Syntactic Theory

A Unified Approach

Robert Borsley

• Publisher: Routledge
• Year: 2014
• Print ISBN: 9780340706107
• eBook ISBN: 9781317858942
• Topic: Languages & Linguistics
• Subtopic: Linguistics
• Index: Languages & Linguistics

Syntactic theory is central to the study of language. This innovative book introduces the ideas which underlie most approaches to syntax and shows how they have been developed within two broad frameworks: principles and parameters theory and phrase structure grammar. While other texts either concentrate on one theory or treat them as totally separate, here both approaches are introduced together, highlighting the similarities as well as the differences.



Thoroughly updated in the light of major recent developments, this second edition includes expanded explanations of the main characteristics of the two theories, summaries of the main features, exercises reinforcing key points and suggestions for further investigation.

1

Preliminaries

1.1 Introduction

There are many aspects of language that one might be interested in. One might study how it is acquired, how it reflects the nature of society, or how it is used in literature or in propaganda of various kinds. Syntactic theory is concerned with the ways words are combined to form sentences. This might sound like a dry and unglamorous study, and it has to be admitted that some people find it so. However, it is at the heart of modern linguistics. It is widely regarded as the most important branch of the subject and as the branch in which the most important achievements have been made. Hence it must be a major focus of interest for anyone who is interested in modern linguistics. In this chapter, we will consider a number of preliminary matters, and then in the next chapter, we will begin the real work.

1.2 The goals of syntactic theory

Syntactic theory, as the term is used here, has its origins in Noam Chomsky’s 1957 book Syntactic Structures. It can be said to have two goals. On the one hand, it is concerned to develop precise descriptions of aspects of the syntax of various languages, the ways in which specific languages combine words to form sentences. On the other hand, it aims to develop a general theory of syntax, specifying what languages have in common in this area and how they can vary. This is often known as a theory of universal grammar.

One point that should be stressed immediately is that the two goals are not pursued separately. It is not a matter of first describing the syntax of individual languages and then developing a theory of universal grammar. Rather, syntacticians are always concerned both with individual languages and with language in general. Investigations of individual languages are always guided by ideas about what languages are like and how they should be described, and in turn they contribute towards the evaluation and refinement of these ideas.

The second of these aims entails a rejection of the view expressed by one linguist in the 1950s that ‘languages can vary without limit’ (Joos 1957). If languages really could vary without limit, there could be no theory of universal grammar. We could have a theory of English syntax, a theory of Welsh syntax, and so on, but no general theory. It is in fact fairly clear that languages do not vary without limit in their syntax. Differences between languages which seem massive to the learner of a foreign language can be seen to be rather minor when one has a sophisticated descriptive framework that can identify all the similarities.

One thing that is clear is that there are all sorts of situations that do not occur in languages. For example, it is clear that there are no languages in which an interrogative, the sentence-type standardly used to ask a question, is formed from a declarative, the sentence-type standardly used to make a statement, by turning the declarative back to front. In other words, there are no languages where the interrogative related to (1) would be (2).

(1) The boy ate the beefburger.

(2) Beefburger the ate boy the?

Perhaps more interestingly, it is fairly clear that there are no languages where an interrogative is formed from a declarative by moving the second word of the declarative to the front. English might seem to be such a language if one looks at a pair of sentences like the following:

(3) Hobbs will be here.

(4) Will Hobbs be here?

Notice, however, that the interrogative related to (5) is (7), not (6).

(5) The boy will be here.

(6) * Boy the will be here?

(7) Will the boy be here?

Similarly, the declarative related to (8) is (10) and not (9).

(8) Hobbs and Rhodes will be here.

(9) * And Hobbs Rhodes will be here?

(10) Will Hobbs and Rhodes be here?

Following standard practice, I use an asterisk in (6) and (9) to mark an ungrammatical sequence.

Of course, it sometimes happens that things that were thought to be impossible do in fact occur. For example, it was widely assumed in the 1970s that there are no languages whose normal order is object–verb–subject or object–subject–verb; no languages, that is, where the meaning of (1) is expressed by a sentence like (11) or a sentence like (12).

(11) The beefburger ate the boy.

(12) The beefburger the boy ate.

In the late 1970s, however, it became clear that there are languages with these word orders, especially in the Amazon basin.

There is perhaps a moral here. In general, it seems likely that proposed universals which relate to relatively concrete and superficial features of language are unlikely to be tenable. Viable universals are likely to be rather abstract. The various concepts and principles discussed in the following chapters are plausible candidates for universals.

A final point that we should note here is that the following pages will largely ignore languages other than English. The ideas that we will be concerned with will generally be illustrated with data from English. This is simply because English examples will be more accessible to most readers than examples in other languages.

1.3 Languages

We have used the term ‘language’ a number of times in the preceding discussion. Before we go any further we should ask what exactly a language is. A language is normally seen as something that is shared by a group of people, possibly a very large group. We think of English as something shared by a vast number of people and Welsh as something shared by a smaller number. However, this concept is quite problematic.

The problem is that languages in this sense are not clearcut objects. There is often no sharp break between one language in this sense and another. If we ask when Middle English stopped and Modern English started, there is no precise answer that we can give. People did not go to bed one day speaking Middle English and wake up the next day speaking Modern English. If we ask where Dutch stops and German starts, we might suggest the Dutch–German border. If we do, however, we are using sociopolitical criteria and not linguistic criteria to decide whether someone speaks one language or the other. It seems, then, that the commonsense concept of a language is a fuzzy one crucially involving non-linguistic, especially sociopolitical, factors. Hence, it is necessary to replace it with a technical concept.

What sort of technical concept? Chomsky and others argue that the main focus of linguistics should be I-(internalized) languages. An I-language is a set of rules and principles in the mind of a speaker. In earlier work I-language is referred to as linguistic competence, or knowledge of language. Probably no two individuals have the same language in this sense.

As well as an I-language, speakers have a mental lexicon, that is, a set of words which they can use. The I-language and the mental lexicon make available an infinite set of sentences, some of which will never be used because they are too long or too complex or too absurd. We will say that a sentence is grammatical for some speaker if it conforms to the rules that constitute the speaker’s I-language. Since many sentences will never be used, ‘grammatical’ is not the same as ‘likely to be used’.

A rather different concept of language has sometimes been assumed within syntactic theory. Chomsky (1957) defines a language as a set of sentences. Gazdar et al. (1985) see a language as a set of linguistic expressions of various kinds. A similar concept of language is assumed in what is known as formal language theory, a branch of mathematics which studies the properties of language like systems. An obvious question about such concepts is what sort of sets are we talking about? Clearly a language is not just any set of sentences or any set of linguistic expressions. For example, the set of sentences or linguistic expressions in this paragraph is not a language. The obvious answer is that a language is the set of sentences or linguistic expressions made available by some individual’s I-language and mental lexicon. However, this means that this concept of language is dependent on the concept of an I-language.

Chomsky argues not just that a language is a body of rules and principles in the mind of the speaker, but also that universal grammar is a body of principles and parameters which is an innate component of the mind. The former are operative in all languages, while the latter allow a limited amount of variation between languages. Thus, for Chomsky, syntactic theory is ultimately about the human mind. Even if one isn’t prepared to go all the way with Chomsky, it is reasonable to assume that syntactic theory can offer some insight into the workings of the human mind. For many people, this is a major attraction of syntactic theory.

1.4 Acceptability and grammaticality

If we agree that the main focus of linguistics is I-languages, an obvious question is: how can we investigate an I-language? We can do this by investigating what sentences are allowed by its rules and principles, in other words what sentences are grammatical, and developing hypotheses about the underlying rules and principles.

We do this by eliciting speaker’s judgements or intuitions about sentences. These intuitions establish that certain sentences are acceptable or unacceptable to that speaker. Acceptability, however, is not the same thing as grammaticality. This is mainly because sentences can be unacceptable for a variety of reasons.

Sentences can be unacceptable because they cause problems for the perceptual mechanisms. Consider first the following example:

(13) The horse raced past the barn fell.

This is unacceptable, but it is clear that it is grammatical. Notice that the very similar sentence in (14) is perfectly acceptable.

(14) The horse ridden past the barn fell.

Notice also that (13) is a reduced version of (15) in essentially the same way as (14) is a reduced version of (16).

(15) The horse which was raced past the barn fell.

(16) The horse which was ridden past the barn fell.

Why, then, is (13) unacceptable? The problem is that it leads the perceptua...

Table of contents

• Cover
• Title Page
• Copyright Page
• Dedication
• Table of Contents
• Preface to the second edition
• Preface to the first edition
• 1 Preliminaries
• 2 Constituent structure
• 3 Syntactic rules
• 4 Syntactic categories
• 5 Heads and complements
• 6 Subjects and predicates
• 7 Anaphora
• 8 Noncanonical complements and subjects
• 9 Grammatical functions
• 10 Passives
• 11 Raising sentences
• 12 Control
• 13 Wh-dependencies
• 14 Island constraints
• 15 Concluding remarks
• Glossary
• Bibliography
• Index