Analyzing Syntax: A Lexical-Functional Approach / Edition 1

by Paul R. Kroeger

Overview

Designed for students of linguistics at the advanced undergraduate or graduate level, this volume is a comprehensive and accessible textbook on syntactic analysis. Working within the "Lexical Functional Grammar" (LFG) approach, it provides students with a framework for analyzing and describing grammatical structure, using extensive examples from European as well as non-European languages. While its primary focus is on syntactic structure, it also covers aspects of meaning, function and word-structure directly relevant to syntax. It is ideal for one-semester courses in syntax and grammatical analysis.

Product Details

ISBN-13:	9780521016544
Publisher:	Cambridge University Press
Publication date:	04/08/2004
Series:	Cambridge Textbooks in Linguistics
Edition description:	New Edition
Pages:	340
Product dimensions:	6.69(w) x 9.61(h) x 0.91(d)

About the Author

Paul R. Kroeger is Assistant Professor and Chairman at the Graduate Institute of Applied Linguistics, Dallas.

Read an Excerpt

Cambridge University Press
0521816238 - Analyzing Syntax - A Lexical-Functional Approach - by Paul R. Kroeger
Excerpt

1 Three aspects of syntactic structure

Probably no one has ever before said or heard the following sentence, yet any normal adult speaker of English will understand it:

(1) John Adams could have been elected to a fourth term as President, if his step-sister had not been so ugly.

In the same way, a speaker of any language will say and hear many sentences during the course of a normal day which he has never said or heard before. Moreover, other speakers of the same language will not only recognize these original creations as being well-formed sentences but will also (usually) understand what they mean.

These observations tell us something important about the nature of language. A person who knows how to speak a language does not have to memorize every possible sentence in that language. Rather, speakers produce sentences CREATIVELY. Some common phrases and sentences may be repeated so often that they are memorized as a single unit, e.g., idioms and proverbs. But, for the most part, we do not memorize sentences; rather, we construct them when we need them, to express a particular idea.

This creative use of language is possible because the patterns of a language are determined by a set of RULES. A speaker who (unconsciously) "knows" these rules can use them to create and understand any number of new sentences. Linguists use the term GRAMMAR to refer to the set of rules needed to produce well-formed utterances in a particular language. The chief concern of this book will be to help you analyze and describe the kinds of grammatical patterns commonly found in human languages, in particular those aspects of the grammar that are relevant to SYNTAX, i.e., the arrangement of words in a sentence.

1.1 Grammar and grammaticality

Our ultimate goal, as linguists, is to understand the speaker's "internal grammar," i.e., the system of rules which a speaker unconsciously uses in speaking and listening. Since speakers are not aware of using these rules, they cannot simply tell us what the rules are. (As noted below, we are not speaking here of PRESCRIPTIVE rules, like those commonly taught in school. Those would be easier for speakers to describe, because they are consciously learned.) We begin by making observations about what people say, what they do not say, how they interpret what other speakers say, etc. Our description of the grammar, based on careful analysis of these observations, will involve formulating explicit rules that can model (i.e., produce the same grammatical patterns as) the speaker's internal grammar.

Of course, in order to speak and understand a language we must know not only rules but also words. Linguists use the term LEXICON to refer to the set of all the words (or, more generally, meaningful elements) in the language. We can think of the lexicon as the speaker's "mental dictionary." Much of the information in the lexicon is unpredictable, such as the fact that the pronunciation /kæt/ refers (in English) to a small carnivorous mammal with whiskers. But the lexicon contains a number of regularities as well. Precisely how much regularity is a hotly debated question; we will return to this issue in chapter 3. In this chapter, we introduce some basic concepts needed for discussing grammatical systems.

1.1.1 "Grammaticality" and variation

In claiming that sentence (1) is a well-formed English sentence we do not imply that it is either true or sensible. In fact, it is stunningly illogical and historically inaccurate on several counts. (That is why no one else is likely ever to have said it before.) This distinction between the form of a sentence and its meaning is an important one. In some cases we might consider a sentence to be well-formed even when it has no sensible meaning at all. Chomsky (1957) used the following pair of sentences, which have become famous through countless repetitions, to illustrate this point:

(2)

1. Colorless green ideas sleep furiously.

2. Furiously sleep ideas green colorless.

Chomsky claimed: "Sentences [2a] and [2b] are equally nonsensical, but any speaker of English will recognize that only the former is grammatical." The second part of this claim has sometimes been disputed, but the essential point is that speakers have an intuitive sense of "grammaticality," or correctness of form, which does not depend on our ability to interpret the meaning of the sentence.

Chomsky argued that any speaker of English would consider sentence (2a) to be "grammatical," even though it makes no sense. (More specifically, the meanings of the individual words in this sentence are not compatible with each other, and so this combination of word-meanings produces contradictory information.) Conversely, as the following sentences show, we can often understand a sentence perfectly well even if it is not "grammatical":

(3)

1. Me Tarzan, you Jane.

2. Those guys was trying to kill me.

3. When he came here?

When speakers reject the sentences in (3) as being "ungrammatical" or "bad English," they mean that one (or more) of the rules of their internal grammar has been violated. One way in which a linguist tries to understand this internal grammar is to formulate a set of rules which will model the judgments of a native speaker: a set of rules which will produce all of the sentences that speakers consider to be grammatical but none of the sentences that speakers consider to be ungrammatical.

This may sound like a fairly straightforward goal, but there are a number of factors which complicate the process. For one thing, speakers of a given language do not always agree about the grammaticality of particular sentences. Each of the sentences in (4a, b, c), for example, would be perfectly acceptable to English speakers from certain geographical areas, but sound quite odd to English speakers from other areas. This kind of variation among regional dialects is found at least to some extent in most of the world's languages. Sentence (4d) involves a different kind of variation, namely variation among individual speakers. It is not associated with a particular region; even speakers from the same dialect area may differ as to whether they would say such sentences, or accept them as being grammatical.¹

(4)

1. I might could be persuaded to try that.

2. My back door needs fixed.

3. The ship is arriving Monday week.

4. They have come visited us every day this month.

Of course language is a means of communication, and if two speakers have radically different internal grammars, communication will be extremely difficult if not impossible. For many purposes, we can speak of "the grammar" of a language as if it were a body of knowledge which all speakers of that language must share; and to a very large extent this is true. But we must always be aware of the variation among dialects and individuals which this over-simplified view ignores. Indeed, the variation itself is also something that linguists seek to document and explain.

Another challenge to the linguist who wants to describe and model the rule system of a language is that languages are always changing. Moreover, community attitudes about language often change more slowly than the actual practice of the community in speaking the language. Consider the examples in (5).

(5)

1. With two things hath God man's soul endowed.

2. I know not what course others may take, but as for me . . .

3. The problem is, is that no one wants this job.

4. The mission of the Enterprise is to boldly go where no man has gone before.

The word order of (5a) was perfectly normal in Old English (before 1100 AD), but most speakers of modern English will probably consider it extremely odd, if not actually ungrammatical. Example (5b) is taken from a famous speech by Patrick Henry (1775 AD). Similar examples are very common in Shakespeare and the King James Bible (1611 AD), but today they are regarded as archaic or (in modern usage) unnatural.

The construction in (5c) is a much more recent innovation. It is now very widely used in informal speech, at least in American English, but most educated speakers would probably reject it in formal written styles. Moreover, standard reference books on English grammar, school textbooks, etc. do not recognize it as an acceptable way of speaking.

Sentence (5d) is an example of a "split infinitive," because the adverb boldly appears in the middle of the infinitive to go. This pattern emerged in the fourteenth century due to a cluster of morphological and syntactic changes in Middle English, and has been used quite commonly ever since (Hall, 1882; Kiparsky, 1997). However in the nineteenth century, as part of a growing concern for defining "correct" usage in English, influential authorities began to assert that the split infinitive was "bad English" (i.e., ungrammatical), apparently because no such pattern exists in Latin. This judgment is maintained in school textbooks to the present day.

The case of the split infinitive is a notorious example of a PRESCRIPTIVE approach to grammar: grammarians telling other people how to talk. Of course, there are many contexts where a prescriptive approach is appropriate and indeed necessary: e.g., in planning and developing the standard form of a new national language, in teaching adult language learners to speak correctly, in teaching high school students to write acceptable essays, etc. But these areas of "applied linguistics" will not be our primary focus.

The approach we adopt in this book will be DESCRIPTIVE rather than prescriptive. This means that we take it as our goal to observe, describe, and analyze what speakers of a language actually say, rather than trying to tell them what they should or should not say. Most of our examples will come from published sources. As a result, these examples often represent a standardized or high-prestige variety of the language.² But the same approach can be applied to data from non-standard dialects, or even languages which have no written form at all. Indeed, one of the goals of this book is to equip you to undertake this kind of research.

Our approach will be primarily SYNCHRONIC; that is, we will be primarily interested in describing the structure of a particular language or dialect at a particular time (normally the present), rather than comparing it with related dialects or investigating how the language has changed over time.

As we noted in the introduction to this chapter, SYNTAX is the branch of linguistics which seeks to describe and account for the arrangement of words in a sentence. In order to do this we will often need to look at the structure of the words themselves, i.e., at certain aspects of the MORPHOLOGY. For example, in some languages the presence of a certain prefix or suffix on the verb determines where other words in the sentence may or may not occur. Moreover, even though form and meaning are partially independent of each other, as we have seen, they are also intimately connected. So in studying the form of sentences we will often be led to consider their SEMANTICS (i.e., meaning) as well.

Finally, in addition to the form and meaning of individual sentences, we will sometimes need to consider connections between one sentence and another, or the function of a particular sentence in a specific context. (The study of these aspects of language use is called PRAGMATICS.) And we may occasionally mention stylistic or social factors which influence the kind of sentence patterns a speaker might use; but our primary focus will be on the sentence patterns themselves. Morphology, semantics, pragmatics, and sociolinguistics are major fields of study in their own right, but in this book we will touch on them only where they are directly relevant to the syntactic issues that we consider.

1.1.2 Sentence structure

A sentence is not simply a string of words, one after another. An overwhelming body of linguistic and psycho-linguistic evidence shows that speakers think of sentences as having a fairly complex structure, with certain words grouped together to function as units, larger groups formed from smaller groups, and important relationships defined between one group and another.

Of course, these structural relationships are "invisible," because all we hear is the string of words. In some ways the task of the linguist is similar to that of early chemists. By observing changes in the physical properties of substances when they were combined in various ways, these chemists were able to develop theories about the unseen structures of atoms and molecules which could account for their observations. In the same way, the linguist tries to understand unseen linguistic structures based on observations about what can be heard.

One reason for thinking that sentences do in fact have this kind of abstract structure is that a given string of words may be AMBIGUOUS, i.e., allow two different interpretations, even when none of the individual words in the string is itself ambiguous. Consider the following examples (adapted from Huddleston, 1984).

(6)

1. Liz attacked the man with a knife.

2. Ed likes Sue more than Jill.

3. The proposal that Hitler was advancing seemed preposterous.

The meaning of example (6a) depends on the structural relationships of the phrase with a knife. This sentence could be used to answer two different questions: Who did Liz attack? The man with a knife or What did Liz attack the man with? With a knife.

The meaning of example (6b) depends on the relationship of the word Jill to the verb like: is Jill to be understood as the subject of like (as in Ed likes Sue more than Jill likes Sue), or as the object of like (as in Ed likes Sue more than Ed likes Jill)? The meaning of example (6c) depends on the relationship between the word proposal and the phrase that Hitler was advancing. Is the proposal something which is being advanced by Hitler? Or is Hitler himself supposedly advancing, and the proposal merely a report (by some other person) of this event?³ Such examples of STRUCTURAL AMBIGUITY can provide important evidence about the structures which speakers assign to sentences.

Languages show great variety in terms of the strings of words that they use; that is one of the reasons why word-by-word translation fails so miserably. But when we compare abstract structural properties, languages turn out to be much more similar to each other than we might have guessed. In some respects the variation among languages is surprisingly limited. For this reason, studying data from a wide range of languages can be a great help to us in knowing what to look for when we tackle a new language, particularly if that language has not been analyzed before.

1.2 Outline of a framework

The job of the syntactician can be divided into two main steps: first, determine the correct structure(s) for each grammatical sentence; and second, formulate a set of rules which will distinguish between grammatical and ungrammatical structures for that language (i.e., allow us to predict which structures will be grammatical and which ungrammatical). Of course, the two tasks are closely related, and we will be concerned with both. But we will devote most of our attention in this book to the first of these issues, in particular to the kinds of evidence that are relevant for determining linguistic structure.

In order to discuss the details of syntactic structure with any kind of precision we will need to develop (i) a technical vocabulary; (ii) a system for representing structural relations; and (iii) a set of concepts which are relevant to this task. Such an inventory of vocabulary, notation and concepts is called a syntactic FRAMEWORK.

A good framework must do at least three things. First, it should make it easy to describe the syntactic patterns found in any particular language. Second, it should make it easy to compare syntactic patterns between languages. Third, it should allow us to make generalizations about human language in general, i.e., to state THEORIES (factual claims about how language works).

In the remainder of this chapter, we will sketch out the beginnings of a framework for syntactic analysis. As a way of introducing some of the concepts which we will need to use in talking about syntactic structure, let us first think about structural complexity. What makes a sentence "complex"? Which of the following Malay sentences is the most complex? Try to rank them, from simplest to most complex.

(7)

1. Dia mandi.
   3sg bathe
   'He is bathing.'

2. Saya makan nasi.
   1sg eat cooked.rice
   'I eat/am eating rice.'

3. Orang tua itu makan nasi goreng setiap hari.
   person old that eat rice fry each day
   'That old person eats fried rice every day.'

4. Dia belajar untuk menjadi pensyarah.⁴
   3sg study in.order become lecturer
   'He is studying to become a lecturer.'

One way to measure complexity, though perhaps not the most revealing way, is by absolute length (number of words). On this basis, (7a) is clearly the simplest and (7c) the most complex. Of course, there are other, and more persuasive, reasons for considering (7a) to be the simplest. Semantically it names a single event, 'bathe', which involves just one participant. Sentence (7b), on the other hand, names an event which involves two participants, the "eater" and the "eaten."

Sentence (7c) names an event of exactly the same type as (7b). It uses the same verb (makan 'eat') and involves the same number of participants playing the same roles. So in this respect we might evaluate them as being equally simple. But (7c) also contains an additional piece of information, namely the fact that this event occurs 'every day.' This time phrase is, in a sense, added on to the basic description of the event.

In discussing the meaning of a clause, we will use the term PREDICATE to refer to the word which names the action, event, or state described by that clause. Typically this word will be a verb.

Now any event named by the predicate 'eat' must involve at least two participants, the "eater" and the "eaten." (This is true even though one or the other of these participants may not be mentioned in a particular description of the event, e.g., John is still eating or The fish was eaten.) For this reason we say that the predicate 'eat' takes two ARGUMENTS. But a time phrase like 'every day' is not an inherent part of the meaning of 'eat.' This kind of phrase can be added freely to virtually any clause that describes an event. An extra piece of information of this kind, something that is not an argument of the predicate, is called an ADJUNCT.

The ARGUMENT STRUCTURE of a predicate is a representation of the number and type of arguments associated with that predicate, as illustrated in (8). We will use the general term AGENT to represent the participant who performs a certain action, and the term PATIENT for the participant that something happens to; see section 1.2.1 for further discussion and examples.

(8)	bathe < agent >
	eat < agent, patient >

Argument structure is important to the syntax, because it determines many of the basic grammatical properties of the clause in which the predicate occurs. Argument structure is closely related to meaning, but it is obviously not intended to represent the full meaning of a sentence, or even of a predicate. For example, it is true that (7b) and (7c) have the same argument structure; but that does not mean that the meaning of the two sentences is equivalent. One reason for this is that adjuncts are not a part of the argument structure. Another reason is that the argument structure indicates only the role which each argument plays in the event, but does not give any information about the inherent properties of the arguments themselves, e.g., the fact that the agent in (7c) is an old person.

This last point brings us back to another important difference between (7b) and (7c). Sentence (7b), repeated below as (9a), consists of three basic elements: a subject, a verb, and a direct object. (What we mean by "subject" and "direct object" will be discussed in section 1.2.3.) Each of these elements is named by a single word. Sentence (7c) (ignoring the adjunct phrase) has the same three basic elements; but here the subject and object are each expressed by a phrase containing more than one word, as shown by the brackets in (9b).

(9)

1. [Saya] makan [nasi].
   1sg eat cooked.rice
   'I eat/am eating rice.'

2. [Orang tua itu] makan [nasi goreng].
   person old that eat rice fry
   'That old person eats fried rice.'

The fact that groups of words can function as units (or CONSTITUENTS) within sentences is an important aspect of the grammar of every human language. In analyzing the structure of a sentence, it is very important to identify the constituent boundaries (i.e., to determine which words group together as units), to specify the linear order of constituents in the sentence, and to specify the ordering of the words within each constituent. The aspect of syntactic structure which represents these kinds of information is called CONSTITUENT STRUCTURE (or PHRASE STRUCTURE).

Now let us return to sentence (7d), repeated below as (10). This sentence is shorter than (7c), in that it contains fewer words; and it involves only a single participant (the one who is studying). So at first glance it may look simpler than (7c). But the meaning is more complex. This sentence actually describes two events, 'studying' and 'becoming a lecturer.' The two events have a certain logical relationship to each other: the agent does X in order to achieve Y. Each of these events is named by a distinct predicate, belajar 'study' and menjadi 'become,' and each predicate has its own argument structure.

(10)	Dia belajar untuk menjadi pensyarah. (= 7d)
	3sg study in.order become lecturer
	'He is studying to become a lecturer.'

Corresponding to this semantic complexity, the grammatical structure of (7d) is considerably more complex than that of (7c). This kind of sentence structure will be discussed in some detail in chapter 5. For now we will just observe that in (7d) we find one CLAUSE (or simple sentence) embedded within another. That is, one clause functions as a constituent of another, specifically in this case as an adjunct. Moreover, the subject of the embedded clause is understood to be identical with the subject of the main clause. The function of the embedded clause within the larger sentence, and the relationship between the subject of the embedded clause and the subject of the main clause, are part of the FUNCTIONAL STRUCTURE of the sentence.

We have mentioned three aspects of the structural complexity of sentence (7d): argument structure, constituent structure, and functional structure. These three aspects of syntactic structure will be important for our analysis of a wide variety of constructions; so in the remainder of this chapter we will briefly review some of the basic features of each.

1.2.1 Argument structure

As stated above, argument structure is a representation of the number and type of arguments associated with a particular predicate. Determining the number of arguments is not always as easy as one might expect; but identifying the "type" of these arguments may seem even more difficult. What exactly do we mean by this?

In fact this question has been a hotly debated issue among linguists. The approach which we will adopt here is to assign participants to broad semantic or conceptual categories according to the role they play in the described event or situation: "agent" for participants that do something; "patient" for participants to whom something is done; "experiencer" for participants who think or feel something, etc. Unfortunately (but not surprisingly) there is no one set of semantic role labels which all linguists agree on, and different linguists sometimes use the same labels in different ways. But in this book we will refer to (at least) the following semantic roles:

(11)	INVENTORY OF SEMANTIC ROLES:
	AGENT: causer or initiator of events
	EXPERIENCER: animate entity which perceives a stimulus or registers a particular mental or emotional process or state
	RECIPIENT: animate entity which receives or acquires something
	BENEFICIARY: entity (usually animate) for whose benefit an action is performed
	INSTRUMENT: inanimate entity used by an agent to perform some action
	THEME: entity which undergoes a change of location or possession, or whose location is being specified
	PATIENT: entity which is acted upon, affected, or created; or of which a state or change of state is predicated
	STIMULUS: object of perception, cognition, or emotion; entity which is seen, heard, known, remembered, loved, hated, etc.
	LOCATION: spatial reference point of the event. The LOCATION role includes the sub-types SOURCE, GOAL, and PATH, which respectively describe the origin (or beginning-point), destination (or end-point), and pathway of a motion
	ACCOMPANIMENT (or COMITATIVE): entity which accompanies or is associated with the performance of an action

Some examples of the most common of these roles are given in (12):

(12)

1. John gave Mary a bouquet of roses.
   AGENT RECIPIENT THEME

2. John baked Mary a chocolate cake.
   AGENT BENEFICIARY PATIENT

3. John opened the lock with a key.
   AGENT PATIENT INSTRUMENT

4. The key opened the lock.
   INSTRUMENT PATIENT

5. Sherlock Holmes hearda piercing scream.
   EXPERIENCER STIMULUS

For many verbs, it is not too difficult to identify the arguments with one or another of these roles; some examples are given in (13). But in other cases a particular argument may not seem to fit naturally into any of these categories; for example, what is the semantic role of mother in Susan resembles her mother? Or an argument may appear to bear two roles; for example, John seems to be both an agent and a theme in John jumped into the well.

(13)	dance <agent>
	eat <agent, patient>
	love <experiencer, stimulus>
	give <agent, theme, recipient>

These kinds of issues have been discussed at considerable length, but for our purposes they do not represent a major problem. The primary function of these role labels is to allow us to distinguish among the arguments of a particular predicate.⁵ Thus some linguists prefer to use unique labels for each predicate, as we did above in referring to the arguments of eat as the "eater" and the "eaten." But it is also true that the categories in (11) are quite useful for describing a wide variety of grammatical patterns. That is why the use of these labels remains so popular, in spite of certain well-known problems.

Notice that the list in (11) is restricted to ARGUMENT roles. Some other commonly expressed types of semantic information, e.g., time, manner, purpose, etc. are not included here, because the elements which express these concepts are almost always ADJUNCTS rather than arguments. This distinction between arguments and adjuncts is important, but not always easy to make. The basic difference is that arguments are closely associated with the meaning of the predicate itself, while adjuncts are not.

Adjuncts contribute to the meaning of the sentence as a whole, but are never necessary to complete the meaning of the predicate. Thus adjuncts are always optional, whereas arguments are frequently obligatory. Sentence (14a) shows that the object of use is obligatory, and therefore an argument. As (14b) illustrates, even when an argument is grammatically optional it may be semantically obligatory; for example, even when the patient of eat is not expressed, we know that something gets eaten. But adjuncts can always be omitted without creating this kind of implication, as seen in (15).

(14)	ARGUMENTS
a	Mary used my shirt for a hand towel.
	*Mary used for a hand towel.
b	John ate an apple.
	John ate. (implies that John ate something)

---

© Cambridge University Press

Table of Contents

1. Three aspects of syntactic structure; 2. Identifying constituents and categories; 3. Passives, applicatives, and the 'dative shift'; 4. Reflexives; 5. Control; 6. Pragmatic functions: topic and focus; 7. Filler-gap dependencies and relativization; 8. Causative constructions; 9. Serial verbs and related constructions; 10. 'Quirky case' and subjecthood; 11. Syntactic ergativity.