Overview

In the previous chapter, we argued that native speakers' competence in their native language(s) is reflected in their intuitions about sentence well-formedness (derived from their acceptability judgments) on the one hand, and their intuitions about sentence-structure on the other. We have argued that the ability to make judgments about well-formedness and structure holds at all four major linguistic levels – Phonology, Morphology, Syntax, and Semantics. Hence, it follows that a native speaker's syntactic competence will be reflected in his intuitions about the grammaticality (= syntactic well-formedness) of sentences on the one hand, and their syntactic structure on the other. We have discussed in some detail the problems of deciding whether a given sentence is grammatical or not. In this chapter, we examine the question of what it means to say that a sentence has a syntactic structure, we discuss the evidence in support of that claim; and we look at ways of representing syntactic structure.

Intuitions about Structure

Part of the evidence for claiming that sentences have a syntactic structure in language comes from the native speaker's intuitions about the structure of sentences in his language. The structural intuitions which native speakers have about the Syntax of their languages are of two types, namely (i) intuitions about how sound-sequences in sentences are structured into successively larger structural units which we call constituents; and (ii) intuitions about whether particular sets of constituents (i.e. structural units) belong to the same category or not.

Overview

So far, we have been assuming that there is only one level of structure in syntax – that is, the level of S-structure which represents the superficial syntactic structure of sentences. And we have been assuming that the task of any adequate syntax of a language is to characterise (= generate) all the well-formed S-structures of the language. We have assumed that the class of well-formed S-structures in a given language will be determined by the Base of the Grammar, which comprises (i) a Categorial Component, and (ii) a Lexicon. The Categorial Component comprises a set of category-neutral X-bar rule schemas; the Lexicon specifies the categorial status and contextual restrictions associated with individual lexical items in a given language (e.g. the fact that an inherently transitive Verb such as devour cannot be used in an intransitive V-bar in which it lacks an NP complement).

In this chapter, however, we are going to argue that our existing model of Syntax is not adequate to handle certain characteristic constructions in natural languages, and that in order to provide a principled account of the syntax of these constructions, we need to posit an additional level of structure known as D-structure (corresponding to the level of Deep Structure in earlier models).

Overview

Thus far, we have been assuming that the syntactic component of a Grammar contains three main subcomponents, namely:

1. (1) (i) a Categorial Component comprising a set of X-bar principles (e.g. category-neutral rule-schemas and rule-constraints)

2. (ii) a Lexicon (or dictionary) containing a list of all the words in a language, together with a specification of their idiosyncratic syntactic, semantic, phonological, and morphological properties

3. (iii) a Lexicalisation Principle, attaching lexical items (= words) under appropriate word-level category nodes (e.g. inserting man under N)

Let us say that the three components listed in (1) together constitute the Base Component of our grammar. In earlier chapters, we have looked in some detail at the role and form of the Categorial Component (1) (i) in a Grammar; in this chapter, we turn instead to consider the kind of information contained in the Lexicon (1) (ii). We also take a closer look at the Lexicalisation Principle (1) (iii), suggesting that it should be revised, and that perhaps it can ultimately be subsumed under a more general principle, and hence eliminated.

Categorial information

So far, we have considered only one kind of syntactic information which the Lexicon must provide us with – namely categorial information. That is, we have assumed that for each item listed in the Lexicon, our grammar must specify which syntactic category (or categories) it belongs to.

Overview

In this chapter, we focus on one particular transformation which we have not looked at hitherto – namely the rule of WH movement. Our aim is partly methodological (to show the depth of argumentation which can be used to provide empirical support for the postulation of a single transformation), and partly theoretical (to show how detailed analysis of the operation of one rule can cause us to radically rethink our ideas about rules and structures). Since the rule of WH movement plays a key role in the syntax of so-called wh-questions, we'll begin by establishing what exactly a wh-question is. Later, we'll look at the role played by WH movement in Relative Clauses, and in other wh-constructions (Exclamatives, etc.). Finally, we'll turn to consider in some detail the question of the landing-site for moved wh-phrases, and see that this question forces us into a critical reappraisal and radical revision of our earlier assumptions about the constituent structure of Clauses.

Question types

Questions in natural languages can be classified into a number of types. One major typological division, for example, is between yes-no questions and wh-questions. Yes-no questions are so called because they permit ‘Yes’ and ‘No’ (or their counterparts in other languages) as appropriate replies – as in the following dialogue:

1. (1) speaker a: Are you going out tonight?

2. speaker b: Yes/No

Of course, speaker B does not have to reply yes or no to such a question: he might instead reply ‘Maybe’, ‘I don't think so’, ‘That's right’, ‘Why do you ask?’, ‘Mind your own business’, and so forth - but at least he has the option of answering ‘Yes’ or ‘No’.

[This prologue is a short, one-act play for reader and author]

reader [groaning in protest]: Oh no, not another book on Transformational Grammar! I've never done any grammar before: will I be able to follow it?

author: Of course you will! I've aimed the book at the absolute beginner who's not done any Syntax before.

reader: What's the general aim of the book?

author: To get beginners to the point where they can understand some of the ideas and issues debated in current work on Transformational Syntax such as Chomsky's Knowledge of Language, or Barriers. I've deliberately set out to de-bug the text of the unnecessary technical jargon which plagues so much of the literature in the field. In that respect it's like my earlier Transformational Syntax book. But this isn't a second edition of that: the two are very different.

reader: Different how?

author: There are three main differences – theoretical, descriptive, and pedagogical. From a theoretical viewpoint, this book is much more up to date than the earlier one, and so uses a more recent framework, which takes into account major works published since 1981 (e.g. Chomsky's Barriers monograph). At a descriptive level, the present book has a greater data coverage than its predecessor – i.e. it discusses a wider range of constructions and rules.

Overview

The aim of this chapter is to give you some idea of the goals of linguistic theory, and to introduce you to some simple concepts which will be used throughout the rest of the book. Among the notions which will be explained in this chapter are terms such as theory of language, grammar of a language, particular/universal grammar, competence, performance, grammatically, linguistic intuition, rule-governed creativity, generate, observational/descriptive/explanatory adequacy, constraint, markedness, and innateness.

Grammatical competence

Linguistics is the study of Language. But why should we be interested in the phenomenon of Language? Chomsky gives an avowedly mentalist answer to this question. For him, the most fundamental reason for studying language is that language is a mirror of the mind – i.e. by detailed study of language, we might hope to reach a better understanding of how the human mind produces and processes language. As Chomsky remarks (Language and Mind (1972a), p. 103):

Overview

In the previous chapter, we took a detailed look at the internal structure of various different types of Phrases (Noun Phrases, Verb Phrases, Adjectival Phrases, Adverbial Phrases, and Prepositional Phrases). Now, by contrast, we shall turn to look in some detail at the internal structure of Clauses. Since much of our subsequent discussion will focus on similarities and differences between finite and nonfinite Clauses, we shall begin with a brief characterisation of the essential differences between these two Clause types.

Finite and nonfinite Clauses

The distinction between finite and nonfinite Clauses is based partly (though not wholly) on morphological criteria: thus, a Clause is finite if it contains a finite Verb (i.e. a Verb inflected for Tense/Agreement), and nonfinite if it lacks a finite Verb (e.g. if it is a verbless Clause, or if it is a Clause containing a nonfinite tenseless and agreementless Verb). Let's look briefly at the relevant Tense and Agreement properties which characterise finite Verbs.

From a morphological point of view, English can be said to have a binary (i.e. two-way) Tense contrast between present and past tense forms (in finite Clauses, obviously). The relevant set of inflections for finite regular Verbs are listed in (1) below:

1. (1) past tense: -(e)d for all persons and numbers

2. present tense: -(e)s for 3rd person singular forms

3. -ø for all other forms

(Note that ø represents a ‘zero inflectional morpheme’, and hence indicates that no overt ending is added to mark Tense in the relevant forms.)

Introduction

It does not seem far-fetched to hold Chomsky indirectly responsible for the accelerated development of sociolinguistics and ethnolinguistics at the end of the 1960s and for the emphasis laid upon pragmatics and discourse analysis in the mid 1970s. Paradoxical as it may seem, his revival of the Saussurean langue–parole dichotomy (under the names ‘competence’ and ‘performance’), and, even more important, his assertion of the autonomy of syntax, sparked a renewed interest in the study of language in its socio-cultural context. Both these twin pillars of Chomskyan linguistics seemed to many to shut out most of the more interesting questions about language, in particular those relating to its functioning in society. As a consequence, a sizeable number of linguists struck out on their own, as it were, and devoted themselves to building alternative conceptions of language, in which its social function was regarded as paramount.

The reaction to Chomsky's position that the systematicity of language is confined to competence took a number of different forms. Some, seeing systematicity outside of competence in Chomsky's narrow use of the term, attempted to extend the notion of competence to cover most of the aspects that Chomsky ascribed to performance. An example is Hymes's (1972) ‘communicative competence,’ which he defined as the knowledge of the abstract rules of a language required to produce sound/meaning correspondences, and the ability to use those correspondences between sound, meaning, and form in socially and culturally appropriate ways.

Introduction

The topic of bilingualism (a term that generally includes multilingualism; cf. Weinreich 1953: 1) fits appropriately under many headings in a taxonomy of linguistic fields such as the one that forms the basis for this Survey. That people can and do know and use more than one language sets interesting and critical challenges to numerous subfields of the language sciences. One could therefore expect to find it treated in chapters dealing with neurolinguistics (are bilingual brains different?), with language processing (when a person is bilingual, are the two languages stored and processed separately?), with second language learning (by definition the study of how some people become bilingual), with pidgin studies (for speakers of pidgins are bilingual), with language death (as bilinguals give up on the use of one language in favor of another), and with language and education (for the usual aim of language education is to add control of one or more extra varieties of language).

The importance of the topic also means that there is a vast literature to review. Anyone seeking such an overview might safely start with any one of a number of reviews of the field: those which shaped its study, such as Weinreich 1953 or Haugen 1956, or which have charted its progress since then, such as Haugen 1973, Mackey 1976, Spolsky 1976 or, most recently, Grosjean 1982.

Introduction

Dialectology is ‘the study of dialect and dialects’ (Chambers & Trudgill 1980: 3). Not surprisingly, the major issues in the field of dialectology during the century of its existence have been deciding which varieties of language count as dialect and how best to study them. Seen in this light, the history of the discipline can, without too much exaggeration, be characterized as the shift from collecting facts about geographically distributed, (mostly) rural varieties of language to analyzing the distribution of features of (mostly) urban varieties as these features correlate with social and linguistic factors.

This shift in methodology has been accompanied by important changes in the goals of the discipline. Whereas the dialectologists of the late nineteenth century collected data in order to test the contentions of the Neogrammarians about the exceptionless nature of sound change, later researchers, especially those engaged in linguistic atlas projects, collected and studied data mainly as a contribution to the description and history of a particular language. More recently, however, practitioners of urban dialectology, often referred to as sociolinguists or quantitative sociolinguists, have conceived of their work as one of the several ‘way(s) of doing linguistics.’ As Trudgill puts it, such research

Introduction: the Whorf hypothesis

The human, social world is a meaningful one, and speech is the most important device by which humans create and distribute meaning. The determination of the constraints on this creation and distribution, and the role of language in the relationship between cognizing human beings and the cognized environment, are central concerns of scholars in many disciplines.

A thorough review would require attention at least to the literatures of philosophy, psychology, anthropology, and the literary theory of translation. The role of language as a map for social action is a practical as well as an intellectual concern, as evidenced in work by feminist scholars (see McConnell-Ginet, Borker & Furman 1980), to note only one example. The present review will concentrate primarily on the shapes that these concerns have assumed within anthropology during the last decade.

Modern anthropologists concerned with the relationship between language, culture, and world view trace their intellectual genealogy through a ‘Whorf hypothesis’: that the forms of meaning created in the syntactic, morphological, and phonological patterns of language can vary more or less without limit, and that these forms, which constitute reifications of the world, are powerful mediators of human understanding, which should in its own turn assume a more or less unlimited range of forms.

Introduction

Some of the very same qualities that make conversation an important topic for linguistic attention also make conversation a difficult topic for linguistic analysis. First, in having a conversation – something which we all do virtually every day – we draw upon our communicative competence (Hymes 1972): our tacit knowledge of the abstract rules of a language, which is required both to produce sound/meaning correspondences within grammatical sentences and to use those correspondences between sound, meaning, and form in socially and culturally appropriate ways. Such competence includes ‘the knowledge of linguistic and related communicative conventions that speakers must have to create and sustain conversational cooperation’ (Gumperz 1982a: 209). Despite the ubiquity of our communicative competence, however, neither the nature of that knowledge, nor the ways in which it can be put to use for social and expressive purposes, is readily accessible to our own introspection. Rather, such knowledge has to be inferred through the analysis of structures and patterns in that which it is able to produce, i.e. in conversation. The fact that inferences about what is causing a particular phenomenon have to be made through observation of its outcome is one reason why conversation is a difficult topic for linguistic analysis.

Introduction

The study of discourse belongs to the study of language in use, which means that it is concerned not just with the properties of linguistic representations but also with the nonlinguistic factors that determine what message is conveyed by the use of a linguistic form and whether it counts as an acceptable contribution to the communicative enterprise. While the linguistic properties of an utterance may determine a range of possible interpretations, the actual message recovered by the hearer depends on its nonlinguistic properties. Consider, for example, the exchange in (1):

1. (1) A: You're not eating.

2. B: It's too hot.

A grammar that pairs phonetic and semantic representations of sentences cannot determine the reference of you or it, what meaning of hot is intended, or what it is too hot for. Moreover, it is not evident from the linguistic properties of A's utterance whether he is informing the hearer that she is to refrain from eating, asking the hearer to confirm that she is not eating, or expressing his disapproval of the fact that the hearer is not eating. The hearer is expected to recover a specific message on the basis of nonlinguistic or contextual information.

Some writers, recognizing that the interpretation of an utterance is not fully determined by its linguistic properties as they are defined by a traditional sentence grammar, have argued that the grammar should be extended to include a pragmatic component that assigns interpretations to sentence–context pairs.