In the course of the 1980s and 1990s, historical syntax in general, and English historical syntax in particular, developed into a thriving field of research. Much of the credit for this renaissance must go to the advent of sophisticated models of language variation and of linguistic theory. It is perhaps in the domain of syntax that modern theoretical work has most clearly sharpened the traditional questions of historical linguistics, leading to a surge of novel and interesting insights. Happily, this interest in theoretical questions has gone hand in hand with a continued interest in philological matters and, perhaps even more importantly, the creation of ever larger and more sophisticated computerized databases. For these reasons, it seems a particularly felicitous moment for a textbook to appear in which questions concerning the historical syntax of English are consistently addressed from the perspective of a model of syntactic theory.

The model of syntactic theory adopted in this book is the one known as the Principles and Parameters framework. This has important consequences for the way in which we view historical change. In the Principles and Parameters framework, the focus of investigation is the grammar internalized by the native speaker rather than the language output. Consequently, we will attempt throughout the book to make a distinction between language change and grammar change. In the first chapter, we outline the view of grammar change that we try to establish in the book, and set out our arguments and methodology for making the distinction between language change and grammar change. Chapters 2 and 3 are devoted to descriptive overviews of the most important features of the syntax of Old English and Middle English respectively.

Introductory remarks

This chapter presents a broad outline of the syntax of Middle English, i.e. English in the period 1100–1500. Many of the syntactic phenomena found in Old English, as described in chapter 2, continue in this period, but there is also a great deal of change. In fact, it has often been said that, while Old English is to all intents and purposes a foreign language to present-day speakers of English, (late) Middle English writings, such as those of Chaucer, Gower and Malory, do not confront modern readers with any major syntactic obstacles to comprehension. As the editors of a widely used anthology of Middle English literature put it: ‘There are many subtle differences in syntax between Middle English and Modern English, but few will present any difficulty to the reader’ (Dunn and Byrnes 1973: 13). The main reason for this difference is no doubt the occurrence of change in many areas of grammar between the Old and Middle English periods.

In chapters 4 to 9, we shall trace some of the individual changes in detail. It is with the aim of providing a framework against which to interpret these changes that this chapter sketches the basics of Middle English syntax, corresponding to the sketch of Old English in chapter 2. In section 3.2 we consider inflections in Middle English, and look at two constructions (impersonals and passives) characterized by special inflectional marking of grammatical roles. Section 3.3 deals with word order, both within the NP and within the clause.

As the descriptions in the following sections and chapters will make clear, Middle English syntax is characterized by greater variability than Old English syntax.

Introduction

In this chapter we will discuss two cases of long-term change in the history of English which have been interpreted by a number of linguists as cases of grammaticalization. We will present an account of the historical development of the periphrastic construction with have to and of the development of sentential negation, better known as Jespersen's negative cycle. We will, in the spirit of this book, analyse these cases explicitly in terms of grammar change. We want to show what role the (synchronic state of the) grammar and the language acquisition process play in a development which grammaticalization theorists – looking at it from a language-historical rather than a grammar-theoretical angle – generally regard as a universal, gradual, long-term development with an impetus of its own. This is a view that, as such, is foreign to the general approach adopted in this book, and we shall therefore propose reanalyses of the two changes, suggesting that they can be understood using the notion of grammar change that we also employed in analysing the empirical developments dealt with in the earlier chapters.

Before explicating the differences between these two approaches, we should say that the term ‘grammaticalization’ itself is used in two different ways, which is potentially confusing. According to the most traditional use of the term, grammaticalization is an empirical phenomenon, for which Meillet's (1912: 13) definition is appropriate: it involves the ‘[a]ttribution du caractère grammatical à un mot jadis autonome’ (the attribution of grammatical character to an erstwhile autonomous word). Grammaticalization in this sense refers to the countless instances of language change whereby lexical elements lose their lexical status and come to be employed as grammatical function words.

Introduction

The aim of this chapter is to give a descriptive overview of a number of important features of the syntax of Old English, i.e. English from the earliest texts (c. 800) to about 1100. The material in this chapter is primarily based on the evidence from the two main bodies of prose text in Old English: the prose of King Alfred (ninth century) and that of Ælfric, abbot of Eynsham (tenth century). In the final section of the chapter, we will touch on some of the ways in which the syntax of the prose differs from that of the Old English poetry.

Old English is the language imported into the British isles by the immigrations from the continent in the fourth, fifth and sixth centuries. It therefore evolved from a number of continental West Germanic languages/dialects. The syntactic study of Old English can be reliably based only on one dialect: the West Saxon standard written language or Schriftsprache, in which the bulk of Old English writing has come down to us. There is little scope for work on dialect syntax in Old English; almost all the texts are in the West Saxon dialect, while those works of any length that were not written in West Saxon consist mostly of interlinear glosses on parts of the Vulgate bible, and are therefore of limited use for syntactic purposes. We will nevertheless be commenting upon those aspects where they do prove revealing, which is mostly in those cases where the gloss deviates from the original.

The view of Old English presented in this chapter should be seen as a digest of the consensus view in the literature.

Introduction

A comparison of the position of the finite verb in Modern English with its precursors in Old English shows up some striking contrasts. As is well known, Modern English has regular SVO word orders in declarative clauses, regardless of whether the verb is finite or not. In questions and negative-initial clauses, we see the phenomenon of ‘subject–aux inversion’:

1. (1) a. Who did you talk to at the party yesterday?

2. b. Never have I talked to such a strange character.

Subject–aux inversion – as the name says – is restricted to auxiliaries. Main verbs do not invert in questions, even when they are finite. Rather, a ‘dummy’ auxiliary do is employed as a finite verb when the first verb in a sentence is not an auxiliary. Old English is different in most of these respects, although some of the patterns look rather familiar.

As we saw in chapter 2, Old English had many SVO word orders like those in Modern English, but at least as many SOV word orders, or orders that seem to be a mix of SOV and SVO. The position of the finite verb plays a special role in this: especially in main clauses, it is very often found following one initial constituent, regardless of whether that first constituent is the subject or not, and regardless of the rest of the word order of the clause.

Introduction

In this chapter we consider in detail the history of a specific construction, the so-called ‘easy-to-please’ construction or ‘tough’ movement. Some modern examples are (1)–(3).

1. (1) John was easy to convince.

2. (2) The problem was tough to deal with.

3. (3) He is hard to get a straight answer from.

In each of (1)–(3), there is a sequence of an adjective plus an infinitival clause which is predicated on a noun phrase. The infinitival clause contains a nonsubject gap (in (1), the gap functions as direct object, in (2) and (3) as complement of a preposition), and the noun phrase in the superordinate clause provides the interpretation for this gap. Thus in (1), the speaker is talking about convincing John, not someone else, and (2) and (3) are about dealing with the problem and getting a straight answer from him, respectively. Although (1)–(3) all have the relevant NP functioning as subject of the verb be, this is not a necessary characteristic of the construction. Instead of be, the verb may also be another copula, like seem, appear, turn out or become, and there are also examples like (4).

1. (4) I consider Mary impossible to get along with.

In this sentence, although there is no copula between the noun phrase Mary and the sequence adjective–to-infinitive, it is nevertheless common practice to say that there is a subject–predicate relation holding between Mary and impossible to get along with (in terms of the analysis of Stowell 1981, 1983 and a great deal of subsequent work, the two elements would form a small clause).

Introduction

Constructions with infinitival verb forms are a prominent feature of present-day English syntax. The infinitive without to occurs routinely after modal auxiliaries (e.g. I will/might/could do my homework now) and after verbs of perception and causation (e.g. I saw/heard/made him do his homework), while the infinitive with to occurs after a range of verbs with widely varying properties (I tried to do it; I promised Peter to do it; I ordered Peter to do it; I persuaded Peter to do it; I believe him to be innocent). If we compare the distribution and properties of infinitives in the present-day language with their Old English counterparts, the conclusion must be that some pervasive changes have taken place. Some of these changes seem to be largely quantitative: Manabe (1989) has shown, on the basis of a large collection of data, that infinitival clauses replaced finite ones at a fast rate in Middle English. Thus, in Old English (as in present-day German and Dutch), the usual construction after verbs expressing purpose or intention was a poœ-clause, as in (1):

I am finishing this book at the University of Coimbra, the first and oldest university of Portugal, where I first came as a student of Portuguese in 1988 in order to gain better access to the literature on the Portuguese-based Creoles and the Brazilian vernacular. This seems entirely proper, since it was a Fulbright award in 1993-1994 that first allowed me to teach here and to begin work on this book. I am grateful to the J. William Fulbright Foreign Scholarship Board, to the colleagues who were so helpful, especially Ana Luis and Clara Keating, and to those professors – Maria Irene Ramalho de Sousa Santos, Martin Kayman and Jorge Morais Barbosa – whose extraordinary efforts led to the creation of the chair I now hold here.

It is an honour to be part of one of the great medieval universities of Europe, but that is not the whole story of this book. From 1980 until 1998,1 taught at the City University of New York (CUNY). Few people outside that institution can appreciate the riches of its cultural diversity, such as the thousands of its students who speak Creole and semi-creole languages. Mitchell (1997) notes that ‘More students of color earn their bachelor degrees from the City University of New York than any other institution in the country.’ Although many of these students speak standard English as their first language, many also speak Creole English or French, African American Vernacular English or nonstandard Caribbean Spanish.

Introduction

This chapter traces the development of the major ideas that have shaped the study of pidgin and Creole languages. It also gives an overview of the history of the discipline itself, but its primary objective is to provide a better understanding of the climate of ideas in which the main theoretical advances were made.

Before European expansion

Although most of the known pidgin and Creole languages arose after western Europeans began establishing overseas colonies in the fifteenth century, there is ample reason to believe that more existed in earlier times than the two that have been documented: Lingua Franca and Pidgin Arabic (see below). Indeed, language contact seems likely to be nearly as old as language itself. However, languages have not been recorded in writing until the last few millennia and mixed languages have usually been among the last to be written down. Zyhlarz (1932–3) considered the language of ancient Egypt, first recorded in hieroglyphs in the third millennium BC, to have grown out of a trade language, i.e. a pidgin that developed among several Afro-Asiatic languages which came into contact in the Nile valley. If this is the case, it was essentially a Creole language (Reinecke et al. 1975:53). In any case the languages of ancient empires from China to Sumer expanded along with their military, commercial and cultural influence and it is quite likely that this happened via pidgin-ized varieties, although no known records of such speech remain.

Introduction

This chapter is a study of some of the phonological features found in a number of Creoles but not their lexical source languages. The theoretical orientation of this study is that discussed in section 1.4: the position that Creole languages resulted from a number of forces and that their features reflect the influence of both superstrate and substrate languages, universals of adult second-language acquisition, borrowing from adstrate languages, creole-internal innovations, and the convergence of all or some of these.

Sorting out which of these influences may have resulted in particular phonological features is by no means an easy task. It is especially difficult to determine the degree of continuity from the superstrate language (‘internal’ phonological development – if this concept is indeed applicable to creolized languages) as opposed to influence of the substrate languages (sound substitution conditioned by systems external to the language creolized). For the Atlantic Creoles, one of the major difficulties is the lack of detailed information about the phonology of the sixteenth- and seventeenth-century varieties of the particular superstrate dialects involved, to say nothing of the nearly total lack of documentation of the relevant substrate languages of this period. Still, the challenge of reconstructing the phonological development of Creole languages has led to some fascinating linguistic detective work, e.g. Smith (1987).

Superstrate influence is problematical because there is less continuity between the European languages and their Creoles than there is between them and their overseas regional varieties that have not undergone creolization.

Introduction

Pidgin and Creole languages cannot be defined, nor can their genesis and development be understood, without taking into account the social factors that shaped them. Pidgins, for example, are defined in part by such sociolinguistic factors as their being no one's first language, their arising in a particular social context such as trade, and their evolving as the result of non-intimate social contact between groups of unequal power. Some aspects of the definition of a pidgin are purely linguistic (e.g. reduction and simplification), but even the crucial quality of stability is claimed to depend on the sociolinguistic factor of tertiary hybridization.

The crucial element in the definition of a Creole is also sociolinguistic: that it grew out of a pidgin (or possibly an unstable pre-pidgin) that had become nativized in a particular speech community. The purely linguistic elements in the definition of Creoles (e.g. structural complexity) do not distinguish them from other natural languages. While one could draw up a list of structural features shared by most of the Atlantic Creoles (cf. chapters 5 and 6), there is little agreement that these could be used to determine whether a language is a Creole without reference to its sociolinguistic history.

The validity of the theories put forward to explain the genesis and development of pidgin and Creole languages crucially depends on whether these theories can satisfactorily take into account the many, various and complex sociolinguistic circumstances under which the known pidgin and Creole languages came into being and developed.