An introduction to english morphology

Go and went are also different word forms that realize a single LEXEME [GO].. A single word may be composed of one or more morphemes: nation 1 morpheme, nation.al 2 morphemes, nation.al

Hue University College of Foreign Languages

Hue University College of Foreign Languages

Can we always tell precisely what a word is? Do motet, motion and motive have anything to

do with each other? What ways do we have of making new words in English? Are the same ways of forming new words found in all languages? Is it just coincidence that although you

can have a word like people which means much the same as ‘a lot of persons’, and a word

peoples which means, more or less, ‘a lot of lots of persons’, you cannot have a word personss meaning the same thing? Is it just coincidence that the ablative plural of the Latin

word re:x ‘king’, re:gibus, meaning ‘by/ from/ with the kings’ is so much longer than the

nominative singular re:x? (I use the phonetic length mark rather than the traditional macron to

show long vowels in Latin.) All of these questions relate to morphology, the study of words

and their structure

It is a well-established observation that words occur in different forms It is quite clear to anyone who has studied almost any of the Indo-European languages Students of these languages learn paradigms like those below as models so that they can control the form-changes that are required As illustrations, consider a verb paradigm from Latin and a noun paradigm from Icelandic (The word ‘paradigm’ means ‘pattern’ or ‘example’.)

(1) amo; ‘I love’

amais ‘you (singular) love’

amat ‘he/she/it loves’

amaitis ‘you (plural) love’

amant ‘they love’

we might say that Latin has a more complex morphology than English Again by extension,

‘morphology’ is also used for the sequence of rules which are postulated by the linguist to account for the changes in the shapes of words In this sense we might contrast the morphology of language L with the syntax of language L (where the syntax is the sequence of rules postulated by the linguist to account for the ways in which words are strung together) In this sense we might also say that something is part of the job of ‘the morphology of language L’ or, more generally, of ‘morphology’, implying that this is true for all languages We shall see later how all these senses fit together; such extensions of meaning are common within linguistics, and do not usually cause problems of interpretation

Many traditional ‘grammars’ (in the sense ‘grammar books’) deal largely with such morphology as can be laid out in paradigms like those presented above, and have little to say about syntax This has led to the situation where many lay people today still believe that languages like Chinese or English do not have much grammar, because they do not have extensive morphological paradigms That is, for many people the term ‘grammar’ is equated with morphology For most linguists today, however, ‘grammar’ includes both morphology and syntax, and most of the linguistic study of ‘grammar’ in this sense has, since the middle

of this century, not been of morphology, but of syntax This is understandable Syntax, especially from 1957 onwards, was a relatively new field of study, while morphology was considered well-researched and well-under-stood It did not seem at that time as if there was a great deal that was new to say about morphology Morphological descriptions of hundreds of languages were available, but all the languages differed in what appeared to be essentially random ways There did not seem to be any cross-linguistic generalizations to be made in morphology Syntax, in the middle of this century, was a far richer ground for linguistic discoveries It was the excitement of the progress being made in the study of syntax which gave Linguistics such a boost in the 1960’s It was also progress in the study of syntax which eventually led to the realization that there were still questions to be answered in morphology

As a result, there has in recent years been a resurgence of interest in morphology

The theoretical background to this new interest in morphology comes from three distinct sources Firstly, there is the philological study of grammar in the last century and the early years of this century Secondly, there is the study of diverse languages under the influence of one or another of the structuralist schools of Linguistics In particular the work of the American structuralists, especially Bloomfield and his followers, is important here Finally, there is the influence of transformational grammar and the school of thought that emerged from the work of Chomsky It is not always easy to separate out these three strands in current morphological theory, and sometimes one dominates, sometimes another Nonetheless, all three influences can be strongly felt This book provides an introduction to the study of

3

morphology covering the input from these various sources, and attempting some kind of synthesis in the light of the most recent research It discusses both the general background to all morphological study, and also some of the detail of recent theories of morphology

(Laurie Bauer 1992: 3-5)

As with any other area of linguistic theory, we must distinguish between general

morphological theory that applies to all languages and the morphology of a particular language General morphological theory is concerned with delimiting exactly what types of

morphological rules can be found in natural languages The morphology of a particular language, on the other hand, is a set of rules with a dual function First, these rules are

responsible for word formation, the formation of new words Second, they represent the speakers’ unconscious knowledge of the internal structure of the already existing words of

their language

2 Definition

Morphology is the study of internal structure of words and of the rules by which words are formed

Deinstitutionalization: practices of releasing patients from hospitals for the mentally ill

Reinstitutionalization: practices of returning them to these institutions

By means of morphological rules we all understand that the above two words are derived

from the root institution and the affixes de-/re-, -al, - ize, -ation

Questions:

1 How is morphology of a particular language understood/meant?

2 What is meant by English morphology?

76

APPENDIX 1

The following is chapter 3 extracted from Laurie Bauer Introducing Linguistic Morphology

(1988: 19-41) It aims at providing readers with more detailed information on morphological structure of words

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

(Laurie Bauer 1992: 19-41)

4

CHAPTER 2

MORPHEME AND TYPES OF MORPHEME

1 Words

In any science, one of the basic problems is to identify the minimal units, the basic parts out

of which more complex units are constructed In language, we must distinguish the basic units

of sound, which in themselves are meaningless, from the basic meaningful units, which are made up of individually meaningless sounds

Most people, if asked what the minimal meaningful units of language are, would have a ready

answer - words Indeed, of all the units of linguistic analysis, the word is the most familiar

In fact, its existence is taken for granted by most of us We rarely have difficulty picking out the words in a stream of speech sounds or deciding where to leave spaces when writing a sentence But what, precisely, is a word? A word need not have any special phonetic properties: some words bear stress but others do not; some words are set off by intonational cues but others are not The two syllables in the following examples have exactly the same pronunciation even though they are separate words in the first case but part of the same word

in the second case

a door - adore

It is also difficult to distinguish words from other linguistic units in terms of the types of

meaning they express Bachelor and unmarried adult male, for example, seem to have the same meaning, even though one is a word and the other a phrase Similarly, builder and

someone who builds mean about the same thing even though one is a word and the other a

A free form is an element that can occur in isolation and whose position with respect to

neighboring elements is not entirely fixed Thus, we would say that hunters is a word since it can occur in isolation (as in answer to the question, Who are they?) and can occur in different

positions within the sentence, as shown below:

The hunters chased the elephant

The elephant chased the hunters

In contrast, the units -er and -s do not count as words here since they cannot occur in isolation

and their positioning with respect to adjacent elements is completely fixed Thus, we cannot

say *erhunts or *serhunt, only hunters

The reference to minimal above is necessary to ensure that we do not identify phrases such as

the hunters as a single word Although this unit can occur in isolation and can occupy

different positions, it is not a minimal free form since it consists of two smaller free forms -

the and hunter (We know that the is a separate word because its positioning with respect to hunters is not entirely fixed; thus, another word can appear between the two, as in the courageous hunters.)

5

How many words are there in

1 The cook was a good cook as cooks go, and as cooks go, she went

2 I’ve been in hot water so often I feel like a tea-bag

How many different words does (1) contain?

11 different word forms: the, cook, was, a, good, cooks, go, and, as, she, went

On the other hand, there is another sense in which cook and cooks are different forms of ‘the

same word [COOK]’ (i.e LEXEME) Go and went are also different word forms that realize

a single LEXEME [GO]

A Lee walked home

B Lee has walked home

The word walked in A & B realizes the same lexeme [WALK] Yet walked in these two sentences is not precisely the same element In A, walked realizes WALK + past tense, while

in B, it realizes WALK + past participle We might want to say that walked in A & B are

different words, even though they realize the same lexeme [WALK]

In the printed text, words are separated out for us But even if they were not, we would still

be able to discover the beginnings and the ends of words fairly simply

Menbecomeoldbuttheyneverbecomegood

Menareconservativeafterdinner

Menlosetheirtempersindefendingtheirtaste

Afterfortymenhavemarriedtheirhabits

2 Elements smaller than the word

Words, though they may be definable as minimal free forms, are not the minimal meaningful units of language we are looking for, since they can often be broken down further The word

hunters, which as we have just seen can stand alone and is thus a free form, nonetheless

consists of three meaningful parts: hunt, er, and s

Now consider the following sentence:

(1) He was born stupid, and greatly increased his birthright

We could isolate all the word-forms in this sentence in the ways outlined above, but we can also look within the word-forms and isolate recurrent forms within the word-form

For instance, if we consider the word-form birthright, we can divide that into two parts; for

the first part, birth we could substitute within the word-form things like copy and water, for the second part, right, we could substitute things like day, place and rate

In a similar way, we could divide the word-form greatly up into two parts For the first part,

great, we could substitute other forms like vast and incredible, and within the word-form

(though not within this particular sentence) we could substitute items like ness, er, est for

-ly

The units which we arrive at within the word-form we will call morphs A word-form may

contain only one morph (stupid, and) or it may contain several (great-ly, increase-d,

birth-right)

A morph, then, is a unit which is a segment of a word-form It has a constant form and realizes or is related to a constant meaning

Some morphs have the potential of being word-forms on their own In

(2) Every-one live-s by sell-ing some-thing

this applies to the morphs every, one, live, by, sell, some and thing Such morphs are called

potentially free morphs Notice that the potentiality is not actually exploited for all of these

morphs in (2) Indeed, only by is actually free in (2), but the others listed are potentially free

Morphs which cannot be word-forms by themselves but which need to be attached to other

morphs are termed obligatorily bound morphs In (2) only the morphs -s and -ing are

obligatorily bound

3 Allomorphs and morphemes

Sometimes two or more morphs which have the same meaning are in complementary

distribution That is, the two can never occur in precisely the same environment or context,

and between them they exhaust the possible contexts in which the morpheme can appear For

example, there are two morphs in English which can be glossed as 'indefinite article': a and

an Some examples of their distribution can be seen below

In this case the choice between the two morphs a and an is determined or conditioned by the

following phonetic sound We can say that their distribution is phonetically conditioned

In other cases the distribution of morphs may be determined by other factors For example,

there are various ways of marking plurality in English The most common way is with an -s (variously pronounced), there is -en in ox-en, -ren in child- ren, and a few other ways as well

The choice of these various ways does not depend on the phonetic environment, but on the

WORD involved It is a peculiarity of the word OX (as opposed to BOX, COX, FOX) that it takes a plural marker -en No other word in English marks its plural in just this way The

plural marker -en is determined by the particular WORD involved: it is lexically conditioned

7

Morphs can also be grammatically conditioned In languages such as German, or Russian,

adjectives change their form depending on the gender and the number of the noun they modify

But if English a and an, or -s, -en, -ren … are clearly separate morphs because of their

different shapes, they nevertheless have things in common They have their meaning in common: ‘indefinite article’ or ‘plural’ between them, they divide up a single distribution: always before a singular countable noun, or always on the end of a countable noun In the

clearest cases, they even have a similarity in form There is a sense, therefore in which a and

an are ‘the same thing’ We will say that these various sets of morphs realize the same

morpheme

Morphemes, then, are abstract units They are the minimal (smallest) meaningful units of a

language Morpheme is an arbitrary union of a sound and a meaning and cannot be further analyzed (into meaningful units) A single word may be composed of one or more morphemes: nation (1 morpheme), nation.al (2 morphemes), nation.al.ize (3 morphemes), nation.al.iz.ation (4 morphemes), …

Morphs which realize a particular morpheme and which are conditioned (whether

phonetically or lexically or grammatically) are called the allomorphs of that morpheme

A and an are the two phonetically conditioned allomorphs of the (indefinite) morpheme [a]

Allomorphs can be understood as any of the different forms of a morpheme (i.e alternative

representations of a morpheme) Notice that every allomorph is a morph

In most cases in English any morph which can realize a word (lexeme) and which is not further analyzable (except in terms of phonemes) is termed a ROOT Obligatorily bound

morphs which do not realize words (lexemes) and which are attached to roots to produce

word-forms are called AFFIXES (The basic single morpheme form to which affixes are

attached is called a ROOT)

In a word like dealings, deal is the root, and -ing and -s are affixes In something in sentence

(2) there are two roots

Note that this implies that some and thing in something realize the words SOME and THING,

respectively, even though SOMETHING is also a word in its own right

Most roots are free morphemes, but some are not: Euro-crat, octo-pus, phil-anthrop-y,

phonet-ic, quadra-phon-ic, wis-dom, and so on

Affixes can be added directly to a root, as in fool-ish or they can be added to a root and some already attached affix, as is the case with -ness in fool.ish- ness

We can call anything we attach affixes to, whether it is just a root or something bigger than a

root, a BASE

So in the formation of dealings the root is deal, but the base to which the -s is added is

dealing Note that in this case the final -s was not added to a root

8

Any lexical unit which an affix is attached to is called a base

-ize attached to a Adjective-base to form Verb national-ize

-ion attached to a Verb-base to form Noun denationaliz-ation

-al attached to a Noun-base to form Adjective denationalization-al

If an affix is attached before a base it is called a prefix, if it is attached after a base it is called

a suffix, and if it is attached in the middle of a base it is called an infix In the word prepacked

there is a root pack, a prefix pre-, and a suffix -ed All of the affixes that have been illustrated

have been suffixes, which are more common in English than prefixes are There are no infixes

in English

STEM is the actual form to which an inflectional affix is attached to Stem may consist of

more than one morpheme For example, the form sing is the stem of the word sing.s, but singer is the stem of the word singer.s

4 Free and Bound Morphemes

The analysis of morphological structure of words is based on the distinction between free morphemes, which can constitute a word by itself, and bound morpheme, which must be

attached to another element The morpheme learn, for example is free since it can be used as

a word on its own; the morpheme -er, as in learner, on the other hand is bound since it cannot

stand on its own but must be attached to other elements such as teach, sing, work, etc Bound

morphemes are never words but always parts of words

5 Derivation and Inflection

5.1 Derivational morphemes

Derivational morphemes are (bound) morphemes that add new meaning to an existing word

They are called derivational morphemes because when they are conjoined to other morphemes (or words), a new word is derived, or formed

+ the derived words may be in a different grammatical class: nation.al

+ the derived words may be in the same grammatical class:

Inflectional morphemes are bound morphemes that are purely grammatical markers,

representing such concepts as tense, number, gender, person, case, …

9

Inflectional morphemes do not change the syntactic/grammatical category of the word or morphemes to which they are attached On the other hand, they modify a word’s form in order to mark the grammatical subclass to which it belongs, i.e complete the grammatical category of the word form

Sing - sings - sang - sung - singing ( ≠ singer)

Reading

2.3 Inflection and derivation

Affixes can be of two kinds, inflectional or derivational An inflectional affix is one which produces a new word-form of a lexeme from a base A derivational affix is one

which produces a new lexeme from a base Take a word-form like recreates This can

be analyzed into a prefix re-, a root create, and a suffix -s The prefix makes a new lexeme recreate from the base create But the suffix -s just provides another word-form

of the lexeme recreate The prefix re- is derivational, but the suffix -s is inflectional In

English (though not in every language) prefixes are always derivational Suffixes in

English, though, may be either derivational or inflectional In the word-form formalizes the root is form and there are three suffixes: -al, -ize and -s Formal belongs to a different lexeme from form, so -al is a derivational suffix; formalize belongs to a different lexeme from formal, so -ize is a derivational suffix; but formalizes belongs to the same lexeme as formalize, so -s is an inflectional affix There are a number of ways

of telling whether a suffix is inflectional or derivational if you are not sure whether or not it produces a new lexeme

(a) If an affix changes the part of speech of the base, it is derivational Affixes which do not change the part of speech of the base are usually (though not invariably) inflectional

So form is a noun, formal is an adjective; -al has changed the part of speech; it is thus a derivational affix Formal is an adjective, formalize is a verb; -ize has changed the part

of speech; it is a derivational suffix Formalize is a verb, formalizes is still a verb; -s has

not changed the part of speech; -s is likely to be an inflectional affix Note, however, that while all prefixes in English are derivational, very few of them change the part of speech of the base

(b) Inflectional affixes always have a regular meaning Derivational affixes may have an irregular meaning If we consider an inflectional affix like the plural -s in word-forms

like bicycles, dogs, shoes, tins, trees, and so on, the difference in meaning between the

base and the affixed form is always the same: 'more than one' If, however, we consider

the change in meaning caused by a derivational affix like -age in words like bandage,

cleavage, coinage, dotage, drainage, haulage, herbage, mileage, orphanage, peerage, shortage, spillage, and so on, it is difficult to sort out any fixed change in meaning, or

even a small set of meaning changes

(c) As a general rule, if you can add an inflectional affix to one member of a class, you can add it to all members of the class, while with a derivational affix, it is not generally

possible to add it to all members That is, inflectional affixes are fully productive, while

derivational affixes are not For example, you can add -s to any non-modal verb in

English to make the 'third person singular of the present indicative', but you cannot add

-ation to any non-modal verb to make a noun: n-ationaliz- -ation is a perfectly good word,

10

so it works some of the time, but none of* corn (e)- ation, * inflect-ation, *produc(e)~

ation or *walk-ation are words of English We can summarize this criterion in the

following way: affixes which show limited productivity with large numbers of gaps are derivational; affixes which are fully productive (can be used with all members of a class) may be either inflectional or derivational In fact, the distinction between inflectional and derivational affixes is more complex than this suggests, and the matter will be taken up again in Chapter 6 The criteria provided here, though, will cover most

of the straightforward cases

2.4 Allomorphs and morphemes

Sometimes two or more morphs which have the same meaning are in complementary distribution That is, the two can never occur in precisely the same environment or context, and between them they exhaust the possible contexts in which the morpheme can appear […]

… the choice between the two morphs a and an is determined or conditioned by the

following phonetic sound We can say that their distribution is phonetically

conditioned In other cases the distribution of morphs may be determined by other

factors […] The plural marker -en is determined by the particular lexeme involved: it

is lexically conditioned Morphs can also be grammatically conditioned In a

language like German, adjectives change their form depending on the gender of the noun they modify Thus in the nominative singular, we find the following pattern:

(14)

ein gross-er Wagen ‘a big car (masculine)’

ein gross-er Fisch ‘a big fish (masculine)’

ein gross-es Haus ‘a big house (neuter)’

ein gross-es Tier ‘a big animal (neuter)’

eine gross-e Feder ‘a big feather (feminine)’

eine gross-e Schlange ‘a big snake (feminine)’

The suffix on the adjective is not determined by the phonetic shape of the base or of the next word, nor is it determined by the particular lexeme following, but by the grammatical gender class that lexeme belongs to The conditioning is thus neither

phonetic nor lexical, but grammatical But if English a and an, or -s, -en, -ren and -im,

or German -er, -es, and -e are clearly separate morphs because of their different shapes,

they nevertheless have things in common They have their meaning in common:

‘indefinite article’, ‘plural’ or ‘nominative singular’ Between them, they divide up a single distribution: always before a singular countable noun, always on the end of a countable noun, or always on the end of an attributive adjective following an indefinite determiner In the clearest cases, they even have a similarity in form There is a sense,

therefore in which a and an (and the other sets) are ‘the same thing’ We will say that

these various sets of morphs realize the same morpheme

[…]

Notice that the morpheme, like the lexeme and the phoneme, is realized by something else You cannot hear a morpheme or say a morpheme (just as you cannot say or hear a lexeme or a phoneme): you can only say or hear something which realizes a morpheme

11

(or a lexeme or a phoneme) You can hear or say a morph (or a word-form or a phone), but not what it realizes Morphemes (like lexemes and phonemes) are abstract units The morpheme {a} is whatever all the morphs which can realize {a} have in common

Morphs which realize a particular morpheme and which are conditioned (whether

phonetically or lexically or grammatically) are called the allomorphs of that morpheme

(If we consider the written form of the language, it is also possible to talk about

orthographically conditioned allomorphs of a morpheme, as in come and com-, the latter

of which occurs in coming.) A and an are the two phonetically conditioned allomorphs

of the morpheme {a} Notice that every allomorph is a morph The term allomorph is simply more informative than is morph on its own, because it says that the morph is one

of several realizations of the same morpheme

It should be noted that the terms ‘morpheme’ and ‘allomorph’ are frequently used by other linguists in rather different senses from the ones they have been given here While the meanings used in this book are relatively common, they are not in universal use, and you will have to take care when you meet these terms in other works that you know precisely what is intended

This abstractness of morphemes frequently causes problems of understanding for students who are new to morphology

A parallel might be helpful Consider the various symbols presented below:

(15) R r r

What do these symbols have in common? The answer, fairly simply, is that they are all

kinds of ‘R’ What is it about them that shows, then, that they are all kinds of ‘R’? How

can you tell that they are ‘R’s? The answer lies not in their particular shape, which varies depending on the kind of script they happen to occur in The answer lies in their

function; they all function as ‘R’, they all have the same value in the system, they all,

for instance, can be used at the beginning of the word rat But this is an abstract quality

The ‘R-ness’ of the symbols presented in (15) is something abstract The ‘R-ness’

corresponds, in the analogy, to the morpheme The different forms for these ‘R’s which

have the same function in the system correspond to morphs The symbol ‘r’ is just one

of the morphs that can realize the morpheme, the abstract notion of what ‘R’ is

2.5 Recapitulation

Armed with all this terminology, consider the word-form was in the sentence

(16) While he was not dumber than an ox, he was not any smarter

Was is a form which is probably not further analyzable into morphs This

word-form realizes, among other things, the lexeme BE, as we can see if we change the syntactic environment of (16), but not its meaning:

(17) Being not dumber than an ox, but being no smarter, he was not a brilliant conversationalist

(18) I hope, at least, to be not dumber than an ox

12

But there are also other morphemes realized by the word-form was Firstly, was is singular, because if it had a plural subject it would be replaced by were:

(19) While they were not dumber than oxen, they were not any smarter

You cannot say in standard English

(20) * While they was not dumber than an ox, they was not any smarter

Secondly, was also marks past tense You cannot say

(21) While he was not any dumber than an ox, he was not any smarter at the moment You have to say

(22) While he is not any dumber than an ox, he is not any smarter at the moment

In other verbs, pastness is shown by a separate morph (as, for example, in the difference

between compliment and compliment- ed), but here this difference is included in the single word-form So was is a single morph, but, realizes not only the lexeme BE (which

contains a single morpheme {be}), but also the morphemes {singular} and {past tense}

A morph which realizes more than one morpheme in this way is called a portmanteau

morph

To summaries, we can say that an actual speech event can be analyzed into a series of word-forms, some of which will be made up of a single morph, others of which will be made up of more than one morph The morphs realize (though not necessarily in a one-to-one manner) morphemes Morphemes are abstract units of grammatical and semantic analysis One or more morphemes may form a lexeme, an abstract vocabulary item This

is presented in tabular form (with a certain amount of simplification) in fig 2.1

- LEXEMES made up of one MORPHEMES

1 What is meant by word? How do you understand the term word?

2 What is morpheme? What is morph?

3 How do morphemes relate to word?

4 How are English plural morphemes represented?

5 How are morphemes classified? What are they?

6 What are derivational morphemes and inflectional morphemes?

13

7 What is root? stem? and base?

8 When can (and can’t) a root be a stem and/or a base, and vice versa?

9 What is portmanteau morph? Give example to illustrate

10 How many words are there in the sentence He cut the meat and got a deep cut on the

finger

11 How many different words are there in the above sentence?

12 How many morphs and morphemes are there in the above sentence?

14

CHAPTER 3

WORD CLASSES

A natural first step in a scientific approach to words is to seek to establish the different types

of words which appear in languages That such information is readily available to native speakers and, furthermore, is predictively useful for them is easy to demonstrate Suppose, for instance, that you hear the sentence in (1):

(1) A plingle has arrived

Of course, you don't know what plingle means, but you can immediately infer that plingle is the sort of expression which occurs in the constructions the plingle, two plingles, every plingle

which has ever existed, etc In short, (1) enables you to assign plingle to a particular class of

words, and once you know what class of words it belongs to, you know a great deal about its potential for occurrence within the language It is reasonable, then, to suppose that the word class to which a word belongs is specified in that word’s lexical entry The immediate task facing us in this section is that of developing criteria for assigning words to classes

1. Lexical categories

The categories which we will look at in this section are four major lexical categories: nouns, verbs, adjectives and adverbs We will briefly look at prepositions although these do not have any characteristic grammatical endings

A familiar distinction is that between nouns (N) and verbs (V), and there are several ways in

which we can justify this for English For instance, nouns often refer to types of concrete

objects in the world (e.g cake, engine, moon, waiter, and, we might now suppose, plingle), while verbs typically refer to activities (collide, steal, applaud, snore) Furthermore, verbs and

nouns exhibit a different range of forms: most nouns have a special form for the plural

(engine ~ engines), while verbs have a larger number of forms, as shown by the following

sentences

(1.) a Harriet applauds my suggestion

b Tom and Harriet applaud my suggestion

c The children are applauding the clown

d The children applauded the clown

Third, nouns and verbs combine with other words to form phrases in distinct ways For

example, a noun will often be found preceded by a definite (the) or indefinite article (a/an)

(the moon, an engine) A verb cannot be preceded by these articles (*the applauds, *an applauded) If we form a phrase consisting of an article and a noun, this can often follow a

verb to form a larger phrase (steal an engine, applaud the waiter) - we say that an engine and

the waiter function as complements of the verbs steal and applaud in these constructions

However, verbs themselves cannot generally function as complements of other verbs (*steal

applauding, *applaud collides) Alternatively, an article-noun sequence may combine with a

verb to form a whole sentence (the waiter snores) - here the waiter functions as the subject of

the sentence (see section 17 for further discussion of subject and complement) Again, verbs

themselves cannot generally fulfill this role (*stole collides, *collides snores) Generally, we

say that subjects and complements are arguments of verbs and a typical simple sentence,

such as that in (2) consists of a verb (stole) and its arguments (the waiter, a cake):

15

(2) The waiter stole a cake

A third major word class recognized in traditional grammar is adjectives (A) These typically refer to properties which people or things possess and they are used to modify nouns, e.g

happy man, noisy engine Although they share with articles the property of appearing in front

of a noun, if an article and an adjective both combine with a noun, they do so in a fixed order

(a happy man, *happy a man, the noisy engine, *noisy the engine) We can also ascribe a

property by putting the adjective after a form of the verb be to form a sentence (the man is

happy, the engine was noisy) Like nouns and verbs, many adjectives have special forms

indicating the extent to which a property is true of some-thing, the comparative form,

happier, ‘happy to a greater degree than’, and the superlative form, happiest, ‘happy to the

greatest degree’ A fourth class of word is adverbs (ADV) While an adjective modifies a

noun, an adverb typically modifies a verb, adjective or another adverb, indicating how, when

or why something happened or the degree to which a property characterizes an individual or event Examples illustrating these three uses appear in (3) - the modifying adverbs are in italics and the modified item is in bold:

(3) a The waiter carelessly dropped the plate

b The engine is really noisy

c The audience applauded the clown very enthusiastically

(note that in 3c, the adverb enthusiastically, itself modified by very, modifies the verb

applauded)

Adverbs can readily be formed from a majority of adjectives by the addition of -ly: happily,

slowly, reluctantly, etc However, adverbs which do not fit this characterization are far from

uncommon: very, well, yesterday Another important word class is illustrated in (4):

(4) a Harriet was sitting under a tree

b They're due to arrive before noon

c That is the end o/the news

d There was a debate about economic policy

The italicized words in (4) precede nouns (or phrases centered around nouns, such as a tree or

economic policy) They typically serve to relate objects, people or events in space or time (under I be fore), though often the relationship is more abstract as in (4c, d) Words of this

type are called prepositions (P), and they do not have the capacity to appear in a range of

different forms (*unders *abouted, *ofest, *beforely)

Up to now, we have distinguished five word classes or lexical categories In doing this, we

have appealed to three types of criteria for establishing a category: semantic (relying on meaning), morphological (relying on word forms) and syntactic (taking account of behavior in phrases) Taken together, these criteria identify our separate classes quite well However, it is important to be clear that there are plenty of cases where one or other type of criterion fails to

work For instance, some nouns refer to abstract ideas rather than concrete objects (justice,

idea, quantity); worse still, there are nouns such as game and speech which refer to types of

activities, the semantic criterion we introduced for recognizing verbs For some nouns the pluralization criterion does not work in a straightforward fashion, either because their plural

forms are irregular (men, women, children) or because they lack a plural form entirely

(*furnitures, *sakes) Likewise, there are verbs which refer to states rather than, activities (fear, border (on)), and other difficulties with applying these criteria too rigidly will become