Linking Form and Meaning

Studies on Selected Control Patterns in Recent English

J. Rudanko

• Publisher: Palgrave Pivot
• Year: 2015
• ISBN:
  9781137509499
• Topic: Languages & Linguistics
• Subtopic: Linguistics
• Index: Languages & Linguistics

This book documents changes and trends in English predicate complementation. In-depth case studies of grammatical patterns presented here uncover new links between form and meaning in these constructions, offering fresh insights into explanatory principles to account for variation and change in the system of English predicate complementation.

1

Introduction

Abstract: Control constructions are central in the book, and Chapter 1 introduces the concept of control, beginning with subject control. A central assumption concerns the postulation of understood subjects, in line with other work on infinitival and gerundial complements. The assumption is motivated in the chapter. The sentential complements introduced include the to infinitival and to -ing patterns, and it is noted that the word to is homonymous in current English. The chapter also provides a brief description of the central research questions investigated in the body of the book. Chapters 2 through 4 deal with subject control, and Chapters 5 through 8 with object control. The chapter concludes with a comment on the method of focusing on individual verbs and adjectives in work on complementation in English.

Keywords: infinitival to; object control; subject control

Rudanko, Juhani. Linking Form and Meaning: Studies on Selected Control Patterns in Recent English. Basingstoke: Palgrave Macmillan, 2015. DOI: 10.1057/9781137509499.0004.

The purpose of this book is to investigate a number of new ideas in the area of the system of English predicate complementation, with an emphasis on variation and change in that system in recent English. The data for the investigation come mainly from large electronic corpora.

To introduce some of the main assumptions that are made in this investigation, it is helpful to start by considering sentences (1a–b) from the British National Corpus, the BNC:

(1)	a.	We sought to measure attitudes in several ways. (A62)
	b.	... he occasionally resorted to taking one of my sleeping-tablets. (CES)

Sentences (1a–b) exemplify subject control. In (1a) the matrix, or higher, verb is seek, and its complement is a to infinitive, and in (1b) the matrix verb is resort and its complement is what is here termed a to -ing construction, which consists of the word to and a following -ing clause. The -ing clause in question may be termed a gerund. An assumption made here is that in both (1a) and (1b) the complements of the matrix verbs are sentential, with their own subjects. This assumption is not accepted by all linguists today, but there are a number of reasons for thinking that to infinitive and -ing complements are indeed sentential. One line of reasoning was expressed by Otto Jespersen, a traditional grammarian, many years ago as follows:

Very often a gerund stands alone without any subject, but as in other nexuses (nexus-substantives, infinitives, etc.) the connexion of a subject with the verbal idea is always implied. (Jespersen [1940] 1961, 140)

The use of a verb in a sentence thus implies the presence of someone or something that the verbal idea is predicated of. Given the presence of two verbs in (1a–b) it is therefore reasonable to think of it as implying the presence of two such entities. The first of them, the subject of the higher verb, is overtly represented, and the second, the subject of the lower verb, is covertly represented. Given that the higher subjects of (1a–b) are assigned theta roles by their matrix verbs, the sentences in (1a–b) are control structures. It is customary in much current work to represent the covert – or understood or implicit – subjects of (1a–b) with the symbol PRO.

The argument for the postulation of an understood subject outlined in the paragraph above is strengthened when it is borne in mind that an understood subject can provide an antecedent to bind a reflexive in the lower sentence. For instance, consider (2a–b), from the Corpus of Contemporary American English, COCA:

(2)	a.	Vice President Joe Biden has sought to separate himself from his boss, ... (2010, MAG)
	b.	... he had to stealthily resort to identifying himself as “Dave, a caller from Washington, D.C”. (1992, MAG)

Assuming, then, that both (1a) and (1b) contain understood subjects, they may be represented as in (1a´) and (1b´), in their essential respects.

(1)	a.	[[We]NP [sought]Verb1 [[PRO]NP [to]Infl [[measure]Verb2 attitudes]VP]S2]S1
	b.	[[he]NP occasionally [resorted]Verb1 [to]Prep [[[PRO]NP taking one of my sleeping-tablets]S2]NP]S1

The representations provided contain a minimum of analysis, but they are suitable for the study of argument structure. There are two other features of the representations that are worth mentioning here. First, it is observed that the representation in (1b´) makes use of the traditional idea of a nominal clause. This is simply a sentence dominated by an NP.

Second, the structural representations given in (1a´–b´) also embody the hypothesis that there are two different types of to in English. The to preceding the -ing form of a verb in (1b´) is represented as a preposition, but the infinitival to in (1a´) is under the Infl node, corresponding to the Aux node in more traditional terminology. There are some scholars today who consider both types of to as prepositions, but it is observed that they are not in free variation, at least not in the case of the verbs seek and resort, since both (3a) and (3b), modified from the authentic examples in (2a–b), are ill-formed:

(3)	a.	*Vice President Joe Biden has sought to separating himself from his boss.
	b.	*He had to resort to identify himself as Dave.

The matrix verb seek thus only selects infinitival to and the matrix verb resort only selects prepositional to, and it is necessary to distinguish the two types in the grammar of English.

From a historical perspective, it is clear that in Old English the word to found with infinitives was a preposition. However, the infinitive has drifted from a “nominal to a verbal character,” as Denison (1998, 266) puts it. This drift is “now virtually complete,” and the drift has also involved the “concomitant dissociation of the infinitive marker to from the homonymous preposition” (Denison 1998, 266). The approach adopted here shares Denison’s basic view of the analysis of to infinitives, and treats the word to as homonymous in current English. At the same time, the term “infinitive marker” has in some treatments involved the implication that the constituent in question is “merely” a marker, devoid of meaning. This implication is not shared by the present author. More than 25 years ago he pointed to the historical associations of infinitival to (Rudanko 1989, 35), and the fact that it is placed under the Infl – or Aux – node does not mean that it is devoid of meaning. On the contrary, like other items under the Infl node, it may carry a meaning.

As for the to that precedes -ing clauses, or gerunds, as in (1b), it is probably fair to say that there is a consensus in the literature regarding its prepositional status. When infinitival to is distinguished from prepositional to, the behavior of the latter is regular: similarly to other prepositions, prepositional to co-occurs with -ing forms, or gerunds, rather than with infinitives. (Apart from prepositional gerunds, there are of course also simple or nonprepositional gerunds in English, as in the complement of avoid in I avoided looking at her, COCA, 2011, FIC).

Once it is recognized that there are two types of to in recent English, it is fascinating to investigate the variation affecting the use of the two patterns in recent English. There is little doubt in the light of a considerable body of recent work that it is the -ing variant that has been spreading at the expense of the to infinitive. Indeed the spread of the gerund, whether prepositional or nonprepositional, at the expense of to infinitives is one of the more important features of what in recent work has been called the Great Complement Shift (Rohdenburg 2006a, Vosberg 2006, Vosberg 2009, Rudanko 2010b, Rudanko 2011, Davies 2012, Rudanko 2012).

In work investigating variation and change affecting to infinitives and -ing complements it is essential to document the frequencies of the two variants and to note any trends in their incidence, but a concomitant and more interesting research task is to explore the factors that can have an impact on the variation and change observed. One of the aims of the present volume is to contribute to this ongoing area of research. Chapter 2 sets the stage for exploring an idea that the present author has mooted in recent work. The idea involves a semantic distinction that was proposed as an explanatory principle separating the two constructions at a time of considerable variation. This hypothesis is grounded in what has been called Bolinger’s Generalization. This says that a “difference in syntactic form always spells a difference in meaning” (Bolinger 1968, 127). The principle is heuristic. It does not specify what the difference is in any one case where variation involving two or more patterns is found. Instead, it sets up a research agenda for zeroing in on the semantics of different types of complementation patterns. Applying the heuristic principle to the variation between to infinitive and to -ing complements of the adjective accustomed, Rudanko (2010a) proposed that a contrast between predicates that are [+Choice] and those that are [−Choice] can shed light on variation between the two types of complement at a time when both are found in sizable numbers. At the back of the distinction is the hypothesis that the study of semantic roles, especially the Agent role, is salient in the analysis of syntactic variation and change.

The distinction between [+Choice] and [−Choice] predicates as an explanatory principle bearing on complement selection was based in Rudanko (2010a) on data from three decades of the TIME Corpus, the 1930s, the 1940s, and the 1950s. (The 1930s is the first full decade of the TIME Corpus.) It was argued in that study that during this period of three decades the complementation structure of the adjective accustomed underwent a major restructuring and that the distinction between [+Choice] and [−Choice] predicates can be used as an explanatory principle to shed light on variation between two major patterns of sentential complementation during a time of considerable variation between the patterns.

Chapter 2 of the present book takes up the same adjective accustomed, but in a different text type. While the TIME Corpus represents the text type of Magazines, the data investigated in the chapter concern the text type of Fiction. The data come from the Corpus of Historical American English, COHA, and the composition of the corpus makes it possible to examine not only the three decades covered in the study of the TIME Corpus, but to extend the investigation to the preceding two decades. One purpose is to examine whether the change in the complementation of the adjective accustomed observed in the text type of Magazines also took place in the text type of Fiction, and if so, whether it happened at the same rate. Another purpose is to find out whether the explanatory principle based on the [+Choice] and [−Choice] distinction can be applied to data from the text type of Fiction and whether it can shed light on variation between sentential complements of the adjective in that text type.

Chapter 3 extends the investigation of the semantic distinction between [+Choice] and [−Choice] predicates in a...

Table of contents

• Cover
• Title
• 1  Introduction
• 2  Tracking a Change over Five Decades: to Infinitive and to -ing Complements of Accustomed in American Fiction from the 1910s to the 1950s
• 3  A New Angle on Infinitival and of -ing Complements of Afraid, with Evidence from the TIME Corpus
• 4  Additional Data on Nonfinite Complements of Afraid
• 5  On the Semantics of Object Control in English, with Evidence from the Corpus of Contemporary American English
• 6  The Transitive into -ing Pattern as a Caused Motion Construction: The Case of Force
• 7  Exploring the Creative Potential of the Transitive into -ing Pattern
• 8  On a Class of Exceptions to Bachs Generalization
• 9  Concluding Observations
• References
• Index