Introduction

With the exception of Afrikaans, the partially restructured languages examined here have the subject-verb-object word order for declarative sentences that is found in both their Western European superstrate and Niger-Congo substrate. In fact, this is also the basic word order in all of the Atlantic creoles. Brazilian Vernacular Portuguese favors the SVO order even for object pronouns, where European Portuguese can have SOV order. Afrikaans is unlike the other varieties in that it has basic SOV word order, but it is like them in so far as it follows the word order of its superstrate, Dutch, which coincides with the SOV order of some of Afrikaans' substrate languages.

While the European lexical source languages can require the inversion of the subject and the verb (or auxiliary) to transform a statement into a question, this is not a part of the syntax of Niger-Congo languages or full-fledged creoles. Instances of creole-like non-inversion of subject and verb (or auxiliary) that would be unacceptable in the source language can occur freely in all of the partially restructured varieties under discussion except Afrikaans, which adheres strictly to Dutch word order.

The structure of AAE clauses

AAE word order

African American English has the usual English subject-auxiliary inversion (or lack of it) in questions that can be answered “yes” or “no,” e.g. “Can I go?” (Burling 1973:68).

Introduction

Language is a kind of social behavior, one of the many ways in which individuals interact with those around them. Thus linguistics is a social science, and linguists take pride in thinking of themselves as scientists, with all the objectivity that word denotes. Unfortunately, objectivity is very hard to achieve, especially in the social sciences, and linguistics is no exception. It is hard to imagine any study of language which manages to put away all ideology, but in the case of the languages discussed in this book, the task is unimaginable.

African American English – also called AAE, Ebonics, or just Black English – is a good case in point. Until at least the middle of the twentieth century, the overwhelming majority of white Americans saw their country and its culture as the product of their European roots flourishing in a new land. This ideology allowed very little room for the contribution of other cultures, so that even the distinctiveness of the folk ways and speech of African Americans was attributed to their frequent lack of access to education and general ignorance – if not to their very intelligence. Thus well into the 1950s Negro Nonstandard English (as AAE was then called) was usually considered bad English in need of eradication rather than study. In so far as its origins were considered at all, it was assumed to have descended solely from British dialects that had been left untended in America.

Introduction

This concluding chapter will attempt to relate the social information in chapter 2 to the linguistic information in chapters 3–5 in order to provide a basis for a theory that can account for the facts known about the five languages examined here: what led to their partial restructuring and how this process affected their structure.

Social factors in partial restructuring

Chapters 2 pointed to a single, overriding social factor in the development of African American English, Afrikaans, Brazilian Vernacular Portuguese, Nonstandard Caribbean Spanish, and the Vernacular Lects of Réunion French as varieties distinct from both unrestructured overseas varieties of their source languages (e.g. the English of Ontario, the extinct Dutch of New York and New Jersey, the Portuguese of Madeira, the Spanish of Chile, or the French of Quebec) and completely restructured creole languages (e.g. Guyanese Creole English, the extinct Creole Dutch of the Virgin Islands, Guiné-Bissau Creole Portuguese, Palenquero Creole Spanish, or Mauritian Creole French). That social factor is the demographic balance, during the first century of a new language's development, of native speakers versus non-native speakers of the European source language.

Parkvall (2000) is certainly correct in his conclusion that this demographic ratio is not the only social factor that determines the degree of language restructuring: there are other relevant factors, such as an incoming population already having some fluency in a common restructured language brought in from elsewhere, such as the English-Creole-speaking slaves imported into the southern American colonies during the seventeenth century (introduction to chapter 2 and section 2.1.1).

Introduction

The kind of partial restructuring of languages examined here is clearly related to the more complete kind of restructuring called creolization, and there is a widespread consensus that the defining characteristics of creolization include social as well as linguistic factors (Holm 2000a:68–71). In fact, Thomason and Kaufman (1988:35) demonstrated that “it is the sociolinguistic history of the speakers, and not the structure of their language, that is the primary determinant of the linguistic outcome of language contact.” It will be argued here that partial restructuring is also a sociolinguistic process, and to understand it we must consider phenomena that are both social and linguistic.

One of the crucial social factors in creolization has to do with the proportion of native versus non-native speakers of the lexical source language present at the time the creole language emerged. Although Parkvall (2000) has questioned the centrality of this demographic ratio in determining the degree to which creoles are restructured, his comparative work actually reaffirms the validity of this correlation, which is the topic of the following discussion.

Bickerton (1981:4) claimed that the maximum percentage of native speakers of the superstrate language associated with full creolization was 20 percent. He was not the first to ponder this part of the equation; over a century earlier, Van Name surmised that

Introduction

The verb phrase has been of central importance in contact linguistics. While it is true that no particular set of syntactic features can identify a language as having undergone restructuring without reference to its sociolinguistic history, it is also true that the structure of the verb phrase has been of primary importance in distinguishing creole varieties (e.g. Jamaican) from non-creole varieties (e.g. Caymanian English) of the same lexical base. In the Caribbean, the non-creoles have their European system of tense marking (e.g. auxiliary verbs and verbal inflections) more or less intact, whereas the creoles have a radically different way of dealing with tense and aspect. With few exceptions, basilectal Atlantic creole verbs have no inflections; instead, they are preceded by particles indicating tense (the time of an action's occurrence) or aspect (referring to its duration, recurrence, completion, etc.). These often have the outer form of auxiliary verbs from the lexical source language (which occupy a similar position and usually serve a similar function), but semantically and syntactically they are much more like the preverbal tense and aspect markers in many of the creoles' African substrate languages.

The data below show that verbs in the partially restructured languages have few if any inflections – far fewer than their European source languages. In some cases their auxiliary verbs can take on the semantic or even syntactic uses of preverbal markers in fully creolized languages of the same lexical base, giving rise to constructions quite unlike those in their source languages.

Introduction

None of the inflectional morphology of the noun phrase in the European source languages is preserved – at least functioning as such – in fully creolized languages. This inflectional morphology can be relatively complex. In the English NP, nouns are marked for plural number; while they take no inflectional marking for grammatical gender, they do take a possessive inflection. Dutch nouns, in addition to taking plural inflections, co-occur with articles marked for gender as well as number, and these and other modifiers agree with the head; however, Dutch possessive inflections have largely given way to periphrastic constructions. In Spanish and Portuguese modifiers such as determiners and adjectives must be marked by inflections to agree with the head of the NP in gender (usually -o for masculine and -a for feminine) and number (zero for singular, -s for plural). While number agreement is audible in spoken Spanish and Portuguese, it is often not audible in spoken French, in which the pluralizing inflection (usually -s) is often silent. Still the covert system of plural marking of French, in which nouns co-occur with articles and other modifiers marked for plurality, is nearly as salient as the overt system of plural marking in Spanish and Portuguese. Furthermore, personal pronouns in all five European languages must agree with the noun to which they refer in number and gender – the latter not only in the third person singular, but also in all persons of the plural in Spanish.

The naming of books is a tricky business, and it has consequences. In the case of this book – a comparison of five languages with a surprising commonality in structure and social history, and an account of the linguistic processes that formed them – the first thing that may strike anyone who knows my work is not so much the title, as what the title is not.

Since the late 1980s, I have used the term “semi-creolization” for the process I discuss here. This book, however, is not called “semi-creolization”. The change is in part strategic: I wish to make an argument which is, I think, original about the nature of these language varieties, which all raise current and important issues of politics and culture. I am not interested in exercising an imperial right to label, especially if it obscures discussion about the issues raised by this book. Creole language studies was one of the first post-colonial disciplines: Reinecke was surely far ahead of his times in seeing creoles from the perspective of their speakers rather than those who sought to be their speakers' imperial masters.

I am very aware that the social sciences' long demand to label at will is always problematical. I also know that the use of a new or unfamiliar term in an established field is an irritant to others who, perhaps, have a word they like better or even a certain resistance to thinking again about issues of taxonomy.

Introduction

In Chapter 2 we discussed three well-known cases which involve reanalysis of a verb to an auxiliary element, an affix, or a particle. All three cases share reanalysis of V to a T element. In the Romance case, though, habeo further reduced to a suffix, while in Greek thelo became a particle arguably in the C system, thus following the V > T > C reanalysis path. In this chapter, we turn to the grammaticalization of C elements. In the first three sections (3.1–3.3) we will consider the development of the subjunctive particle na in Greek, of Southern Italian mu and of the infinitival marker to in English. In section 3.4, we look at the accounts of the development of that-complementizers in Germanic (cf. Ferraresi 1991, 1997, Kiparsky 1995, Longobardi 1991) and in connection to this we also briefly discuss the Greek complementizer pou. Finally, in section 3.5 we consider the development of complementizers out of lexical verbs, and in particular out of a serial verb construction. Our analysis heavily relies on the data discussed in Klamer (2000). In this case we also show that lexical to functional reanalysis is upwards.

Section 3.1 starts with the particle na in Greek, as its discussion is crucial for the analysis of the elements mu and to. The development of na is seen in the light of the changes that took place in the history of Greek and were discussed in the previous chapter in relation to tha.

This book has two related goals. On the one hand, we wish to address the question of syntactic change in the context of the minimalist programme, by using (variants of) some of the technical devices that have been proposed in order to provide a general analysis of a pervasive diachronic phenomenon, grammaticalization. On the other hand, we wish to address a deeper question raised by the nature of the minimalist programme itself. A central idea behind the minimalist programme is the idea that language is in some sense a perfect system (the strong minimalist thesis: see Chomsky (1995:1–10), (2000:96f.), (2001:1–2)). Now, perfect systems do not vary over time, so the very existence of syntactic change appears to be a challenge to this thesis. The existence of synchronic variation among grammatical systems also poses an apparent problem for the strong minimalist thesis. The account of grammaticalization that we develop will lead to what we believe to be an interesting response to this problem, and an explanation for the existence of apparent variation and change in syntactic systems which we believe to be consistent with the strong minimalist thesis.

The term grammaticalization was first introduced by Meillet (1912) to describe the development of new grammatical (functional) material out of ‘autonomous’ words. Since then the topic has received much attention in the literature on language change, especially amongst typologists (see the references and citations in Janda (2001), and the impressive compendium of examples of the phenomenon in Heine & Kuteva (2002)).

Introduction

The purpose of this chapter is to provide empirical evidence for the claim that grammaticalization involves reanalysis of functional categories. The central idea is that whenever grammaticalization takes place, the content of at least one functional category is reanalysed, in such a way that new morphophonological realizations of functional features are created. In the notation of Chapter 1, section 1.3, new cases of F* develop for some feature F (usually, but not always, F*Merge is innovated – see below and Chapters 3 and 4 for empirical evidence). This can mean that a lexical item or class of lexical items is reanalysed as functional, or that one functional category develops into another. We will see cases of both kinds in what follows. Crucially, our approach to grammaticalization implies that it is not a structural change: functional structure is present both before and after grammaticalization takes place; what changes is the way the features associated with functional heads are realized. More precisely, assuming a universal hierarchy of functional heads, as mentioned in Chapter 1, section 1.2 (see also Cinque 1999), the change involves the overt realization of these heads.

The chapter is organised as follows: section 2.1 deals with the grammaticalization of V elements to T markers, focussing on the development of modals in the history of English. This has been treated as a typical case of grammaticalization whereby a lexical verb is reanalysed to an auxiliary element.

Introduction

In this chapter we turn our attention to the third type of functional category: D elements. We proceed much as in the previous two chapters, in that we will examine certain well-known cases of grammaticalization, in particular the development of definite determiners out of demonstratives (section 4.1), the development of n-words (sections 4.2), the development of wh-words out of indefinites (sections 4.3) and of universal quantifiers (sections 4.4). We show that several of these cases involve loss of movement inside the DP, and the formation (or loss of) an Agree relation with another head in the clause structure. In so far as they involve loss of movement, these examples of grammaticalization are therefore the D system counterpart of the development of auxiliaries in the T system, as discussed in Chapter 2. In sections 4.5 and 4.6 we discuss the development of pronouns as agreement markers (clitics and affixes), further showing how features typically associated with the DP become associated with functional heads in the clausal domain. In some respects this reanalysis is similar to the development of complementizers as discussed in Chapter 3.

In general, then, the grammaticalization of D elements offers examples of loss of movement combined with loss of morphology of a type familiar from our study of T elements in Chapter 2, the reanalysis of heads due to structural reanalysis and simplification, similar to what we saw in Chapter 3, and a new phenomenon: the addition of a formal feature to a head.