigated in the present
research. Despite the existence of that-clauses with various verbs with the
meaning ‘to refuse’ in Old English, refuse has always been followed by
Summary and conclusions 179

to-infinitives since it was borrowed into English. In the historical data of the
OED, there were some examples of refuse followed by gerunds, especially
during the first few hundred years of the Modern English period, but they are
marginal throughout the entire history of English.
Although the use of to-infinitives seems to be fully established with this
verb, however, there are some features which could be a sign of to-infinitives
losing ground in a way or another. The construction “refuse + object + to-
infinitives”, which was once used, is no longer a usual construction in Present-
day English. In other words, a certain usage of to-infinitives has been lost in
the history of English. The confirmation of this would, however, require further
research. Unfortunately, the number of the examples obtained in the present
study was too small.

6.1.2. Verbs of implicit negation and gerunds

The verbs discussed in Chapter 3 are forbear and avoid, which are both fol-
lowed by gerunds in Present-day English. The paragraphs in the following
summarise the discussion in this chapter.
The historical development of forbear displays fairly clearly the rise of to-
infinitives at the cost of the decline of that-clauses, and the rise of gerunds side
by side with the recession of to-infinitives. As with other verbs of implicit
negation analysed in the present study, the former shift or the first complement
shift takes place during the early Modern English period, while the latter from
the eighteenth century onwards with this verb. The second complement shift is
particularly noteworthy with forbear in negative contexts, which often occurs
in the form of “cannot forbear + –ing”. Although the gerundial construction of
forbear has commonly been accompanied by the preposition from in recent
years, as in the form “from + –ing”, the same verb is dealt with in Chapter 3 in
the present study, viz. the chapter on verbs of implicit negation followed by
gerunds without from. This is because the employment of from is much restrict-
ed with forbear when compared with the verbs discussed in Chapter 4, i.e.
prohibit, prevent, hinder, and refrain, at least in the historical data of the OED
quotations.
It is, however, noticeable that more than half of the relevant examples of
forbear present the preposition from as far as the data in the BNC are con-
cerned. This is most likely related to the fact that the meanings of ‘avoiding’,
‘shunning’, and ‘refraining’ have increasingly been prominent with forbear
rather than the meanings of the opposite direction, i.e. ‘bearing’, ‘enduring’,
180 Chapter 6

and ‘submitting’ in recent years. In fact, the OED states that the latter meanings
are obsolete in contemporary English. Here, the construction in which the verb
refrain occurs may be of some influence upon the syntax of forbear. Finally,
the employment of the verb forbear itself is restricted in use in current English.
There are only scanty examples of forbear in the twentieth-century quotations
in the OED, while none of LOB, FLOB, Brown, and Frown provides relevant
examples of forbear. Furthermore, the Cambridge Advanced Learner’s Dic-
tionary notes that it is a verb used only in formal contexts in contemporary
English.
By contrast, the employment of avoid, the second verb discussed in
Chapter 3, seems to be increasingly common these days. While it is a newer
word than forbear in the history of English, it undergoes a faster shift of com-
plements and an earlier establishment of the construction with –ing. Despite the
fact that relevant examples of avoid itself are rather limited from later Middle
English to early Modern English in the data of the present study, the to-
infinitival construction was most probably available to some noticeable extent
with the same verb in the sixteenth century. By the time of the eighteenth
century, however, the use of the form “avoid + –ing” was more or less
established in spite of the existence of some comments in favour of the use of
to-infinitives in grammars in the past. In Present-day English, the gerundial
construction is practically the only type available to this verb.
There may be some factors behind the earlier establishment of the gerund
with avoid than forbear in their historical developments. The fact that avoid
was a newcomer in English may be relevant to this issue. Since forbear is not
free from the longer history it has in English, where older constructions were
fully used, it may have been more difficult for it to discard them abruptly in
their entirety. Also, the fact that the frequency of avoid itself has expanded in
the history of English may be relevant. It is perhaps a reasonable conjecture
that the growing process of the construction of –ing has been accelerated by the
growing frequency of the verb avoid itself as I discussed in this chapter. And
this was not really the case with forbear.
Finally, Chapter 3 discussed a phenomenon often called horror aequi in
the literature, whereby the occurrence of similar constructions side by side is
inclined to be avoided in use. This is considered to be a general rule applicable
to various phases of language in general as a number of previous studies state.
In the case of complementation patterns, the horror aequi principle is evi-
denced in the avoidance of –ing followed by –ing and to-infinitives followed
by to-infinitives.
Considering the fact that avoid is almost always followed by –ing in
Summary and conclusions 181

contemporary English, there is no room, at least theoretically, for the horror


aequi principle to work with this verb. The employment of –ing is practically
the only possible option, in whichever form the verb avoid itself may occur.
Interestingly enough, however, more than half of the relevant examples of
avoid in the OED quotations are found in the to-infinitival form, i.e. to avoid,
and as a result “avoiding + –ing”, where similar forms co-occur side by side, is
extremely rare, satisfying perhaps unawares the horror aequi principle.
Furthermore, the historical dataset provides some additional examples in sup-
port of the principle of horror aequi, where the form avoiding is inclined to be
followed by to-infinitives rather than by –ing. This suggests that horror aequi
was functionally relevant with this verb when both to-infinitives and gerunds
were possible options with this verb.
Incidentally, the verb refuse investigated in Chapter 2 is a verb almost reg-
ularly followed by to-infinitives throughout the history of English. Interestingly
again, it often occurs in the gerundial form refusing, consequently providing
the “refusing + to-infinitive” construction. In other words, the horror aequi
principle is again satisfied, perhaps as a result, in many cases where this verb is
involved, although the possibility for this to be merely accidental may not be
eliminated.

6.1.3. Verbs of implicit negation and gerunds with prepositions

Chapter 4 dealt with the historical development of the four verbs prohibit, pre-
vent, hinder, and refrain. These verbs often, though not always, display the
gerundial construction with the preposition from in Present-day English. The
discussion there may be summarised in the following manner.
Although “prohibit + from + –ing” is a fully established construction in
contemporary English, prohibit presents that-clauses and to-infinitives in the
history of English. As far as the OED quotations are concerned, relevant ex-
amples of this verb are available only from the sixteenth century, when the to-
infinitive is the dominant form followed by that-clauses. Thus, the complement
shift from that-clauses to to-infinitives was already taking place at the begin-
ning of the Modern English period with prohibit, which is in accordance with
the development of some other verbs of implicit negation examined in the
present study.
The recession of that-clauses is even clearer by the time of the seventeenth
century with this verb, when the second complement shift or the rise of gerunds
at the cost of to-infinitives launches. In the seventeenth century, both
182 Chapter 6

to-infinitives and gerunds are frequent with prohibit, whereas from the
eighteenth century onwards to-infinitives are increasingly infrequent and
gerunds take their place. Prohibit is indeed a verb which typically displays the
complement shift from that-clauses to to-infinitives and then the subsequent
shift from to-infinitives to gerunds in a most clear form. The first shift takes
place in the early Modern English period and the second shift in the late
Modern English period with this verb.
Finally, there is an additional phase with prohibit during later periods, i.e.
the increasing use of the preposition from. When the rise of gerunds occurred
with prohibit in the seventeenth and eighteenth centuries, they were still used
more frequently without from than with from. The eighteenth century observes
a gradual rise of “prohibit + from + –ing”, which becomes the dominant form
from the nineteenth century onwards. It is, therefore, safe to conclude that this
is a phrase which became prominent only after the use of gerunds expanded to
some noticeable extent.
One thing which merits our attention in respect of the verb prohibit is that
it does not provide any examples of but-clauses or lest-clauses in its historical
development, at least as far as the OED quotations are concerned. It is also
noticeable with the same verb that the limited examples of that-clauses have a
strong tendency to include expletive negation. These facts suggest that the
complement shift from that-clauses to to-infinitives was a fairly quick and
successful one in the case of this verb. It is probable that that-clauses of
prohibit did not really have a chance to struggle for survival by utilising other
finite forms like but-clauses and lest-clauses when the weakening of that-
clauses took place with the same verb.
Since this is a mere assumption based upon the OED quotations data only,
however, I do not deny the entire possibility of finding prohibit but … and
prohibit lest … in other historical data. I would simply postulate that they are
most probably uncommon if they are ever attested. In any case, the relevant
quotations of prohibit in the OED illustrate fairly clear-cut shifts of com-
plements which took place in the history of the English language.118
The second verb discussed in Chapter 4 was prevent, which in my view
followed a similar path to the historical development of prohibit. As far as the
OED quotations are concerned, however, relevant examples are limited in

118
Graeme Davis has kindly pointed out that prohibit followed by lest is found in Mackenzie
(1825: 413) (personal communication). Thus, examples of prohibit lest … are available,
though not common, outside the OED quotations. In any case, the gist here is that the
complement shift was fairly smooth in the case of this verb.
Summary and conclusions 183

number in the sixteenth century, when gerunds are already the most frequent
form of complementation of this verb. By the time of the seventeenth century,
the use of gerunds is almost fully established with prevent. The focus of the
discussion in Chapter 4 was, therefore, placed upon the variation among the
three gerundial patterns, all of which are available in Present-day English as
well as in historical data: (a) “prevent + possessive + –ing”, (b) “prevent +
object + –ing”, and (c) “prevent + object + from + –ing”. Gerunds without from
are still left more abundantly with prevent than with prohibit discussed above,
which is indexical of the possibility that the process of the development of
gerunds themselves has perhaps been slower with prevent than with prohibit,
although the predominance of gerunds as against to-infinitives was attained
earlier with prevent than with prohibit, as far as the dataset of the OED
quotations is concerned.
Of the three patterns of gerunds, “prevent + possessive + –ing” is on a
gradual decline in current English, although it is still observed well especially
with personal pronouns today. In fact, it is most significant in the analysis of
the historical development of prevent to make a separate treatment between the
cases of nouns and personal pronouns used as the semantic subject of the
gerund, since there is a clear distinction in terms of the variation of the three
gerundial patterns between the two cases. With nouns (rather than personal pro-
nouns), the decline of the possessive form occurred much earlier, whereas with
personal pronouns it lingered much longer in the history of English, almost
hindering the expansion of “prevent + object + –ing”.
Finally, whether “prevent + object + –ing” is a short form of “prevent +
object + from + –ing” was considered. The discussion in Chapter 4 strongly
suggests that the objective form of the former construction arises at the cost of
the decline of the possessive form of “prevent + possessive + –ing” rather than
from the ellipsis of the form “prevent + object + from + –ing”, and this
conjecture is consistent with what Rohdenburg (1996) calls the Complexity
Principle, which is applicable to various aspects of language. In the con-
temporary setting, however, the inference that the omission of from is involved
in the development of “prevent + object + –ing” cannot entirely be eliminated,
since it is not a matter of surprise if language users consider in this manner
once the decline of the possessive construction is much in progress, leaving the
situation in which only the patterns with the objective with or without from are
available. Under such circumstances, the increase of one form is possible only
at the cost of the decrease of the other in the end.
Hinder, the third verb discussed in Chapter 4, is in many respects similar
to prevent. The historical shift to the gerund was so quick with it that the rise of
184 Chapter 6

to-infinitives was almost invisible in its historical development. Gerunds were


the most frequent form already in the sixteenth century with hinder, and by the
time of the seventeenth century they were already predominant. As for the rise
of the form “hinder + object + from + –ing”, there is again a marked difference
between the cases of personal pronouns and nouns employed as the semantic
subject of the gerund. Both the introduction and the establishment of from were
earlier when the object of hinder was a noun than when it was a personal pro-
noun.
The chief difference between hinder and the first two verbs discussed in
Chapter 4, i.e. prohibit and prevent, is found in the timing of the historical de-
velopment thus far described. Especially, the development of the preposition
from takes place much earlier with hinder than with prohibit and prevent.
Already in the seventeenth century, the gerund accompanied by from is pre-
dominant with hinder, while this is not at all the case with prohibit and prevent.
Hinder is about two centuries earlier in this respect than prohibit, and perhaps
about one or two centuries earlier than prevent. In this sense, it may have had a
leading function in the development of the gerundial construction with the
preposition from in the history of English. Or, at least one could state that it
was one of the leading verbs in this matter.
An additional point to note in the historical development of hinder is that
various older constructions lingered with this verb to a fairly larger extent in
later years, despite the earlier development of hinder in general. That-clauses
and but-clauses are remnant, though to a very minor extent, with hinder till the
nineteenth century. This may be ascribable to the fact that hinder is an older
verb than prohibit and prevent, both of which are loan words in English, and
therefore it was more difficult for hinder to eliminate the long-existing
syntactic variations in a complete manner. Thus, hinder turns out to be a verb
which led the historical development of the “from + –ing” form on the one
hand, while on the other hand it is a verb which kept showing older and
remnant constructions for a longer period.
Lastly, Chapter 4 investigated refrain. It also developed the gerund with
the preposition from in its historical development but is different in nature from
the other three verbs in the sense that it is often intransitive in Present-day
English. Despite the usage of this verb in contemporary English, however, the
analysis of the OED quotations reveals that the transitive use of refrain was not
uncommon in the past, although its object was often a reflexive one when it
occurred transitively. This is perhaps one of the reasons why refrain followed a
similar path to those followed by prohibit, prevent, and hinder. As the other
three verbs, refrain also underwent the rise of to-infinitives and gerunds in its
Summary and conclusions 185

historical development, although the whole process took place within such a
short period of time that the complement shift from that-clauses to to-
infinitives was almost invisible in occurrences.
The rise of gerunds was also early with this verb. The data present some
notable number of to-infinitives used as its complement during the early
Modern English period, but gerunds are always predominant from the sixteenth
century onwards. Furthermore, the employment of the preposition from was
established fairly early with the same verb. As far as the quotations in the OED
are concerned, the construction with the preposition from is dominant already
in the early Modern English period, and it is the sole form observed with it in
the nineteenth and twentieth centuries, although there are some irregularities in
the data. Thus, the timing of the historical development of refrain is relatively
early and almost comparable to that of hinder, which also reveals early shifts of
complements in the history of English. Despite the early development of
gerundial constructions with refrain, however, it is uncertain whether it really
performed a leading function among the same group of verbs, since its tran-
sitive use quickly receded in the course of its historical development, which
eventually differentiated the same verb from other verbs like prohibit and
prevent.

6.1.4. Verbs of implicit negation and subordinate clauses

The three verbs of implicit negation investigated and discussed in Chapter 5 are
most commonly followed by finite clauses even today. The verbs are fear,
doubt, and deny. The gist of the discussion in this chapter can be summarised
in the following manner.
In respect of the verb fear, there is a sharp contrast between the verb used
in negation and the same verb employed in affirmative sentences, in terms of
the historical development of its complements. The occurrence of fear in
negative contexts is rather restricted in frequency in general, but its limited
examples in the data of the present study illustrate a relatively prototypical
form of complement shifts in the history of English. The early Modern English
period saw the rise of to-infinitives with this verb, which was later on caught
up with by the rise of gerunds. Hence, the first and the second complement
shifts are evidenced with this verb when it is used in negation.
These shifts are, by contrast, not at all prominent with fear used in af-
firmative sentences. This contrast is most likely due to the fact that the same
verb in affirmative sentences took an entirely different path of development
186 Chapter 6

from the path taken by the same verb in negative environments, from the early
Modern English period onwards. What is outstanding in the case of affirmative
sentences is the development of its parenthetical use, which is first charac-
terised by the ellipsis of the conjunction that and then by the dislocation of the
fixed forms I fear, etc., i.e. the dislocation from the original position to sen-
tence-medial and final positions. In the parenthetical use, the subject of the
verb fear tends to be the first person (especially singular). It also tends to be in
the present tense. Then the fixed form functions almost as a sentential adverb.
The dataset of the OED presents copious examples of potential candidates
for the parenthetical use with fear in affirmative sentences from the early Mod-
ern English period onwards, despite the contention by Wierzbicka (2006: 2007)
that the development of parentheticals is relatively late in the history of English
and that it is observed from the eighteenth century onwards. The present study
demonstrates that tendencies evidenced in the OED are more in accordance
with the argument by Brinton (2008: 223–224), who considers that the de-
velopment at issue is much earlier and that it goes back at least to the Middle
English period.
One interesting point revealed in the present research is that there seems to
be the opposite inclination to use that-clauses of the original kind in the twen-
tieth century. This may indicate a possible re-establishment of clauses of this
type in recent years with fear, although further research is certainly called for
to prove this point. It is difficult to eliminate the possibility that the
twentieth-century materials in the OED quotations reflect relatively strong
editorial intervention, which may have led to the “expression” (rather than
ellipsis) of the conjunction that in them.
Side by side with that-clauses, to-infinitives, and gerunds, one also notices
that the verb fear is occasionally followed by but-clauses and lest-clauses. This
phenomenon is evidenced in some measure at least up to the nineteenth century.
Here again, there is a contrast between fear used in negation and the same verb
used in affirmative sentences. Clauses introduced by but occur with the former,
whereas lest-clauses are particularly favoured in the latter context. The use of
but-clauses and lest-clauses suddenly declines around the turn of the twentieth
century, however. Indeed, fear in Present-day English may need further in-
vestigation, but one could possibly postulate on the other hand that the twen-
tieth century was a period of “consolidation” of the various possible patterns of
complementation of this verb. Hence, it observed the decline or obliteration of
some forms and the revived establishment of others. The returning of that-
clauses with fear can be understood in this linguistic environment.
Doubt, the second verb investigated in Chapter 5, followed a similar path
Summary and conclusions 187

to that of fear in the history of English, in the sense that it also developed, in
the process of its historical development, epistemic parentheticals in af-
firmative sentences. The period when that-clauses were increasingly unstable,
which was perhaps due to the demise of the subjunctive, saw the rise of non-
finite forms like to-infinitives and gerunds with this verb to some extent at least.
On the whole, however, they stayed minor throughout the history of English,
and doubt has always been more comfortable with finite clauses including not
only that-clauses, whether-clauses, and if-clauses, but also but-clauses and lest-
clauses—but-clauses and lest-clauses were mostly found in the early Modern
English period. The discussion of doubt in Chapter 5, therefore, placed more
emphasis upon finite clauses in general than upon non-finite forms of com-
plementation.
Further findings on affirmative doubt include the fact that a fairly constant
use of whether-clauses after affirmative doubt is a gradual development in
Modern English, although the same verb is usually considered in grammars and
dictionaries to yield whether-clauses in affirmative sentences and that-clauses
in negative ones in Present-day English. Also fairly recent is the expanded use
of if-clauses after affirmative doubt. Its expansion in the twentieth century is
noteworthy in the dataset of the OED quotations. At the same time, however,
the absence of if-clauses in the Bible in English on CD-ROM leads us to the
conjecture that its use is somewhat stylistically conditioned.
To turn to negative sentences, clauses introduced by whether and if are
marginal during the entire history of doubt. Negative doubt is often followed
by that-clauses instead. The stability of that-clauses after doubt in negative
environments was, however, not always secure in the history of English. That-
clauses after negative doubt also underwent a critical period from later Middle
English onwards, yielding various other possibilities like to-infinitives, gerunds,
and but-clauses. While subordinate clauses introduced by the conjunction lest
were not available in negation, those introduced by but were much more
abundant in negative contexts than in affirmative ones. This is reflected in the
Present-day English situation where but-clauses are still attested especially in
negative sentences, though restricted in frequency. As in the case of doubt in
the affirmative, the use of that-clauses after doubt in negation came back again
once the critical period was over in the course of the history of English,
particularly in the twentieth century. Today, the use of that-clauses after doubt
in negative contexts is firmly established, or rather re-established.
Finally, doubt in negation followed by that-clauses did not lead to the de-
velopment of the parenthetical use to the same extent as doubt in affirmative
sentences did. It is true that the context of negation itself is not favourable for
188 Chapter 6

the development of the parenthetical use, but does not make it wholly im-
possible. Chapter 5 proposed that there were possibly some additional factors
hindering the rise of this use, one of which was the fact that negative doubt was
busy dealing with the competition between I doubt not and I do not doubt (and
also I don’t doubt) during the Modern English period. It is perhaps difficult for
fluctuating forms like this to become a parenthetical, since the process of the
development of the parenthetical use is also a process of becoming more and
more fixed or frozen in form. Unfortunately, the development of the auxiliary
do was taking place at the same time, and doubt, which was relatively slow in
adopting the auxiliary do, was certainly fluctuating between the forms with do
and without during the critical period.
Another possible factor perhaps relevant is the constant increase of no
doubt in the history of English, which derives from the noun doubt. Since no
doubt often functioned as an adverb at the sentential level, it was not necessary
for I doubt not or I do not doubt to take the responsibility of functioning at the
sentential level. Furthermore, the existence of doubtless may also be a relevant
factor in this concern, although this has not really been proved in the present
study, due to the limitation of the data under analysis. Since the OED is a
collection of quoted examples, it is not suitable for the type of studies that is
interested in the increase or decrease of a particular word, in this case doubtless.
At least, the limited examples of doubtless in the OED quotations show that
they are often used as a sentential adverb in the meaning of ‘no doubt’ in its
historical data. In any case, there were other options available than the
parenthetical use of negative doubt—no doubt and possibly doubtless—, and
the other options were often more attractive than the fluctuating forms of I
doubt not and I do not doubt.
In the last section of doubt, I touched upon the verb question, which is a
different verb but which is occasionally used in the meaning ‘doubt’. Indeed,
there are some similarities or parallelisms between question and doubt in terms
of their syntactic structures and their historical development. For example,
question is, like doubt, more frequently followed by whether-clauses and
if-clauses when it is used in the affirmative than in the negative. In addition,
question in negative environments followed by but-clauses is typically found in
the early Modern English period, which is again a shared feature with doubt.
Furthermore, there are some limited number of to-infinitives and gerunds
encountered with question in the seventeenth and eighteenth centuries, which
phenomenon is attested with doubt as well.
On the other hand, however, there is also a significant difference between
the verbs question and doubt. Question is much more frequently followed by
Summary and conclusions 189

complements of the “open type” (e.g. when-clauses, how-clauses), suggesting


that it is more neutral as to whether or not the content matter of the com-
plement clause is true. In this sense, question can better be categorised under
verbs of asking than under verbs of implicit negation, although the clas-
sification of different verbs is not the central goal of the present study. In the
case of doubt, complements of the “open type” are not considered to take place,
although in practice they are available to a minor extent in the historical data of
the OED.
The last verb discussed in Chapter 5 is deny. Although it is a verb most
typically followed by that-clauses in Present-day English, the historical path
represented by it differs significantly from the paths followed by fear and
doubt, both of which are also followed by that-clauses today. Unlike the cases
of fear and doubt, the development of the parenthetical use is not at all obvious
with deny even in affirmative sentences. This may be due to, or at least relevant
to the fact that the occurrence of first person subjects is rather rare with deny
throughout the history of English. In other words, the conditions for the verb to
become an epistemic parenthetical were hardly fulfilled in the case of this verb.
The other path that deny could have taken then is to undergo the shifts of
complements by which to produce non-finite forms of complementation.
Indeed, to-infinitives were as frequent as that-clauses in later Middle English
and became the dominant form of complements with this verb in the sixteenth
century. Thus, there is indeed a sign of the occurrence of the first complement
shift with deny. The rise of to-infinitives was, however, not at all complete in
the end. They were never so frequent as to obviate the older construction with
that-clauses, in the historical development of this verb. That-clauses gradually
became frequent again, and by the time of the nineteenth century, their use
achieved dominant status and was fully re-established with deny.
Presumably because the rise of to-infinitives at the cost of that-clauses was
incomplete with this verb, the second shift or the rise of gerunds is not clearly
visible, either. The existence of to-infinitives led to the rise of gerunds to some
extent in later Modern English, but this was already a period when the use of
that-clauses was being re-established with deny. After all, the expansion of
gerunds in later Modern English is not at all outstanding with it. There is, how-
ever, one interesting aspect which is perhaps related to the decline of to-
infinitives and to the rise of gerunds. In the course of its historical development,
deny has lost, except in dialectal varieties of English, the construction where it
is directly followed by the to-infinitive (i.e. without the intervening object),
which was once a possible construction with it. As a result, the construction has
to be in the form “deny + object + to-infinitive” in current English, when it
190 Chapter 6

occurs with the to-infinitive. This is certainly a sign for this verb to lose some
of its syntactic possibilities. Interestingly enough, this change coincides with
the rise of constructions with simple gerunds (i.e. without the semantic subject
of gerunds) in historical terms. It may, therefore, be a reasonable postulation
that the capability of the to-infinitive became slightly restricted with this verb
and that the potential of the gerund expanded to some extent in this process. In
other words, the complement shift from to-infinitives to gerunds is observed in
this humble manner with deny. It lost a certain minor construction with to-
infinitives and gained a certain construction with gerunds.
In any case, the verb deny illustrates a mixed case of various developments
so far discussed in the present research. Simultaneously, however, none of the
paths related to historical shifts of complements is clearly represented by this
verb. Neither the first complement shift nor the second shift is evidently visible
with this verb.

6.2. Historical shifts of complements in general


6.2.1. The rise of to-infinitives, gerunds, and epistemic parentheticals

The discussions in the preceding chapters reveal that the period from later Mid-
dle English to early Modern English is one of the key periods in terms of the
chronological shifts of complements which various verbs of implicit negation
experienced, although some verbs, especially those borrowed from foreign
languages into English, provide relevant data only from the early Modern Eng-
lish period onwards, making the situation of Middle English difficult or even
impossible to analyse.
Apparently, there was an unequivocal decline of that-clauses, most proba-
bly coinciding with the decline of the subjunctive in English, during this key
period, which in consequence prepared various paths of development for dif-
ferent verbs to take. The most typical is the rise of to-infinitives in place of
that-clauses, which is observed in a transparent manner with forbid, forbear,
prohibit and to some extent with fear and deny, as far as the verbs investigated
in the present study are concerned. This shift in many cases, though not always,
leads to the second complement shift in later Modern English, namely the shift
from to-infinitives to gerunds. Some of the eleven verbs discussed in the
present research display a sudden increases of gerunds without showing any
clear and prior increase of to-infinitives, which may be ascribed to the fact that
the interval between the two stages of the historical development is rather too
Summary and conclusions 191

short and therefore almost invisible in extant data. This is the case with verbs
like avoid, prevent, hinder, and refrain.
As regards the verbs which reach the stage of the expanded use of gerunds,
there is an additional phase to be considered. Some verbs further develop the
construction of gerunds with the preposition from—other prepositions are also
found on occasions—, whereas others do not. The rise of the structure with
from usually takes place slightly later than the establishment of the gerund it-
self. Verbs typically showing this later development include prohibit, prevent,
hinder, and refrain, all discussed in Chapter 4. Of these four verbs, prohibit,
hinder, and refrain are most frequently encountered in this new form in
Present-day English, whereas prevent gives the older construction without from
to some noticeable extent even today, especially when the semantic subject of
the gerund is in the form of a personal pronoun. In addition, the verb forbear
also exhibits the development of the gerundial construction with from to some
noticeable extent, but it was dealt with in Chapter 3 in the above discussion.
This is because the construction without from is more prominent with this verb
when viewed from the historical perspective.
Furthermore, there is an important group of verbs which took a totally dif-
ferent path from the ones described above, especially when they occur in af-
firmative contexts. As far as the verbs investigated in the present study are con-
cerned, fear and doubt are a case in point. They also experienced the instability
and the decline of that-clauses as other verbs of implicit negation did, from
later Middle English to early Modern English, but instead of seeing the rise of
to-infinitives, the matrix clauses themselves developed into an epistemic paren-
thetical, transforming the original subordinate clauses introduced by the con-
junction that into matrix ones. The original matrix verbs now function in the
forms of I fear and I doubt almost as an adverb at the sentential level. I would
postulate that this is another way to solve the problem caused by the recession
or the weakening of that-clauses in the history of English, which was most
probably caused by the decline or even demise of the subjunctive, although
admittedly there are contentions in existing studies which assert that paren-
theticals are parentheticals from the beginning and not really the result of any
complement shifts.
One characteristic feature of these verbs to note is that they are inclined to
observe in some measure the revival of that-clauses in the twentieth century.
This is at least the case with both fear and doubt, as far as the quotations in the
OED are concerned. Fear and doubt seem to occur commonly in current Eng-
lish with that-clauses where the conjunction that is expressed, although further
research is certainly necessary to confirm this point, especially in view of the
192 Chapter 6

fact that twentieth-century data in the OED are not sufficiently representative
of Present-day English.
Finally, it is relevant to mention that the path which fear and doubt
followed is a path prepared for contexts where pragmatic inference works at a
high level. Thus, the development of the parenthetical use of fear and doubt is
observed mainly in the affirmative. To turn to the same verbs used in negative
environments, relevant examples themselves are not so numerous as to provide
a transparent picture of their historical development. However, at least the verb
fear in negation displays the rise of to-infinitives and gerunds instead in the
course of the Modern English period, which characterises the path followed by
various verbs of implicit negation in general. Hence, a significant gap is exist-
ent between affirmative and negative environments when verbs of this group
are concerned.

6.2.2. The decline of expletive negation and the introduction of but

While some verbs, especially those borrowed from other languages, have a
relatively short history in English, others go back to a much earlier period.
During the Old and Middle English periods, subordinate clauses dominated by
verbs of implicit negation usually included expletive negation, which gradually
declined from later Middle English onwards. From this perspective, the rise of
non-finite complement forms like to-infinitives and gerunds was a favourable
development, which was well in conformity with the historical development of
English, since it removed subordinate clauses themselves, accelerating the
elimination of expletive negation, which was clearly on the decline in any case.
Although there were some occasional examples of to-infinitives and gerunds
including expletive negation during the transitional period from later Middle
English to early Modern English, it was essentially a phenomenon evidenced
within that-clauses. In other words, the obliteration of that-clauses could auto-
matically delete the phenomenon of expletive negation.
On the other hand, when verbs were to retain the use of that-clauses after
the late Middle English period, it was essential for them to deal with the prob-
lem of the decline of expletive negation, which was still included in that-
clauses. In theory, they could simply remove expletive negation and dominate
that-clauses devoid of it, but they often selected but-clauses instead of that-
clauses, preserving the tone of expletive negation after it receded. This was
often the case when verbs themselves occurred in negative environments or
Summary and conclusions 193

perhaps in interrogation,119 whereas after ordinary affirmative propositions,


lest-clauses were more commonly employed instead. Since this is a phe-
nomenon related to the receding of expletive negation, clauses introduced by
but and lest are observed most characteristically during the transitional period
from later Middle English to early Modern English, and become increasingly
restricted in use in later periods. Furthermore, there are some examples of
to-infinitives preceded by but from later Middle English to early Modern
English, but this seems to be no more than an analogical form, and is quite
similar in a way to the occurrence of expletive negation used with to-
infinitives.
Concerning the eleven verbs discussed in the present study, the oc-
currences of but-clauses and lest-clauses are particularly noticeable with those
which preserved the use of that-clauses for a longer period than others, e.g.
fear and doubt. By contrast, some verbs quickly shifted to the use of to-
infinitives in place of that-clauses, and in this case it was presumably un-
necessary for them to utilise transitional forms of complementation. But-
clauses and lest-clauses are simply absent or only marginal with these verbs.
Prohibit, which is treated in Chapter 4, for example, is a verb of this kind.
As for contemporary English, subordinate clauses introduced by but, if
they occur, are often felt to be archaic. Though much restricted in frequency,
this usage of but is not yet obsolete today.

6.2.3. Lexical diffusion and historical timings

Although the term “lexical diffusion” is frequently employed in the context of


phonology, there has been an increasing awareness in recent studies that the
same concept is often applicable to syntactic changes as well. Tottie (1991:
439) maintains that “[m]uch less attention seems to have been paid to the prob-
lem of regularity versus lexical diffusion in syntax”, while at the same time she
argues that “[i]n both morphology and syntax, lexical diffusion seems to have
been implicitly taken for granted by many writers”. Likewise, Nevalainen
(2006: 91) points out in the context of syntactic developments the fact that “the

119
Although relevant examples of interrogative sentences are by no means abundant in the
data investigated in the present study, the limited attestations of relevant examples are in
keeping with this argument. See also Curme (1931: 241) among others on the employment
of the conjunction but after negative and interrogative clauses containing verbs of implicit
negation. According to him, but what is also encountered in the same syntactic environ-
ments in colloquial speech.
194 Chapter 6

incoming form does not spread to all contexts at once but some acquire it
earlier than others”, and says that the phenomenon is called “lexical diffusion”.
In this manner, the concept of lexical diffusion is extendable to various lin-
guistic changes, including syntactic ones.
It is a useful concept for the purpose of the present study as well. While
the overall development of complementation patterns as described above (cf.
6.2.1. and 6.2.2.) is shared by a number of verbs of implicit negation in English,
some shifts take place earlier with some verbs and later with others. When for-
bear and avoid are compared, for instance, the establishment of the gerundial
construction took place much earlier with avoid than with forbear. Con-
sequently, the shift from that-clauses to to-infinitives and the shift from to-
infinitives to gerunds are more clearly evidenced with forbear, which took a
longer time in undergoing the same processes. By the same token, the rise of
to-infinitives and the rise of gerunds are most prominently evidenced with the
verb prohibit among the four verbs discussed in Chapter 4. This is most
probably due to the fact that the predominance of gerunds was reached later
with prohibit than with prevent, hinder, and refrain.
Interestingly enough, the speed of the historical development of various
complementation patterns is not always constant, either. Prohibit is indeed
slower than the verbs prevent, hinder, and refrain in establishing the gerundial
construction. Turning to the expansion of the preposition from attached to the
gerund, however, a slightly different picture emerges, at least as far as the
quotations in the OED are concerned. The establishment of the construction
with from seems to have taken place earlier with prohibit than with prevent, if
not earlier than with hinder and refrain. As a result, the gerundial construction
with from is much more frequent with prohibit than with prevent in Present-day
English. Among the verbs explored above, the earliest on this front are hinder
and refrain, both of which already observe a significant rise of the construction
with from in the very early period of early Modern English.
There are, however, footnotes to their historical developments. Although
hinder is always more progressive than prohibit and prevent, it also retains old-
er constructions like that-clauses, but-clauses, and to-infinitives until a rela-
tively late period, at least in comparison to prevent. That hinder is an older verb
than prevent may be relevant to this situation. Having a longer history in which
older forms were once normally used, hinder may have found it more difficult
to lead them to complete deletion. In respect of refrain, on the other hand, the
early shifts of complements are indeed noteworthy with this verb, but its use is
increasingly intransitive in the history of English, and it gradually ceases to be
Summary and conclusions 195

a leading verb in relation to prohibit and prevent, which are still transitive in
nature.
Finally, deny is a verb which was not really successful in any of the
complement shifts attested in the history of English. Although the rise of to-
infinitives occurred with it from later Middle English to early Modern English,
it did not in the end reach the stage of eliminating that-clauses. On the other
hand, the verb does not really display the development of the parenthetical use,
either. Probably because the first complement shift was imperfect, the second
complement shift was also far from perfect with this verb. Hence, the only
marginal development of gerunds is attested on the one hand and the return of
that-clauses on the other hand, with deny. It displays a mixed feature of various
complement shifts that took place in the English language, and the strict path
taken by it is not at all transparent.

6.2.4. Meaning and form again

Meaning and syntactic constructions are mutually related, especially when dif-
ferent forms of complementation are simultaneously available to a single and
the same verb. Doubt, for example, often gives the meaning ‘to hesitate’ when
it is directly followed by the to-infinitive, which is found in the historical data.
Also, forbear in the meanings ‘to avoid’, ‘to shun’, and ‘to refrain’ shows a
tendency to appear in the construction with the preposition from rather than in
the construction without it. Moreover, deny is inclined to be in the meaning of
‘refuse’ when it is followed by to-infinitives.
This does not imply, however, that there is always a close and predictable
relationship between the meaning of the verb and the construction it takes or
that a choice of a certain construction is always explainable from the perspec-
tive of meaning. The fact that a number of the verbs investigated in the present
research have undergone rather dramatic shifts of constructions in the history
of English proves this point, at least to some extent, since shifts of con-
structions often take place even when meanings stay the same or at least re-
latively stable. Additionally, it is fairly easy, as discussed in the above chapters,
to give some English verbs whose meanings are similar and whose choice of
constructions differs to a significant extent. For example, forbid and prohibit
are similar in meaning, while the usual constructions they present are different
in contemporary English. The former is usually followed by to-infinitives,
though occasionally followed by gerunds, whereas the latter verb is usually fol-
lowed by gerunds with from. Likewise, dictionaries list the meaning ‘to avoid’
196 Chapter 6

under the entry of refuse (e.g. OED, s.v. refuse), while the complementation
patterns revealed by refuse and avoid differ. Refuse almost exclusively occurs
in combination with to-infinitives, whereas avoid is almost always followed by
gerunds in Present-day English.
It is, therefore, essential not to place excessive emphasis upon the link
between meanings and forms. In synchronic terms, De Smet and Cuyckens
(2007) refer to various factors including sociolinguistic, pragmatic, cognitive,
and stylistic ones, and conclude that “the distribution of complement types is
the outcome of a subtle balance between several motivating factors” (p. 188).
In diachronic terms, more dynamic factors such as the decline of the sub-
junctive are most probably relevant. To be more precise, the decline of the
subjunctive, which was ultimately due to the loss of verb endings, led to the
instability of that-clauses in general from later Middle English period onwards.
As repeatedly argued in the above chapters, this was compensated for by
various means, one of which was the employment of non-finite forms, as a
complement. The rise of the parenthetical use of various verbs as in I fear and I
doubt was another.
Furthermore, the internal development of to-infinitives and gerunds per se
is most probably also relevant to the complement shifts under discussion. As
Haspelmath (1989) notes among others, the to-infinitive was in origin allative
and then purposive in function, which later obtained larger flexibility and came
to cover wide range of functions, one of which was the function of com-
plementation.120 Similarly, Hopper and Traugott (2003: 189) refer to Noonan
(1985: 47–48) and state that “[t]he use of an allative-dative marker as a com-
plementizer is common … when the tense of the complement is determined by
the nature of the main clause verb”.
Although both infinitives and gerunds were nominal in origin, it is often
claimed that infinitives are more verbal in character than gerunds (see
Hamawand 2002: 60, 80, and Vosberg 2006: 35, among others).121 A number

120
Haspelmath (1989) discusses this historical development, i.e. the shift of the nature of to-
infinitives, in the context of grammaticalisation (grammaticisation in his term), and treats
the same phenomenon in various languages, paying special attention to Germanic ones.
Among the various phenomena pointed out by him, it is particularly interesting to note that
the new purposive form om te-infinitive in Dutch is following the same path in the sense
that it can also be used in “complement clauses with irrealis modality”, though not in
“complement clauses with realis modality” (p. 303).
121
See also the verbal-to-nominal scale provided by Vosberg (2006: 36), according to which
that-clauses are more verbal than infinitives and infinitives are more verbal than gerunds of
various types. Vosberg refers to Ross (1974: 114–115) and Lehmann (1988: 200), both of
whom provide similar scales.
Summary and conclusions 197

of studies indeed argue that gerundial complements are similar, or at least more
similar than infinitival ones, to nouns or substantives (e.g. Duffley 1995: 10,
Duffley and Joubert 1999: 256). 122 Interestingly enough, however, in the
history of English, gerunds also display the phase of becoming increasingly
verbal in nature and in this process acquire flexibility in terms of the con-
structions that they employ (see Tajima 1985, Fanego 1996b: 32–33, and
Fischer and van der Wurff 2006: 178–179, among others). As claimed by De
Smet (2008), verbal gerunds are now syntactically flexible in contradistinction
to nominal gerunds.123 In other words, gerunds increasingly obtained syn-
tactic flexibility in their historical development as well, by gradually changing
their nature from the nominal one to the verbal one. Both to-infinitives and
gerunds are now able to provide sufficient information which was formerly
conveyed by a finite clause. In other words, they became increasingly ready to
take the place of that-clauses, and eventually they did so indeed, as shown in
the above chapters.

6.3. The OED as a historical corpus

As mentioned in the Introduction of the present study, the use of the OED as a
corpus was not intended when it was originally compiled. It has increasingly
been acknowledged in recent years, however, that the quotations base of the
OED provides excellent resources for linguistic analyses especially in his-
torical terms.124 Through the process of investigating eleven verbs of implicit
negation in this study, I am fully convinced of its usefulness, although I am
also aware that there are still some shortcomings involved in utilising it as a
historical corpus.
One of the main and classic problems of the OED as a corpus is that the
totals of quotations in different centuries are not constant. As Mair (2004: 124)
and Hoffmann (2004: 24–26) observe, the nineteenth century is best re-
presented in the quotations, which is followed by the twentieth century. In re-
122
Houston (1989) summarises various previous studies dealing with the verbalisation of Eng-
lish gerunds. She shows that gerunds started to take direct objects to a noticeable extent
from later Middle English to early Modern English. Thus, the key period for the decline of
that-clauses in fact coincides with the key period for the verbalisation of gerunds.
123
De Smet (2008: 90–95) displays with evidence that verbal gerunds, unlike nominal gerunds,
are inclined to be accompanied by various modifiers. Thus, he postulates that “the original
impetus for the emergence of verbal gerunds may have lain in the drive for greater
syntactic flexibility” (p. 94).
124
For details, see 1.5. in the Introduction.
198 Chapter 6

spect of the eleven verbs of implicit negation discussed in the present study, I
had no difficulty on the whole in obtaining relevant data for the seventeenth
century to the nineteenth century, when some major shifts of complements
occurred in their usage. This is a very fortunate coincidence, since it means that
I was able to collect a sufficient number of examples for the analyses of the
major historical changes of syntactic structures in English, at least within the
framework of the present research.
At the same time, however, I often felt a shortage of relevant examples for
the period earlier than the sixteenth century,125 and in some cases a slight
shortage of relevant examples for the twentieth century, although the shortage
of twentieth-century materials could easily be supplemented by the use of other
corpora of contemporary English. Considering the fact that the twentieth cen-
tury is relatively well-represented in the quotations in the OED, the lack of re-
levant examples there is rather surprising. This may simply be accidental, but a
more likely explanation is that editors selected many recent quotations for the
purpose of illustrating difficult or new lexical items and that they therefore
tended not to include ordinary verbs like the ones discussed in the present study.
In other words, it is possible that the policy of how to select examples itself has
changed within the 100 years of the twentieth century. Whatever the reason
may be, the present study, which principally relied upon the OED for the
matter of data, has left much room to be filled by future studies using other
corpora.
Despite the existence of these problems, however, I am now confident in
stating that the OED provides useful and excellent resources for linguistic
studies so long as it is handled with care and due caveat in analyses. I hope I
have proved this, at least to some convincing extent, in the present study.

125
It is relevant to repeat here that Brewer (2000: 48) comments on this issue and says: “The
period before 1580 is indeed illustrated with comparatively few quotations (not surprisingly,
given the comparatively small number of sources). By contrast, the years c1581–1610 are
heavily quoted, as much as the early decades of the nineteenth century”. The same tend-
ency has also been pointed out by Schäfer (1980: 6, 51–52), who refers to the increase of
source works after Middle English and the over-representation of the Shakespeare’s period.
Appendix

Tabulation of data for Figures 1–38

Figure 1
centuries 9th 10th 11th 12th 13th 14th 15th 16th 17th 18th 19th 20th
that-clauses 0 3 0 2 8 14 8 7 7 0 4 1
infinitives 1 2 0 0 3 15 8 40 74 42 73 33
gerunds 0 1 0 0 1 2 2 11 14 15 32 10

Figure2
OE1 OE2 OE3 OE4 ME1 ME2 ME3 ME4 Mod1 Mod2 Mod3
that-clauses 0 1 5 2 3 0 1 1 0 0 0
infinitives 0 0 2 1 1 0 3 6 1 6 2
gerunds 0 0 0 0 0 0 1 2 0 1 2

Figure 3
1 2 3 4 5 6 7 8 9 10 11
with expletive negation 0 0 0 0 0 0 0 0 0 0 0
without expletive negation 0 0 3 9 13 13 13 10 12 13 2

12 13 14 15 16 17 18 19
0 0 0 0 0 0 0 0
7 4 1 4 0 0 15 4
Figure 4
1 2 3 4 5 6 7 8 9 10 11-19
with expletive negation 1 4 1 4 1 3 3 1 0 0 0
without expletive negation 0 0 0 0 0 0 0 0 2 1 0

Figure 5
centuries 14th 15th 16th 17th 18th 19th 20th
to-infinitives 3 8 75 148 119 362 309
gerunds 1 2 3 1 8 4 0
200 Appendix

Figure 6
centuries 13th 14th 15th 16th 17th 18th 19th 20th
that-clauses 2 0 1 0 0 0 0 0
to-infinitives 1 2 2 20 68 12 12 0
gerunds 0 0 0 5 26 33 9 1

Figure 7
centuries 13th 14th 15th 16th 17th 18th 19th 20th
that-clauses 2 0 0 0 0 0 0 0
to-infinitives 0 1 0 4 15 3 2 0
gerunds 0 0 0 0 4 25 4 0

Figure 8
centuries 16th 17th 18th 19th 20th
gerunds only 4 24 32 6 0
from + gerunds 1 2 1 3 1

Figure 9
centuries 16th 17th 18th 19th 20th
that-clauses 9.1 3.0 0 0 0
to-infinitives 27.3 12.1 0 2.0 0
gerunds 63.6 84.8 100 98.0 100

Figure 10
centuries 13th 14th 15th 16th 17th 18th 19th 20th
forbear (all types) 5.6 3.0 4.5 3.6 5.1 2.6 0.7 0.1
avoid (all types) 0.2 1.2 3.6 9.5 8.7 9.4 7.5 17.6

Figure 11
centuries 16th 17th 18th 19th 20th
that-clauses 6 1 1 0 0
to-infinitives 10 11 6 11 0
gerunds 2 12 14 34 11

Figure 12
centuries 16th 17th 18th 19th 20th
gerunds only 2 11 8 13 2
from + gerunds 0 1 6 21 9

Figure 13
centuries 16th 17th 18th 19th 20th
that-clauses 60.0 5.8 0.3 0.1 0
to-infinitives 10.0 7.2 0.7 0 0
gerunds 30.0 87.0 99.0 99.9 100
Appendix 201

Figure 14
centuries 17th 18th 19th 20th
prevent + possessive + –ing 6 74 122 12
prevent + object + –ing 1 1 20 18
prevent + object + from + –ing 2 35 91 50

Figure 15
centuries 17th 18th 19th 20th
prevent + possessive + –ing 4 4 3 1
prevent + object + –ing 9 27 130 82
prevent + object + from + –ing 8 106 278 177

Figure 16
centuries 17th 18th 19th 20th
prevent + possessive + –ing 66.7 67.3 52.4 15.0
prevent + object + –ing 11.1 0.9 8.6 22.5
prevent + object + from + –ing 22.2 31.8 39.1 62.5

Figure 17
centuries 15th 16th 17th 18th 19th 20th
that-clauses 0 2 2 0 2 0
but-clauses 0 0 3 1 0 0
to-infinitives 1 5 15 9 2 0
gerunds 0 10 92 97 40 4

Figure 18
centuries 15th 16th 17th 18th 19th 20th
gerunds only 0 6 34 26 10 0
from + gerunds 0 2 58 69 30 4

Figure 19
centuries 16th 17th 18th 19th 20th
that-clauses 0 3 1 2 0
to-infinitives 2 2 5 2 0
gerunds 3 30 39 14 3

Figure 20
centuries 16th 17th 18th 19th 20th
that-clauses 2 2 0 0 0
to-infinitives 3 10 3 0 0
gerunds 4 32 45 19 1

Figure 21
centuries 16th 17th 18th 19th 20th
gerunds only 3 7 16 3 0
from + gerunds 0 23 23 11 3
202 Appendix

Figure 22
centuries 16th 17th 18th 19th 20th
gerunds only 0 4 2 2 0
from + gerunds 4 28 43 17 1

Figure 23
centuries 16th 17th 18th 19th 20th
hinder + possessive + –ing 3 7 16 1 0
hinder + object + –ing 0 0 0 2 0
hinder + object + from + –ing 0 23 23 11 3

Figure 24
centuries 16th 17th 18th 19th 20th
hinder + possessive + –ing 0 0 1 0 0
hinder + object + –ing 0 4 1 2 0
hinder + object + from + –ing 4 28 43 17 1

Figure 25
centuries 15th 16th 17th 18th 19th 20th
that-clauses 1 2 0 0 0 0
to-infinitives 1 5 2 1 0 0
gerunds 1 19 10 9 24 22

Figure 26
centuries 15th 16th 17th 18th 19th 20th
gerunds only 0 6 6 3 0 0
from + gerunds 1 13 4 6 24 22

Figure 27
centuries 13th 14th 15th 16th 17th 18th 19th 20th
that-clauses 0 0 6 115 195 106 214 93
to-infinitives 1 0 1 28 53 17 44 7
gerunds 0 0 0 11 14 4 9 6

Figure 28
centuries 13th 14th 15th 16th 17th 18th 19th 20th
that-clauses 0 0 6 113 187 106 212 90
to-infinitives 1 0 1 21 43 13 40 7
gerunds 0 0 0 9 10 1 5 5

Figure 29
centuries 13th 14th 15th 16th 17th 18th 19th 20th
that-clauses 0 0 0 2 8 0 2 3
to-infinitives 0 0 0 7 10 4 4 0
gerunds 0 0 0 2 4 3 4 1
Appendix 203

Figure 30
centuries 13th 14th 15th 16th 17th 18th 19th 20th
whether-clauses 0 40.0 0 26.1 25.7 40.4 55.0 32.8
if-clauses 0 0 0 2.2 5.4 4.3 18.3 44.3

Figure 31
centuries 13th 14th 15th 16th 17th 18th 19th 20th
no doubt (all examples) 2.2 5.1 2.3 18.5 26.0 30.2 44.8 54.0
no doubt (without that) 2.2 2.6 0 12.9 21.3 23.6 37.8 41.6

Figure 32
centuries 14th 15th 16th 17th 18th 19th 20th
that-clauses 4 4 24 51 23 84 77
but-clauses 0 0 11 20 2 0 0
to-infinitives 2 5 39 39 14 9 0
gerunds 0 0 2 3 4 3 7

Figure 33
Centuries 14th 15th 16th 17th 18th 19th 20th
that expressed 4 4 24 39 21 76 70
that unexpressed 0 0 0 12 2 8 7

Figure 34
centuries 14th 15th 16th 17th 18th 19th 20th
that expressed 3 4 20 29 15 50 56
that unexpressed 0 0 0 10 1 3 4

Figure 35
centuries 14th 15th 16th 17th 18th 19th 20th
1st person 25.0 0 20.8 21.6 17.4 15.5 3.9

Figure 36
centuries 14th 15th 16th 17th 18th 19th 20th
1st person 33.3 0 20.0 17.9 18.8 11.3 3.3

Figure 37
centuries 14th 15th 16th 17th 18th 19th 20th
simple infinitives 1 2 21 16 5 0 0
object + infinitives 1 3 18 23 9 9 0

Figure 38
centuries 14th 15th 16th 17th 18th 19th 20th
simple infinitives 0 0 1 0 1 2 6
object + infinitives 0 0 1 3 3 1 1
References

Aijmer, Karin. 1997. “I think—An English Modal Particle”. Modality in the


Germanic Languages: Historical and Comparative Perspectives ed. by
Toril Swan and Olaf Jansen Westvik, 1–47. Berlin: Mouton de Gruyter.
Allerton, David J. 1988. “‘Infinitives’ in English”. Studies in Descriptive Lin-
guistics 18.11–23.
Allerton, David J. 1991. “The Greater Precision of Spoken Language: Four
Examples from English”. English Studies 72.470–478.
The American Heritage Book of English Usage. 1996. Boston: Houghton
Mifflin.
Andersson, Evert. 1985. On Verb Complementation in Written English. Lund
Studies in English, 71. Malmo: C. W. K. Gleerup.
Aston, Guy and Lou Burnard. 1998. The BNC Handbook. Edinburgh: Edin-
burgh University Press.
Baghdikian, Sonia. 1979. “Ne in ME and EModE”. English Studies 60.673–
679.
Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad, and Edward
Finegan. 1999. Longman Grammar of Spoken and Written English. Har-
low: Longman.
Bladon, R. A. W. 1968. “Selecting the to– or –ing Nominal after like, love, hate,
dislike and prefer”. English Studies 49.203–214.
Blake, Norman F. 1988. “Negation in Shakespeare”. An Historic Tongue:
Studies in English Linguistics in Memory of Barbara Strang ed. by
Graham Nixon and John Honey, 89–111. London: Routledge.
Blake, Norman F. 1996. A History of the English Language. Basingstoke: Mac-
millan.
Bolinger, Dwight. 1979. “The Jingle Theory of Double –ing”. Function and
Context in Linguistic Analysis: A Festschrift for William Haas ed. by
David J. Allerton, Edward Carney, and David Holdcroft, 41–56. Cam-
bridge: Cambridge University Press.
206 References

Brewer, Charlotte. 2000. “OED Sources”. Lexicography and the OED:


Pioneers in the Untrodden Forest ed. by Lynda Mugglestone, 40–58. Ox-
ford: Oxford University Press.
Brinton, Laurel J. 1996. Pragmatic Markers in English: Grammaticalization
and Discourse Functions. Topics in English Linguistics, 19. Berlin: Mou-
ton de Gruyter.
Brinton, Laurel J. 2001. “From Matrix Clause to Pragmatic Marker: The
History of look-forms”. Journal of Historical Pragmatics 2.177–199.
Brinton, Laurel J. 2008. The Comment Clause in English: Syntactic Origins
and Pragmatic Development. Studies in English Language. Cambridge:
Cambridge University Press.
Burchfield, Robert W. 1998. The New Fowler’s Modern English Usage.
Revised 3rd edition. Oxford: Clarendon Press.
Cambridge Advanced Learner’s Dictionary. 2008. 3rd edition. Cambridge:
Cambridge University Press.
Curme, George O. 1931. Syntax. Boston: D. C. Heath.
Denison, David. 1998. “Syntax”. The Cambridge History of the English Lan-
guage, IV: 1776–1997 ed. by Suzanne Romaine, 92–329. Cambridge:
Cambridge University Press.
Dekeyser, Xavier. 1975. Number and Case Relations in 19th Century British
English: A Comparative Study of Grammar and Usage. Antwerpen:
Uitgeverij de Nederlandsche Boekhandel.
De Smet, Hendrik. 2004. “Semantics and Variation in Complement Con-
structions: Gerunds and Infinitives Following the Verb like”. Belgian
Journal of English Language and Literature 2.247–260.
De Smet, Hendrik. 2007. “For … to-infinitives as Verbal Complements in Late
Modern and Present-day English: Between Motivation and Change”.
English Studies 88.67–94.
De Smet, Hendrik. 2008. “Functional Motivations in the Development of
Nominal and Verbal Gerunds in Middle and Early Modern English”.
English Language and Linguistics 12.55–102.
De Smet, Hendrik and Hubert Cuyckens. 2005. “Pragmatic Strengthening and
the Meaning of Complement Constructions: The Case of like and love
with the to-infinitive”. Journal of English Linguistics 33.3–34.
De Smet, Hendrik and Hubert Cuyckens. 2007. “Diachronic Aspects of Com-
plementation: Constructions, Entrenchment, and the Matching Problem”.
Studies in the History of the English Language, III: Managing Chaos,
Strategies for Identifying Change in English ed. by Christopher M. Cain
and Geoffrey Russom, 187–213. Berlin: Mouton de Gruyter.
References 207

Dixon, Robert M. W. 1991. A New Approach to English Grammar, on Semantic


Principles. Oxford: Clarendon Press.
Dixon, Robert M. W. 1995. “Complement Clauses and Complementation
Strategies”. Grammar and Meaning: Essays in Honour of Sir John Lyons
ed. by Frank Robert Palmer, 175–220. Cambridge: Cambridge University
Press.
Dixon, Robert M. W. 2005. A Semantic Approach to English Grammar. 2nd
edition. Oxford: Oxford University Press.
Dixon, Robert M. W. 2006. “Complement Clauses and Complementation
Strategies in Typological Perspective”. Complementation: A Cross-ling-
uistic Typology ed. by Robert M. W. Dixon and Alexandra Y. Aikhenvald,
1–48. Oxford: Oxford University Press.
Duffley, Patrick J. 1995. “Defining the Potential Meaning of the English –ing
Form in a Psychomechanical Approach”. Langues et Linguistique 21.1–
11.
Duffley, Patrick J. 1999. “The Use of the Infinitive and the –ing after Verbs
Denoting the Beginning, Middle and End of an Event”. Folia Linguistica
33.295–331.
Duffley, Patrick J. 2000. “Gerund versus Infinitive as Complement of Transi-
tive Verbs in English: The Problems of ‘Tense’ and ‘Control’”. Journal of
English Linguistics 28.221–248.
Duffley, Patrick J. 2003. “The Gerund and the to-infinitive as Subject”. Journal
of English Linguistics 31.324–352.
Duffley, Patrick J. 2006. The English Gerund-Participle: A Comparison with
the Infinitive. Berkeley Insights in Linguistics and Semiotics, 61. New
York: Peter Lang.
Duffley, Patrick J. and Jean-François Joubert. 1999. “The Gerund and the
Infinitive with the Verbs intend, mean, propose and their Close Syn-
onyms”. Canadian Journal of Linguistics 44.251–266.
Duffley, Patrick J. and Rachel Tremblay. 1994. “The Infinitive and the –ing as
Complements of Verbs of Effort”. English Studies 75.566–575.
Egan, Thomas. 2008. Non-finite Complementation: A Usage-based Study of
Infinitive and –ing Clauses in English. Language and Computers: Studies
in Practical Linguistics, 65. Amsterdam: Rodopi.
Ellegård, Alver. 1953. The Auxiliary do: The Establishment and Regulation of
its Use in English. Gothenburg Studies in English, 2. Stockholm:
Almqvist and Wiksell.
Espinal, M. Teresa. 1991. “The Representation of Disjunct Constituents”. Lan-
guage 67.726–762.
208 References

Fanego, Teresa. 1990. “Finite Complement Clauses in Shakespeare’s English, I


and II”. Studia Neophilologica 62.3–21, 129–149.
Fanego, Teresa. 1996a. “On the Historical Development of English Retrospec-
tive Verbs”. Neuphilologische Mitteilungen 97.71–79.
Fanego, Teresa. 1996b. “The Development of Gerunds as Objects of Subject-
Control Verbs in English (1400–1760)”. Diachronica 13.29–62.
Fanego, Teresa. 1997. “On Patterns of Complementation with Verbs of Effort”.
English Studies 78.60–67.
Fanego, Teresa. 2004a. “Some Strategies for Coding Sentential Subjects in
English: From Exaptation to Grammaticalization”. Studies in Language
28.321–361.
Fanego, Teresa. 2004b. “On Reanalysis and Actualization in Syntactic Change:
The Rise and Development of English Verbal Gerunds”. Diachronica 21.
5–55.
Fischer, Andreas. 1997. “The Oxford English Dictionary on CD-ROM as a
Historical Corpus: To wed and to marry Revisited”. From Ælfric to the
New York Times: Studies in English Corpus Linguistics ed. by Udo Fries,
Viviane Müller and Peter Schneider, 161–172. Amsterdam: Rodopi.
Fischer, Olga. 1992. “Syntax”. The Cambridge History of the English Lan-
guage, II: 1066–1476 ed. by Norman Francis Blake, 207–408. Cam-
bridge: Cambridge University Press.
Fischer, Olga. 2003. “The Development of the Modals in English: Radical
Versus Gradual Changes”. English Modality in Context: Diachronic Per-
spectives ed. by David Hart, 19–32. Bern: Peter Lang.
Fischer, Olga, Ans van Kemenade, Willem Koopman, and Wim van der Wurff.
2000. The Syntax of Early English. Cambridge: Cambridge University
Press.
Fischer, Olga and Wim van der Wurff. 2006. “Syntax”. A History of the English
Language ed. by Richard Hogg and David Denison, 109–198. Cam-
bridge: Cambridge University Press.
Fowler, Henry W. 1926. A Dictionary of Modern English Usage. Oxford:
Clarendon Press.
Garner, Bryan A. 2003. Garner’s Modern American Usage. 2nd edition.
Oxford: Oxford University Press.
Givón, Talmy. 1991. “Isomorphism in the Grammatical Code: Cognitive and
Biological Considerations”. Studies in Language 15.85–114.
Görlach, Manfred. 1987. “Colonial Lag?: The Alleged Conservative Character
of American English and Other ‘Colonial’ Varieties”. English World-Wide
8.41–60.
References 209

Gramley, Stephan. 1980. “Infinitive and Gerund Complements with the Verbs
‘Begin’ and ‘Start’”. Arbeiten aus Anglistik und Amerikanistik 5.159–186.
Greenbaum, Sidney. 1988. “Syntactic Devices for Compression in English”.
Essays on the English Language and Applied Linguistics on the Occasion
of Gerhard Nickel’s 60th Birthday ed. by Josef Klegraf and Dietrich
Nehls, 3–10. Heidelberg: Julius Groos.
Greenbaum, Sidney and Janet Whitcut. 1996. Longman Guide to English Us-
age. London: Penguin.
Haegeman, Liliane. 1980. “TRY and the English Catenatives”. Linguistics
18.1095–1097.
Hamawand, Zeki. 2002. Atemporal Complement Clauses in English: A Cogni-
tive Grammar Analysis. München: Lincom Europa.
Hamawand, Zeki. 2004. “Determinants of Complement Clause Variation in
English”. English Studies 85.451–464.
Haspelmath, Martin. 1989. “From Purposive to Infinitive: A Universal Path of
Grammaticalization”. Folia Linguistica Historica 10.287–310.
Herbst, Thomas, David Heath, Ian F. Roe, and Dieter Götz. 2004. A Valency
Dictionary of English: A Corpus-Based Analysis of the Complementation
Patterns of English Verbs, Nouns and Adjectives. Topics in English
Linguistics, 40. With the assistance of Michael Klotz. Berlin: Mouton de
Gruyter.
Heyvaert, Liesbet, Hella Rogiers and Nadine Vermeylen. 2005. “Pronominal
Determiners in Gerundive Nominalization: A ‘Case’ Study”. English
Studies 86.71–88.
Hoffmann, Sebastian. 2004. “Using the OED Quotations Database as a Corpus:
A Linguistic Appraisal”. ICAME Journal 28.17–30.
Hopper, Paul J. and Elizabeth Closs Traugott. 2003. Grammaticalization. Cam-
bridge Textbooks in Linguistics. 2nd edition. Cambridge: Cambridge Uni-
versity Press.
Houston, Ann. 1989. “The English Gerund: Syntactic Change and Discourse
Function”. Language Change and Variation ed. by Ralph W. Fasold and
Deborah Schiffrin, 173–196. Amsterdam: John Benjamins.
Huddleston, Rodney and Geoffrey K. Pullum. 2002. The Cambridge Grammar
of the English Language. Cambridge: Cambridge University Press.
Ishiguro, Taro. 1994. “A Study of the Old English forbeodan and its Negative
Dependent Clause”. Unpublished MA Dissertation submitted to Inter-
national Christian University, Tokyo.
210 References

Ishiguro, Taro. 1998. “Verbs of Negative Import: A Syntactic Study that


Benefited from the Dictionary of Old English Project”. Old English
Newsletter 26.23–32.
Iyeiri, Yoko. 1998. “Multiple Negation in Middle English Verse”. Tieken,
Tottie, and Wurff (eds.) (1998), 121–146.
Iyeiri, Yoko. 2001. Negative Constructions in Middle English. Fukuoka: Kyu-
shu University Press.
Iyeiri, Yoko. 2002a. “Prohibit ni kakawaru Kobun no Rekishiteki Hensen
nitsuite” (History of the English Verb to prohibit). Eigoshi Kenkyukai
Kaiho (The Bulletin of the Japanese Association for Studies in the History
of the English Language) 8.1–4.
Iyeiri, Yoko. 2002b. “Development of any from Middle English to Early
Modern English: A Study Using The Helsinki Corpus of English Texts”.
English Corpus Linguistics in Japan ed. by Toshio Saito, Junsaku Naka-
mura, and Shunji Yamazaki, 211–223. Amsterdam: Rodopi.
Iyeiri, Yoko. 2002c. “The Development of Non-assertive any in Later Middle
English”. And gladly wolde he lerne and gladly teche: Essays on Medie-
val English Presented to Professor Matsuji Tajima on his Sixtieth Birth-
day ed. by Yoko Iyeiri and Margaret Connolly, 127–143. Tokyo: Kaibun-
sha.
Iyeiri, Yoko. 2003. “‘God forbid!’: A Historical Study of the Verb forbid in
Different Versions of the English Bible”. Journal of English Linguistics
31.149–162.
Iyeiri, Yoko. 2004. “The Use of the Auxiliary do in Negation in Tom Jones and
Some Other Literary Works of the Contemporary Period”. New Trends in
English Historical Linguistics: An Atlantic View ed. by Isabel Mosko-
wich-Spiegel Fandiño and Begoña Crespo García, 223–240. Coruña: Uni-
versidade da Coruña.
Iyeiri, Yoko. 2008a. “Doubt nikakawaru Kobun no Rekishiteki Henka nitsuite:
The Oxford English Dictionary no Inyobundeta no Bunseki kara” (The
Verb doubt and its Changing Patterns of Complementation in the History
of English: An Analysis of the OED Data). Kyudai Eibungaku (English
Studies at Kyushu University) 50.317–338.
Iyeiri, Yoko. 2008b. “Changing Patterns of Complementation of the Verb deny
in the History of English”. Paper presented at the 29th ICAME con-
ference at Ascona in Switzerland. 18 May 2008.
Iyeiri, Yoko. 2008c. “The Verb prevent and its Changing Patterns of Com-
plementation in the History of English”. Historical Englishes in Varieties
of Texts and Contexts: The Global COE Programme, International
References 211

Conference 2007 ed. by Masachiyo Amano, Michiko Ogura, and Masa-


yuki Ohkado, 297–309. Frankfurt am Main: Peter Lang.
Iyeiri, Yoko. 2009a. “I fear: The Historical Development of the Verb fear and
its Changing Patterns of Complementation”. Studies in Modern English
25.19–39.
Iyeiri, Yoko. 2009b. “The Historical Development of the Verb doubt and its
Various Patterns of Complementation”. Exploring the Lexis-Grammar
Interface ed. by Ute Römer and Rainer Schulze, 153–169. Amsterdam:
John Benjamins.
Jack, George B. 1978a. “Negative Adverbs in Early Middle English”. English
Studies 59.295–309.
Jack, George B. 1978b. “Negative Concord in Early Middle English”. Studia
Neophilologica 50.29–39.
Jack, George B. 1978c. “Negation in Later Middle English Prose”. Archivum
Linguisticum n.s. 9.58–72.
Jespersen, Otto. 1909–1949. A Modern English Grammar on Historical Prin-
ciples. 7 vols. London: George Allen & Unwin.
Jespersen, Otto. 1917. Negation in English and Other Languages. Copenhagen:
Andr. Fred. Høst and Søn, Kgl. Hof-Boghandel.
Jørgensen, Erik. 1990. “Remember and forget with Gerund and Infinitive as
Objects”. English Studies 71.147–151.
Jucker, Andreas H. 1994. “New Dimensions in Vocabulary Studies: Review
Article of the Oxford English Dictionary (2nd edition) on CD-ROM”. Lit-
erary and Linguistic Computing 9.149–154.
Kaltenböck, Gunther. 2005. “Charting the Boundaries of Syntax: A Taxonomy
of Spoken Parenthetical Clauses”. View[z]: Vienna Working Papers 14
(1).21–53. <http://www.univie.ac.at/Anglistik/Views0501ALL.pdf>
Kent, Charles W. 1890. “Of the Use of the Negative by Chaucer, with Par-
ticular Reference to the Particle ne”. PMLA 5.109–147.
Kikusawa, Namiko. 2009. “The Subjunctive vs. Modal Auxiliaries: Lest-
clauses in Late Middle English Prose Texts”. Paper presented at MMECL
(Middle and Modern English Corpus Linguistics), 7 July 2009, University
of Innsbruck.
Kiparsky, Paul and Carol Kiparsky. 1970. “Fact”. Progress in Linguistics: A
Collection of Papers ed. by Manfred Bierwisch and Karl Erich Heidolph,
143–173. The Hague: Mouton.
Kirsten, Hans. 1957. “Bemerkungen zu to prevent + Gerundium”. Zeitschrift
für Anglistik und Amerikanistik 1.327–328.
212 References

Kurath, Hans, Sherman M. Kuhn, and Robert E. Lewis (eds.). 1952–2001.


Middle English Dictionary. Ann Arbor: University of Michigan Press.
Kytö, Merja. 1993. Manual to the Diachronic Part of The Helsinki Corpus of
English Texts: Coding Conventions and Lists of Source Texts. 2nd edition.
Helsinki: Department of English, University of Helsinki.
Langacker, Ronald W. 1992. “Prepositions as Grammatical(izing) Elements”.
Leuvense Bijdragen 81.287–309.
Lange, Claudia. 2007. Reflexivity and Intensification in English: A Study of
Texts and Contexts. Arbeiten zur Sprachanalyse, 46. Frankfurt am Main:
Peter Lang.
Leech, Geoffrey. 2004. Meaning and the English Verb. 3rd edition. Harlow:
Pearson.
Lehmann, Christian. 1988. “Towards a Typology of Clause Linkage”. Clause
Combining in Grammar and Discourse ed. by John Haiman and Sandra A.
Thompson, 181–225. Amsterdam: John Benjamins.
Longman Dictionary of Contemporary English. 2003. New edition. Harlow:
Pearson Education.
López-Couso, María José and Belén Méndez-Naya. 1998. “On Minor Declara-
tive Complementizers in the History of English: The Case of but”. Ad-
vances in English Historical Linguistics (1996) ed. by Jacek Fisiak and
Marcin Krygier, 161–171. Berlin: Mouton de Gruyter.
Los, Bettelou. 1998. “Bare and to-infinitives in Old English: Callaway Re-
visited”. Advances in Historical Linguistics (1996) ed. by Jacek Fisiak
and Marcin Krygier, 173–188. Berlin: Mouton de Gruyter.
Los, Bettelou. 2005. The Rise of the to-Infinitive. Oxford: Oxford University
Press.
Los, Bettelou. 2007. “To as a Connective in the History of English”. Connec-
tives in the History of English ed. by Ursula Lenker and Anneli Meurman-
Solin, 31–60. Amsterdam: John Benjamins.
Lothian, John Maule and Thomas Wallace Craik (eds.). 1975. Twelfth Night.
The Arden Shakespeare. London: Methuen and Co. Ltd.
Machida, Naoko. 1979. “Negation in the Two Early Middle English Romances,
King Horn and Havelok the Dane”. Journal of the Faculty of Foreign
Languages (Komazawa University, Tokyo) 8.1–31.
Mackenzie, Eneas. 1825. An Historical, Topographical, and Descriptive View
of the County of Northumberland, and of Those Parts of the County of
Durham Situated North of the River Tyne, with Berwick upon Tweed, and
Brief Notices of Celebrated Places on the Scottish Border, vol. 1. 2nd
edition. Newcastle upon Tyne: Mackenzie and Dent.
References 213

Macmillan English Dictionary for Advanced Learners. 2007. 2nd edition. Ox-
ford: Macmillan Education.
Mair, Christian. 1998. “Corpora and the Study of the Major Varieties of
English: Issues and Results”. The Major Varieties of English: Papers from
MAVEN 97 ed. by Hans Lindquist, Staffan Klintborg, Magnus Levin, and
Maria Estling, 139–157. Växjö: Acta Wexionensia.
Mair, Christian. 1999. “Un-American Activities in British English Syntax:
prevent, stop, save”, in From, Function and Variation in English: Studies
in Honour of Klaus Hansen ed. by Uwe Carls and Peter Lucko, 175–181.
Frankfurt am Main: Peter Lang.
Mair, Christian. 2001. “Early or Late Origin for begin + V-ing?: Using the
OED on CD-ROM to Settle a Dispute between Visser and Jespersen”.
Anglia 119.606–610.
Mair, Christian. 2002. “Three Changing Patterns of Verb Complementation in
Late Modern English: A Real-time Study Based on Matching Text Cor-
pora”. English Language and Linguistics 6.105–131.
Mair, Christian. 2003. “Gerundial Complements after begin and start: Gram-
matical and Sociolinguistic Factors, and How They Work against Each
Other”. Determinants of Grammatical Variation in English ed. by Günter
Rohdenburg and Britta Mondorf, 329–345. Berlin: Mouton de Gruyter.
Mair, Christian. 2004. “Corpus Linguistics and Grammaticalisation Theory:
Statistics, Frequencies, and Beyond”. Corpus Approaches to Gram-
maticalization in English ed. by Hans Lindquist and Christian Mair, 121–
150. Amsterdam: John Benjamins.
Mair, Christian. 2006. “Nonfinite Complement Clauses in the Nineteenth Cen-
tury: The Case of remember”. Nineteenth-century English: Stability and
Change ed. by Merja Kytö, Mats Rydén, and Erik Smitterberg, 215–228.
Cambridge: Cambridge University Press.
Manabe, Kazumi. 1979. Syntax and Style in Early English: Finite and Non-
finite Clauses c900–1600. Tokyo: Kaibunsha.
Martinez, Ignacio M. Palacios. 2003. “Multiple Negation in Modern English: A
Preliminary Corpus-based Study”. Neuphilologische Mitteilungen 104.
477–498.
Middle English Dictionary. See Kurath, Kuhn, and Lewis (1952–2001).
Milsark, Gary. 1972. “Re: Doubl-ing”. Linguistic Inquiry 3.542–549.
Molencki, Rafał. 1999. “A History of the English Perfect Infinitive”. Studia
Anglica Posnaniensia 34.91–121.
214 References

Murray, K. M. Elisabeth. 1977. Caught in the Web of Words: James A. H.


Murray and the Oxford English Dictionary. With a Preface by Robert W.
Burchfield. New Haven: Yale University Press.
Nevalainen, Terttu. 2006. An Introduction to Early Modern English. Edinburgh
Textbooks on the English Language. Edinburgh: Edinburgh University
Press.
Nilsen, Don L. F. 1968. “English Infinitives”. Canadian Journal of Linguistics
13.83–93.
Noël, Dirk. 2003. “Is There Semantics in All Syntax?: The Case of Accusative
and Infinitive Constructions vs. that-clauses”. Determinants of Gram-
matical Variation in English ed. by Günter Rohdenburg and Britta
Mondorf, 347–377. Berlin: Mouton de Gruyter.
Noonan, Michael. 1985. “Complementation”. Language Typology and Syn-
tactic Description, vol. 2 ed. by Timothy Shopen, 42–140. Cambridge:
Cambridge University Press.
Owen, Charles. 1993. “Corpus-based Grammar and the Heineken Effect:
Lexico-grammatical Description for Language Learners”. Applied Lin-
guistics 14.167–187.
Oxford Advanced Learner’s Dictionary of Current English. 2005. 7th edition.
Oxford: Oxford University Press.
The Oxford English Dictionary. See Simpson and Weiner (1989).
Palmer, Frank R. 1988. The English Verb. Longman Linguistics Library. 2nd
edition. London: Longman.
Partridge, Eric. 1994. Usage and Abusage: A Guide to Good English. New
edition. Revised by J. Whitcut. London: Penguin Books.
Potter, Simeon. 1975. Changing English. The Language Library. 2nd revised
edition. London: André Deutsch.
Pullum, Geoffrey K. 1974. “Restating Doubl-ing”. Glossa 8.109–120.
Quirk, Randolph. 1974. The Linguist and the English Language. London:
Edward Arnold.
Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech, and Jan Svartvik. 1985.
A Comprehensive Grammar of the English Language. London: Longman.
Řeřicha, Václav. 1987. “Notes on Infinitival and –ing Participle Complements
of the Verbs begin and start”. Philologica Pragensia 30.129–132.
Rissanen, Matti. 1991. “On the History of that/zero as Object Clause Links in
English”. English Corpus Linguistics: Studies in Honour of Jan Svartvik
ed. by Karin Aijmer and Bengt Altenberg, 272–289. London: Longman.
References 215

Rissanen, Matti. 1999. “Syntax”. The Cambridge History of the English Lan-
guage, III: 1476–1776 ed. by Roger Lass, 187–331. Cambridge: Cam-
bridge University Press.
Rohdenburg, Günter. 1995. “On the Replacement of Finite Complement
Clauses by Infinitives in English”. English Studies 76.367–388.
Rohdenburg, Günter. 1996. “Cognitive Complexity and Increased Grammatical
Explicitness in English”. Cognitive Linguistics 7.149–182.
Rohdenburg, Günter. 2003. “Cognitive Complexity and horror aequi as Factors
Determining the Use of Interrogative Clause Linkers in English”. Deter-
minants of Grammatical Variation in English ed. by Günter Rohdenburg
and Britta Mondorf, 205–249. Berlin: Mouton de Gruyter.
Rohdenburg, Günter. 2006. “The Role of Functinal Constraints in the Evolu-
tion of the English Complementation System”. Syntax, Style and Gram-
matical Norms: English from 1500–2000 ed. by Christiane Dalton-Puffer,
Dieter Kastovsky, Nikolaus Ritt, and Herbert Schendl, 143–166. Bern:
Peter Lang.
Ross, John Robert. 1972. “Doubl-ing”. Linguistic Inquiry 3.61–86.
Ross, John Robert. 1974. “Three Batons for Cognitive Psychology”. Cognition
and the Symbolic Processes ed. by Walter B. Weimer and David S.
Palermo, 63–124. New Jersey: Lawrence Erlbaum.
Rudanko, Juhani. 1984. “On Some Contrasts between Infinitival and that
Complement Clauses in English”. English Studies 65.141–161.
Rudanko, Juhani. 1989. Complementation and Case Grammar: A Syntactic and
Semantic Study of Selected Patterns of Complementation in Present-day
English. SUNY Series in Linguistics. Albany, NY: State University of
New York Press.
Rudanko, Juhani. 1998. Change and Continuity in the English Language:
Studies on Complementation over the Past Three Hundred Years. Lan-
ham: University Press of America.
Schäfer, Jürgen. 1980. Documentation in the O.E.D.: Shakespeare and Nashe
as Test Cases. Oxford: Clarendon Press.
Shorter Oxford English Dictionary on Historical Principles. 2007. 6th edition.
Oxford: Oxford University Press.
Simpson, John A. and Edmund S. C. Weiner, prepared. 1989. The Oxford Eng-
lish Dictionary, 2nd edition on CD-ROM Version 3.2 (2002). Oxford:
Clarendon Press.
Skandera, Paul. 2003. “Start doing or start to do: Is the Gerund Spreading in
American English?”. English Core Linguistics: Essays in Honour of D. J.
Allerton ed. by Cornelia Tschichold, 343–352. Bern: Peter Lang.
216 References

Smithers, G. V. (ed.) 1987. Havelok. Oxford: Clarendon Press.


Söderlind, Johannes. 1958. Verb Syntax in John Dryden’s Prose, Part II. Essays
and Studies on English Language and Literature, 19. Uppsala: A. -B.
Lundequistska Bokhandeln.
Strang, Barbara M. H. 1970. A History of English. London: Methuen.
Swan, Michael. 1995. Practical English Usage. 2nd edition. Oxford: Oxford
University Press.
Swan, Michael. 2005. Practical English Usage. 3rd edition. Oxford: Oxford
University Press.
Sweet, Henry. 1891–1898. A New English Grammar: Logical and Historical. 2
vols. Oxford: Clarendon Press.
Tagliamonte, Sali and Jennifer Smith. 2005. “No momentary fancy!”: The Zero
‘Complementizer’ in English Dialects”. English Language and Lin-
guistics 9.289–309.
Tajima, Matsuji. 1985. The Syntactic Development of the Gerund in Middle
English. Tokyo: Nan’un-do.
Tajima, Matsuji (ed.). 1995. Computer Corpus Riyo niyoru Gendai Eibei Go-
hou Kenkyu (Present-day British and American English Usage: A Cor-
pus-based Study). Tokyo: Kaibunsha.
Tieken-Boon van Ostade, Ingrid. 1987. The Auxiliary do in Eighteenth-century
English: A Sociohistorical-linguistic Approach. Geschiedenis van de
Taalkunde, 6. Dordrecht: Foris Publications.
Tieken-Boon van Ostade, Ingrid. 1989. “Dr Johnson and the Auxiliary do”.
Folia Linguistica Historica 10.145–162.
Tieken-Boon van Ostade, Ingrid, Gunnel Tottie, and Wim van der Wurff (eds.).
1998. Negation in the History of English. Topics in English Linguistics,
26. Berlin: Mouton de Gruyter.
Thompson, Sandra A. 2002. “‘Object Complements’ and Conversation towards
a Realistic Account”. Studies in Language 26.125–164.
Thompson, Sandra A. and Anthony Mulac. 1991a. “The Discourse Conditions
for the Use of the Complementizer that in Conversational English”.
Journal of Pragmatics 15.237–251.
Thompson, Sandra A. and Anthony Mulac. 1991b. “A Quantitative Perspective
on the Grammaticalization of Epistemic Parentheticals in English”. Ap-
proaches to Grammaticalization, II: Focus on Types of Grammatical
Markers ed. by Elizabeth Closs Traugott and Bernd Heine, 313–329.
Amsterdam: John Benjamins.
Tottie, Gunnel. 1991. “Lexical Diffusion in Syntactic Change: Frequency as a
Determinant of Linguistic Conservatism in the Development of Negation
References 217

in English”. Historical English Syntax ed. by Dieter Kastovsky, 439–467.


Berlin: Mouton de Gruyter.
Traugott, Elizabeth Closs. 1992. “Syntax”. The Cambridge History of the
English Language, I: The Beginning to 1066 ed. by Richard M. Hogg,
168–289. Cambrige: Cambridge University Press.
Tregidgo, P. S. 1980. “Some Observations on Verg + ing and Verb + Infinitive”.
English Language Teaching Journal 35.45–48.
Urmson, James O. 1952. “Parenthetical Verbs”. Mind: A Quarterly Review of
Psychology and Philosophy 61.480–496.
Urmson, James O. 1963. “Parenthetical Verbs”. Philosophy and Ordinary Lan-
guage ed. by Charles E. Caton, 220–240. Urbana: University of Illinois
Press.
van der Wurff, Wim. 1998. “On Expletive Negation with Adversative Pred-
icates in the History of English”. Tieken, Tottie, and Wurff (eds.) (1998),
295–327.
van Ek, Jan A. 1966. Four Complementary Structures of Predication in Con-
temporary British English: An Inventory. Groningen Studies in English, 9.
Groningen: J. B. Wolters.
van Gelderen, Elly. 2004. Grammaticalization as Economy. Linguistics Today,
71. Amsterdam: John Benjamins.
van Gelderen, Elly. 2006. A History of the English Language. Amsterdam: John
Benjamins.
Vespoor, Marjolijn. 1999. “To infinitives”. Issues in Cognitive Linguistics:
1993 Proceedings of the International Cognitive Linguistics Conference
ed. by Leon De Stadler and Christoph Eyrich, 505–526. Berlin: Mouton
de Gruyter.
Vespoor, Marjolijn. 2000. “Iconicity in English Complement Constructions:
Conceptual Distance and Cognitive Processing Levels”. Complementa-
tion: Cognitive and Functional Perspectives ed. by Kaoru Horie, 199–225.
Amsterdam: John Benjamins.
Visser, Frederik Th. 1963–1973. An Historical Syntax of the English Lan-
guage. 4 vols. Leiden: E. J. Brill.
Vosberg, Uwe. 2003. “Cognitive Complexity and the Establishment of –ing
Constructions with Retrospective Verbs in Modern English”. Insights into
Late Modern English ed. by Marina Dossena and Charles Jones, 197–220.
Bern: Peter Lang.
Vosberg, Uwe. 2003. “The Role of Extractions and horror aequi in the Evolu-
tion of –ing Complements in Modern English”. Determinants of Gram-
218 References

matical Variation in English ed. by Günter Rohdenburg and Britta Mon-


dorf, 305–327. Berlin: Mouton de Gruyter.
Vosberg, Uwe. 2006. Die Große Komplementverschiebung: Außersemantische
Einflüsse auf die Entwicklung satzwertiger Ergänzungen im Neueng-
lischen. Language in Performance, 34. Tübingen: Gunter Narr Verlag.
Vosberg, Uwe. 2009. “Non-finite Complements”. One Language, Two Gram-
mars?: Differences between British and American English ed. by Günter
Rohdenburg and Julia Schlüter, 212–227. Cambridge: Cambridge Univer-
sity Press.
Warner, Anthony. 1982. Complementation in Middle English and the Meth-
odology of Historical Syntax: A Study of the Wyclifite Sermons. Croom
Helm Linguistics Series. London: Croom Helm.
Weiner, E. S. C. and Andrew Delahunty. 1994. The Oxford Guide to English
Usage. 2nd edition. Oxford: Oxford University Press.
Westney, Paul. 1992. “Revisiting –ing –ing”. Who Climbs the Grammar Tree
ed. by Rosemarie Tracy, 493–503. Tübingen: Max Niemeyer.
Whitcut, Janet (ed). 1994. Eric Partridge, Usage and Abusage: A Guide to
Good English. New York: W. W. Norton and Company.
Wierzbicka, Anna. 1988. The Semantics of Grammar. Studies in Language
Companion Series, 18. Amsterdam: John Benjamins.
Wierzbicka, Anna. 2006. English: Meaning and Culture. Oxford: Oxford Uni-
versity Press.
Wilson, Kenneth G.. 1993. The Columbia Guide to Standard American English.
New York: Columbia University Press.
Wood, Frederick T. 1956. “Gerund versus Infinitive”. ELT Journal 11.11–16.
Wood, Frederick T. 1981. Current English Usage. Revised by Roger H. Flavell
and Linda M. Flavell. London: Macmillan.
Index

A
absit 35, 178 British National Corpus 19 See also
adversative predicates 1 BNC
affirmative sentence(s) 124–126, Brown 19, 21–23
130–133, 135–144, 146–148, Burchfield, R. W. 29, 81, 87–88, 92,
158, 185–187 138-139
Aijmer, K. 129, 132 Burnard, L. 23
Allterton, D. J. 13 but 3–4, 84, 106, 134–135, 138–139,
American English 9, 11, 88–89, 122, 145–146, 148–152, 156–158,
135, 139 166, 182, 186–188, 192–193
Andersson, E. 88
any 41–42, 82–83, 178
Aston, G. 23 C
avoid 24, 64–71, 73–77, 180 cannot forbear 58–59, 179
comment clauses 127
complement shift(s) 1, 5-6, 8, 10, 17,
B 24, 27, 31, 34, 43, 51, 55–57,
Baghdikian, S. 2 80, 83–84, 106, 126, 143, 161,
before 2 173–175, 177–178, 181–182,
The Bible in English 19, 21–22, 24, 28, 185, 190, 195–196 See also
38, 141, 143, 145, 148, 150, the first complement shift, the
187 second complement shift, and
Bladon, R. A. W. 120 shift(s) of complements
Blake, N. F. 3, 11 Complexity Principle 98, 183
BNC 19, 21-23 See also the British Curme, G. O. 3–4, 8, 15, 145, 193
National Corpus Cuyckens, H. 121, 196
Bolinger, D. 72
Brewer, Ch. 19, 198
Brinton, L. J. 127, 129–130, 132, 186
220 Index

D FLOB 19, 21–23 See also Freiburg-


Davies, M. 23 LOB Corpus of British Eng-
De Smet, H. 15, 120–121, 196–197 lish
Dekeyser, X. 95, 97 forbear 24, 52–64, 179–180
Delahunty, A. 88 forbeodan 2
Denison, D. 3, 11, 20, 95, 129 forbid 23, 28–42, 177–178
deny 25, 159–175, 189–190 forget 15
Dixon, R. M. W. 5–6, 97, 100, 137 forsacan 44
do 153, 188 Fowler, H. W. 29, 81
doubt 25, 137–156, 186–188 Freiburg-Brown Corpus of American
doubtless 155, 188 English 19 See also
Duffley, P. J. 6, 12, 14–15, 46, 197 Frown
Freiburg-LOB Corpus of British English
19 See also FLOB
E Frown 19, 21–23 See also Freiburg-
Egan, Th. 12, 14, 29, 43, 52, 64, 120, Brown Corpus of American
122, 160 English
Ellegård, A. 153
Espinal, M. T. 127–128
expletive negation 2–3, 37–41, 55, G
81–84, 102, 114, 135, 139, 144– Garner, B. A. 114
146, 148, 151–152, 165–166, Givon. T. 98
178, 182, 192–193 God forbid 24, 30–31, 35–42, 177–
178 See also forbid
Görlach, M. 88
F Götz, D. 30, 43, 52, 64, 81, 88, 99,
factive 12 122, 157, 159
Fanego, T. 5–6, 11, 16–18, 47, 56, gradience 128
58, 69, 136, 197 Gramley, S. 72
fear 25, 120–136, 185–186 grammaticalisation 196
first complement shift 7–10, 23–25, The Great Complement Shift 6
27, 33–34, 51, 55, 79, 83, 102, Greenbaum, S. 6, 13–15, 81, 95, 97,
107, 119, 126, 163–164, 172, 128–129, 138, 140
179, 185, 189, 195 See also The Guardian 89
the first shift, complement shift
and shift of complements
first person 132–133, 147, 152, 170– H
172, 186, 189 Haegeman, L. 15
the first shift 80, 91, 173, 182 Hamawand, Z. 5, 12–13, 15, 196
Fischer, A. 19 Haspelmath, M. 14, 196
Fischer, O. 2, 9, 32, 197 Havelok 2
Index 221

Heath, D. 30, 43, 52, 64, 81, 88, 99, Kiparsky, P. 12–13
122, 157, 159 Kirsten, H. 92, 95
Helsinki Corpus (of English Texts) Koopman, W. 9
19, 22, 33–34, 41 Kytö, M. 23

Herbst, Th. 30, 43, 52, 64, 81, 88, 99,


122, 157, 159 L
Heyvaert, L. 92–93, 95, 97 Lancaster-Oslo/Bergen Corpus of
hinder 24, 99–111, 183–184 British English 19 See
historical corpus 19–20, 23, 25, 197 also LOB
Hoffmann, S. 20, 197 Langacker, R. W. 13
Hopper, P. J. 196 Lange, C. 132
horror aequi 24, 71–77, 180–181 Leech, G. 6, 13–15, 95, 97, 128–129
Houston, A. 197 lest 134–135, 146, 148, 150, 156,
Huddleston, R. 1, 5–6, 43, 97, 120, 166, 182, 186–187, 193
129, 140 lexical diffusion 193–194
LOB 19, 21–23
López-Couso, M. J. 3, 151
I Los, B. 5, 9, 44–45
I doubt 7, 10. 25, 137, 147–148, 191,
196 See also doubt
I fear 7, 10, 25, 127–134, 136, 186, M
191, 196 See also fear Mair, Ch. 11, 16, 19–21, 23, 63, 88,
if 6, 140–143, 149, 156–157, 187– 92–93, 95, 197
188 Manabe, K. 9
Ishiguro, T. 2, 37 Méndez-Naya, B. 3, 151
Iyeiri, Y. 2, 55, 153 Miami Herald 89
Middle English Dictionary 31
Milsark, G. 72
J Molencki, R. 17
Jack, G. B. 2 More, Thomas 8
Jespersen, O. 1–2 Mulac, A. 128–129, 132
Joubert, J.-F. 197 Murray, K. M. E. 20
Jucker, A. H. 19

N
K ne 2–3, 55, 165
Kaltenböck, G. 127–128 Nevalainen, T. 193
Kent, Ch. W. 2 Nilsen, D. L. F. 43, 48
Kikusawa, N. 134 no doubt 154–155, 188 See also
King James Bible 31 doubt
Kiparsky, C. 12–13 Noël, D. 12
222 Index

non-assertive 41–42, 82–83, 178 Rogiers, H. 92–93, 95, 97


Noonan, M. 196 Rohdenburg, G. 6–8, 71–72, 98, 183
Ross, J. R. 72–73
Rudanko, J. 47, 58, 120, 162
O
OED 20–21, 25, 28, 63–64, 197–198
See also the Oxford English S
Dictionary Schäfer, J. 19–20, 198
The Oxford English Dictionary 2, 19 second complement shift 7, 10, 24,
See also OED 27, 31, 34, 45, 51, 54, 83, 102,
119, 144, 172–173, 179, 181,
185, 190, 195 See also the
P second shift, complement
Palmer, F. R. 15 shift(s), and shift(s) of comple-
paratactic negation 2 ments
parenthetical 7, 10, 25, 127–130, second shift 55, 57, 79, 80, 126, 182,
134, 136–137, 147–148, 150, 189 See also the second
152–154, 158–159, 167–172, complement shift, complement
186, 191, 196 shift(s), and shift(s) of comple-
Partridge, E. 29, 81 ments
Potter, S. 11 Shakespeare, William 20, 37, 41, 136
pragmatic Inference 192 shift(s) of complements 6, 23–25, 30,
prevent 24, 87–99, 182–183 33, 35, 44, 52–53, 91, 105, 113,
prohibit 24, 80–87, 181 120, 126, 133, 174, 178, 180,
Pullum, G. K. 1, 5, 43, 72, 97, 120, 185, 189–190, 194, 198 See
129, 140 also complement shift(s)
Skandera, P. 11
Smith, J. 132
Q Smithers, G. V. 2
question 1, 156–159, 188 Söderlind, J. 66
Quirk, R. 6, 13–15, 95, 97, 128–129 Strang, B. M. H. 11
subjunctive 9–10, 121, 134, 150,
167, 190–191, 196
R Svartvik, J. 6, 13–15, 95, 97, 128–
refrain 24, 112–117, 184–185 129
refuse 23, 43–49, 75–77, 178–179, Swan, M. 64, 122
181 Sweet, H. 12
remember 15–17
require 9
Rissanen, M. 1, 3, 132 T
Roe, I. F. 30, 43, 52, 64, 81, 88, 99, Tagliamonte, S. 132
122, 157, 159 Tajima, M. 32, 89, 197
Index 223

that ne 3 Vermeylen, N. 92–93, 95, 97


Thompson, S. A. 127–129, 132 Vespoor, M. 12–13, 77
Tieken-Boon van Ostade, I. 153 Visser, F. Th. 5, 8, 16, 45, 51, 65–67,
Tottie, G. 193 81, 88, 95, 97, 103, 120, 153
Traugott, E. C. 2, 196 Vosberg, U. 4, 17, 51, 56, 58-60, 65–
Tremblay, R. 14 68, 71–73, 75, 77, 99, 114,
try 15–16 117, 196

U W
unless 2, 134 Warner, A. 3, 55, 151
until 2 Webster, Noah 35
Urmson, J. O. 129, 132 Weiner, E. S. C. 88
Westney, P. 72
whether 6, 134, 137–143, 149,
V 156–157, 187–188
van der Wurff, W. 1–2, 9, 32, 165, Whitcut, J. 81, 138, 140, 143
197 Wierzbicka, A. 11, 13, 132, 186
van Ek, J. A. 92, 95 Wilson, K. G. 89
van Gelderen, E. 9 wiðcweðan 44
van Kemenade, A. 9 wiðsacan 44
verbs of implicit negation 1 Wood, F. T. 12, 95
verbs of negative import 1