This book is primarily a survey, but, unlike some other topic-based books in this series, it surveys two areas. First, it offers an account of the range of variation displayed by languages in relation to definiteness and related grammatical concepts. Most languages do not have “articles”, and in those that do they vary strikingly in both their form and their range of use. All languages have demonstratives, personal pronouns, possessives, and other expressions which either seem to be inherently definite or to interact in interesting ways with definiteness; but again, there is considerable variation in the ways in which these expressions relate to definiteness. Second, the book gives a (very selective) outline of the theoretical literature on definiteness. This literature is vast, consisting both of direct accounts of definiteness and of work mainly concerned with other phenomena on which definiteness impinges. Both the cross-linguistic survey and the theoretical survey are introductory and far from complete, and many of the choices I have made in reducing the material to manageable proportions are no doubt arbitrary. This is true particularly as regards the literature, where I have had to omit much which I see as important, and it is essential that the reader follow up further references given in the works I do refer to.

This is not just a survey, however. I am much too interested in the topic not to want to present my own view of what definiteness is, and I believe the work gains in coherence from the aim of reaching and defending (if in outline) a preferred account.

The informal attempt in Chapter 1 and subsequently to reach a general definition of definiteness ran into a puzzle. Definiteness seems empirically to be a unified phenomenon, on the evidence of the way languages represent it, but it is not straightforwardly characterized. Two characteristics are prominent, but neither is apparently fully adequate as the defining feature. Identifiability is particularly attractive for referential uses, especially where the referent is a physical entity locatable in a physical context, and inclusiveness is particularly attractive for nonreferential uses. Indeed many uses are readily handled by either one of these concepts. But neither works for all uses. In this chapter we survey attempts to analyse definiteness within various theoretical frameworks. As will be seen, writers have variously argued for versions of identifiability or of inclusiveness, or have simply assumed one or the other, as the basic descriptive insight. The general tendency is for logicians and semanticists to prefer inclusiveness (or, very often, uniqueness – thus limiting themselves to accounting for singular definites), and pragmatists to prefer identifiability. But this is by no means a general rule; some have indeed sought to combine the two.

After outlining some major approaches I will argue (following up hints dropped in preceding chapters) that the attempt to find a fully unified characterization of definiteness in semantic or pragmatic terms is misguided. I will propose an account of definiteness as a grammatical category which, like other such categories, cannot be completely defined in semantic or pragmatic terms, though it represents the grammaticalization of some category of meaning.

Introduction

In this chapter we investigate the structure of phrases and clauses in simple sentences. There are two fundamental aspects of structure which every theory must deal with: relational and non-relational structure. As the names imply, relational structure deals with the relations that exist between one syntactic element and another, be they syntactic, semantic or pragmatic in nature, whereas non-relational structure expresses the hierarchical organization of phrases, clauses and sentences, however it may be conceptualized. Semantic relations are the focus of chapters 3 and 4, and pragmatic relations are the main subject of chapter 5. Syntactic relational structure is the main topic of chapter 6, which focuses on grammatical relations. We will concentrate on the non-relational structure in simple phrases and sentences in this chapter. The structure of complex noun phrases and sentences is discussed in chapter 8.

Before we begin, however, there are two general theoretical issues that need to be addressed: how many levels of syntactic representation are there in a grammar, and what aspects of clause structure are universal? These issues will be considered in the next section.

General theoretical issues

Levels of syntactic representation

One of the most important theoretical claims Chomsky made in his early work was that no theory of grammar could approach descriptive or explanatory adequacy if it recognized only a single level of syntactic representation, namely the overt or surface form. He argued that an additional, abstract level of syntactic representation is required.

Introduction

In this chapter we address the details of the linking between syntax and semantics in simple sentences, that is, between the semantic representations introduced in chapters 3 and 4 and the syntactic representations introduced in chapter 2. This will involve syntactic pivots and controllers, as discussed in chapter 6. This will also involve focus structure, as we saw at the end of chapter 5. The system for linking syntax and semantics developed in chapters 3–6 is summarized in figure 7.1. The selection of privileged syntactic arguments, i.e. syntactic pivots and controllers, is based on the actor part of the Actor–Undergoer Hierarchy, as given in (7.1).

1. (7.1) Privileged syntactic argument selection hierarchy

2. arg. of DO > 1st arg. of do′ > 1st arg. of pred′ (x, y) > 2nd arg. of pred′ (x, y) > arg. of pred′ (x)

In syntactically accusative constructions, the highest-ranking argument is the default choice to be the privileged syntactic argument (pivot or controller), whereas in syntactically ergative constructions, the lowest-ranking argument is the default choice. As we will see later in the chapter, languages vary as to whether the privileged syntactic argument must be a macrorole or not. PSA modulation voice constructions permit an alternative choice for pivot or controller. The basics of the linking for a simple sentence in English are illustrated in figure 7.2, repeated from chapter 5; the operator projection is omitted. The double-headed arrows in figures 7.1 and 7.2 indicate that the linking is bidirectional, that is, that linking goes from semantics to syntax and from syntax to semantics.

Introduction

In the previous chapter we presented a theory of morphosyntactic structure which elucidated the structure of simple sentences and noun phrases. At many points in the discussion we made crucial reference to predicates and their arguments and to the semantic representation of sentences. Our task in this chapter and the next is to present a theory of just these things. We begin by presenting a classification of the kinds of events, actions and situations that sentences express and of the roles that the participants in these states of affairs may play. We then turn to the problem of representing the relevant semantic properties of verbs and other predicates that code these states of affairs; these representations will in turn form the basis of the semantic representations of clauses and sentences. From these we will derive the representation of the arguments of the verbs and other predicates, arguments which denote the participants in the states of affairs. In the next chapter we will present the notion of ‘semantic macrorole’ and investigate the semantic representation of adjuncts, operators and noun phrases.

A typology of states of affairs and their participants

In chapter 1 we pointed out that the general perspective from which this book is written maintains that the communicative functions of language are central to the analysis of its structure, and one (but not the only) function of language is reference and predication, that is, representing things that happen in the world (or a possible, fictional world) and the participants involved in those situations.

Introduction

Whenever a sentence is uttered or written, it is done so in a particular communicative context, and for the addressee to correctly interpret the communicative intent of the speaker/writer, the addressee must interpret the sentence in that same context. But as this context goes far beyond the immediate linguistic context to include assumptions of many different types, identification of the proper context by the addressee is not always possible, and so misunderstandings can take place. In order to decrease the chance of misunderstanding, the speaker, in creating the sentence, tailors the form of the sentence to allow the hearer to create the proper context for interpretation with minimal processing effort. For his part, the hearer assumes that the sentence will be tailored in just this way, and so takes the first proposition that comes to mind as the one the speaker intended to communicate, and the first associated set of contextual assumptions that come to his mind as the intended background assumptions. A crucial aspect of this tailoring is the distribution of information in the sentence, which we will call the ‘information structure’ of the sentence (similar to what the Prague School linguists called ‘the functional sentence perspective’). To give one simple example, in the most common type of situation this generally means that the NP referring to the topic that is being spoken about will come first, and the expression of the comment being made about the topic will follow.

Introduction

At the beginning of chapter 2 we stated that there are two types of structure, relational and non-relational. As the labels imply, relational structure deals with the relations that exist between one syntactic element and another, be they syntactic, semantic or pragmatic in nature, whereas non-relational structure expresses the hierarchical organization of phrases, clauses and sentences. Non-relational structure was the focus of chapter 2, while semantic and pragmatic relational structures were the topics of chapters 3–5. In this chapter we turn to syntactic relations, or, as they are better known, grammatical relations. We will begin by looking at some of the conceptions of grammatical relations that have been proposed by different linguistic theories and the implications for theory and analysis of each of the major conceptions, then we will discuss the cross-linguistic diversity of syntactic phenomena related to grammatical relations and propose an account of grammatical relations which deals with this diversity.

Conceptions of grammatical relations

Grammatical relations are a part of traditional grammar. They are important because if one thinks pretheoretically, or as pretheoretically as one can, it is obvious that there are a lot of syntactic phenomena that relate to grammatical relations. For example, if one considers what the -s is doing on the third person singular present tense verb in English, it is clear that it is agreeing with the subject. Notice that this innocuous statement presupposes a theory of grammatical relations of some kind.

Introduction

We now turn to the issue of the syntactic structure of complex sentences and complex NPs. The last six chapters have laid out the essential syntactic, semantic and pragmatic features of simple sentences, and in this chapter and the next we will investigate these aspects of complex sentences, starting with the layered structure of the clause in complex sentences.

Theoretical issues

There are two fundamental questions that every theory must answer about complex sentences; they are given in (8.1).

1. (8.1) a. What are the units involved in complex sentence constructions?

2. b. What are the relationships among the units in the constructions?

A great deal of controversy has surrounded the question of units in contemporary syntactic theory. In GB, all units in complex sentences contain a subject-predicate structure; the theory does not recognize any subclausal units in complex constructions. In GPSG, HPSG, ConG and LFG, on the other hand, both clausal and subclausal (VP) units are posited in complex sentences. In our approach, the answer to (8.1a) is derived from the layered structure of the clause: the fundamental building blocks of complex sentences are the nucleus, core and clause. The traditional answer to (8.1b), the question about the structural relationships among units in a complex sentence, is summarized as follows:

Introduction

In the previous chapter we presented a system of lexical representation for verbs and other predicating elements and their arguments; the logical structures form the basis of the semantic representation for clauses and whole sentences. In this chapter we will fill in the remaining pieces that are needed for semantic representations, in particular the semantic representation of noun phrases and of clausal and NP operators. We will also discuss the lexicon, focusing on what kind of information needs to be represented in lexical entries and in lexical rules. We begin by continuing the discussion of the kinds of semantic relations that an argument can bear to its predicate.

Semantic macroroles

In this book we are presenting a framework for syntactic analysis which directly links the syntactic representation of a sentence, as developed in chapter 2, to the semantic representation, which was developed in chapter 3 (see figure 2.2). The aspects of grammar where these two interact is known as the syntax-semantics interface. The full linking system for simple and complex sentences will be the primary focus of chapters 7 and 9, but we need to introduce an important component of the linking system at this point, since it is tied in with important issues of lexical representation, argument structure and the content of lexical entries for verbs in the lexicon. This is the notion of semantic macroroles. Macroroles are generalizations across the argument-types found with particular verbs which have significant grammatical consequences; it is they, rather than specific arguments in logical structure, that grammatical rules refer to primarily.

Introduction

This book is about some of the devices users of human languages employ to put meaningful elements together to form words, words together to form phrases, phrases together to form clauses, clauses together to form sentences, and sentences together to form texts. The emphasis here will be on the construction of units larger than words, in particular clauses and sentences. This has often been viewed primarily as the domain of syntax. ‘The term “syntax” is from the Ancient Greek sýntaxis, a verbal noun which literally means “arrangement” or “setting out together”. Traditionally, it refers to the branch of grammar dealing with the ways in which words, with or without appropriate inflections, are arranged to show connections of meaning within the sentence.’ (Matthews 1982:1). The expressions of a language involve a relationship between a sequence of sounds and a meaning, and this relationship is mediated by grammar, a core component of which is syntax. In English and many other languages, the arrangement of words is a vital factor in determining the meaning of an utterance, as illustrated in (1.1).

1. (1.1) a. The man saw the woman.

2. b. The woman saw the man.

In Dyirbal (Australia; Dixon 1972) and many other languages, however, the order of words is irrelevant to the determination of the meaning of a sentence; it is, rather, the inflectional form of a phrase which is the crucial factor determining the interpretation of the sentence, as shown in (1.2). (The base forms of each noun are italicized.)