Syntax

A Linguistic Introduction to Sentence Structure

Keith Brown, Jim Miller

• Éditeur: Routledge
• Année: 2020
• ISBN:
  9781000158748
• Édition: 2
• Sujet: Filología
• Sous-sujet: Lingüística

The second edition of this invaluable introductory text takes account of developments in syntactic studies. Dealing with the whole range of syntax, this book explains, in a lucid and approachable way, why linguists have adopted certain solutions to problems and not others.
This book introduces the basic concepts used in the description of syntax, independently of any single model of grammar. Profusely illustrated with diagrams, there are sets of exercises for every chapter which can be used in class or by students working independently.

Part one

Constituent structure

1 Constituent structure

1.1 Determining constituent structure

The analysis of the structure of sentences is traditionally known ‘parsing’. Part of the Oxford English Dictionary’s definition of PARSE is: ‘to resolve (a sentence etc.) into its component parts of speech and describe them grammatically’. In linguistic work each of the ‘component parts’ of a sentence is a ‘constituent’ and the whole procedure is ‘constituent structure analysis’.

In this chapter we concentrate on the two closely related operations suggested by the quotation:

(1) the analysis of the sentence into its constituents;

(2) the grammatical description of these constituents.

We will expect the first operation to tell us how to break a sentence down into its constituent parts, and which strings of words are, and which are not, constituents. Or, seen the other way round, the description will tell us how words can combine to form larger constituents, and how these, in their turn, can form yet larger constituents until we have constructed a sentence.

The second operation will describe how constituents differ from each other, how each type is constructed, how they combine with each other, what order they can, or must, occur in, and so on. This will involve naming the different types of constituent so that we can identify them.

Let us, as an example, consider how we might describe the structure of the sentence:

1 The dog frightened the child

We will make three initial assumptions, all of them justified in due course. First, although we are considering only a single sentence, it is not the only sentence in the language, and we can, indeed must, assume that we can use our knowledge of other potential sentences to guide our analysis. The description we give to this sentence must be compatible with the analysis we would like to make of other sentences we might have chosen; our ultimate aim is the description not of this single sentence but of the language as a whole.

Secondly, we assume that the smallest constituents with which we are concerned are the five words represented in 1. This assumption begs a number of questions. For one thing, it is clearly not quite true, since a word like frightened is recognizable as the ‘past tense’ form of the verb FRIGHTEN and is readily analysable into the smaller constituents frighten and -ed. The analysis is based on the analogy of comparable ‘past tense’ forms of other verbs: kill: kill-ed, walk: walk-ed, cook: cook-ed, and so on. The analysis of words introduces a number of complex issues which we can disregard for our present purposes but to which we will return in Part Two. However, although we will for the time being accept words as unanalysable wholes it will be helpful to introduce three conventions to distinguish between different uses of the term ‘word’ itself. The first is to use italics to quote actual word forms like frightened and killed when they are quoted in the text. The second convention is to use small capitals to refer to ‘dictionary’ or ‘lexical’ entries (the technical term is ‘lexeme’). Thus FRIGHTEN is the lexeme common to all the variant word forms of FRIGHTEN: frighten and frightened, which we have already met, and the other regular forms of FRIGHTEN, frightens and frightening, which we will soon encounter. Our third notational convention arises from the observation that frightened is the ‘past tense’ form of FRIGHTEN. It will sometimes be helpful to use a grammatical description like this and when it is, we will annotate the grammatical category involved as a ‘feature’ subscripted to the relevant lexical item; thus frightened, ‘the past tense form of FRIGHTEN’, will, when necessary, be represented as FRlGHTEN_[past]; similarly killed, the past tense form of KILL, by KlLL_[past]. If we need to refer in general terms to the ‘past tense form of a verb’, we can use the annotation V_[past].

Our third assumption concerns the sentence. We will use the neutral term ‘string’ to refer to any sequence of constituents; so we can refer to the dog, frightened the child, or indeed the whole of 1 as strings with no commitment as to whether they do or do not form a constituent or any identification of the type of constituents they are. What entitles us to refer to the string 1 as a sentence, when the other strings mentioned do not seem to be sentences? Unfortunately, as with the word, there are considerable problems about the identification of sentences; to call a string a sentence implies that it has a certain sort of unity, but it is far from easy to describe exactly what sort of unity a string needs to have for it to be referred to as a sentence. We might regard the study of syntax as an attempt to answer just this question. For the time being we merely accept that 1 is a sentence; again our provisional justification is that it is supposed that we would all agree that 1 is indeed a sentence.

With these three assumptions in mind, let us contemplate the first operation noted at the beginning of the chapter: the analysis of the sentence:

2 The dog frightened the child

into its relevant parts, its constituents.

At an intuitive level the dog, frightened and the child appear to be relevant constituents in a way that strings like frightened the or dog frightened do not. Furthermore, again at an intuitive level, the strings the dog and the child seem to be constituents of the same type. We can back these in...