Making Words in English PDF
Making Words in English PDF
Making Words in English PDF
Making Words
in English
Copying prohibited
Art. No 31008
eISBN 91-44-02481-9
© Magnus Ljung and Studentlitteratur 2003
Printed in Sweden
Studentlitteratur, Lund
Web-address: www.studentlitteratur.se
Printing/year 1 2 3 4 5 6 7 8 9 10 2007 06 05 04 03
Manuella korrigeringar
storlek på noter, not i kap 1 bort.
sidhänvisningar manuellt gjort.
Contents
Preface 7
1 Words 9
The linguistic sign 9
The structure of complex words 16
Defining ‘word’ 19
Reference, denotation, sense 25
The origins of English words 26
Exercises 29
4 Prefixation 61
Characteristics of prefixes 61
Common English prefixes 67
Exercises 77
5 Suffixation 79
General characteristics of suffixes 79
Noun-forming suffixes 85
Adjective-forming suffixes 96
Verb-forming suffixes 103
6 Conversion 109
General characteristics of conversion 109
Productive English conversion patterns 111
Partial conversion 117
Exercises 119
7 Compounding 121
General characteristics of compounding 121
Noun compounds: general characteristics 129
Adjective compounds 134
Compound verbs 138
Compound adverbs 139
Exercises 140
Notes 173
Chapter 1 173
Chapter 2 175
Chapter 3 177
Chapter 4 178
Chapter 5 179
Chapter 6 181
Chapter 7 182
Chapter 8 183
Chapter 9 184
Preface
1 Words
Meaning: ‘anger’
natural connection
Form: growl
There are also non-motivated or arbitrary signs. These are the signs
we find in more complicated information systems like human lan-
guages, typically—but not exclusively—represented by the words in
those languages. Such linguistic signs are not universally understood
and the meaning of each sign has to be learnt separately, because
there is no natural relation between form and meaning.2
There is for example no particular reason why the letter combina-
tions cow and dog and the corresponding sound sequences [kau]
and [dɒg] are tied to the meanings ‘a fully bred female animal of a
domesticated breed of ox, used a a source of milk or beef’ and ‘a
domesticated carnivorous mammal that typically has a long snout,
an acute sense of smell … and a barking … voice’ as one dictionary
puts it rather than the other way round.3
Unlike the dog’s growling, the words cow and dog mean nothing
to those who don’t already know the code. Everybody understands
the meaning of a growl, but in order to understand the meanings of
cow and dog, you have to know English. Non-motivated signs like
cow etc. must accordingly be described in the following way:
Meaning: ‘cow’
no natural connection
Form: [kau]/c+o+w/
of the stem, i.e. the part of a complex word to which the affix is
added.
They know, for instance that words ending in e.g. -ous or -able
like dangerous or readable are in all likelihood adjectives with certain
characteristic meanings. They also know that words ending in e.g.
-er like runner and dish-washer tend to be nouns describing a person
or machine carrying out a certain activity, and that words that are
verbs and begin with un- or re- like untie and recount denote, respec-
tively, the reversal and the repetition of a process or an action. They
have similar detailed knowledge of most other affixes.
LEVEL 1 unreadable
the next lower level.8 The analysis goes on until the entire word has
been divided into morphemes, in this case un-, read and -able.
The two parts isolated on each level are called the immediate con-
stituents of the unit on the level immediately above. Thus the
immediate constituents of the word unreadable, as analysed above,
are the prefix un- and the adjective readable, and the immediate
constituents of readable are read and -able. But why is such an anal-
ysis of the word considered to be ‘the right one’? Why not analyse
unreadable as consisting of unread and able, and then analyse unread
into un- and read as in the figure below?
unreadable
unread- able
un- read
The answer has to do with what we know about the meanings and
combining habits of the bits isolated by the analysis. We know that
there is a prefix un- combining with verbs, found in for instance
unwind, unzip, untie. We are also aware that there is another un- pre-
fix combining with adjectives, in for instance unkind, untrue,
unwise. The two un- prefixes have different meanings: the one
found with verbs means ‘reverse the activity described by the verb’,
while the prefix combining with adjectives simply has negative
meaning: it means ‘not’.
It follows that to unwind something is to reverse the result of a
previous winding process, to untie a knot is to reverse the previous
process of tying that resulted in a knot. This is not a meaning that
makes sense together with the verb read: to unread a book would
have to mean something like ‘to make a book one has previously
read unread’. Meanings like that are impossible in the world as we
know it (but cf. the comments on e.g. unmurder on p. 68) and for
that reason we must reject the analysis in the second figure.
In the analysis in the first figure, on the other hand, we postulate
an adjective readable to which we add negative un-, an operation
resulting in the word unreadable meaning ‘not readable’, a meaning
that makes perfect sense.
Defining ‘word’
So far I have been using the term ‘word’ as if it were a well-defined
entity with a single meaning. But the word word is a slippery cus-
tomer and without wishing to complicate matters more than neces-
sary, I must now point to some of the different meanings of the
term. My discussion will focus on the written word and will avoid
as far as possible the problems raised by words in the stream of
speech.
how many different word-forms there are in the text above. Count-
ing like that, we find that four of the word-forms are used twice, i.e.
cook, cooks, go, and as. Since we are now interested in different word-
forms, we count each of these four word-forms only once. That
gives us a total of 11 words (= different word-forms) in (1).
The words found in the first count—the one in which all repeti-
tions were counted—are called tokens, while the ones found in the
second count—the different words—are known as types. The dis-
tinction between tokens and types is made use of in frequency
word-lists from different texts, like the list below. Such lists are
made up of two columns, one containing the word types, the other
indicating the number of tokens each type has in the text. The
number of tokens for each type is known as the frequency (of occur-
rence) of that type.
In the list above, the types as, cook, cooks, go have two tokens each,
while the types a, and, good, she, the, was, went, only have one token
each. The total number of tokens (15) is higher than the number of
types (11). This is the situation in all normal texts, where the
number of tokens is usually much higher than the number of
types.11 There is a simple reason for that, of course, i.e. the fact that
some words—in particular ‘grammatical’ words like the, a, an and
Clichés: call it a day, call a spade a spade, the devil take the hindmost,
famous last words, go through the roof, go missing, the show ain’t over
till the fat lady sings, beggars can’t be choosers, you are the cream in my
coffee!
Complex prepositions: because of, except for, owing to, in front of, in
back of, in place of, in spite of, as a result of
Idioms: in the doghouse ‘in disgrace’, up shit creek without a paddle ‘in
serious trouble’ (US slang), come a cropper ‘fail’, kick the bucket ‘die’,
as sober as a judge, as pissed as a newt (British slang for ‘dead drunk’).
Phrasal verbs: touch down, fall out (‘quarrel’), give in, come in, go on,
set up, bring up, call off, take in (‘deceive’), turn off, switch on
Prepositional verbs: look at, care for, think of, call on, succumb to, die
of, suffer from, talk about, shudder at, ask for, believe in
Exercises
1 Explain the difference between motivated and non-motivated
(arbitrary) signs. In which category would you put e.g. flag sig-
nals, human sign language and and the way(s) many animals
mark their territory?
(A)
When Chomsky’s ideas spread across into the field of psychol-
ogy in the early 1960s, they made an immediate impact. Psy-
chologists at once started to test the relevance of a transforma-
tional grammar to the way we process sentences. Predictably,
their first instinct was to test whether there was a direct rela-
tionship between the two.
Jean Aitchison The Articulate Mammal p. 183.
(B)
Advice for delegates to next week’s United Nations World Sum-
mit on Sustainable Development in Johannisburg: if someone
approaches your car, pushes a gun in your face, and shouts
“Hijack!” don’t reply “My name’s not Jack.” The good news is
there will be 8000 extra police for protection at the biggest con-
ference ever held in South Africa’s biggest and baddest city.
TIME Magazine, Aug. 26, 2002, p. 29.
9 What are the lexemes in texts (A) and (B) above? Are there cases
in which several answers are possible?
If a word satisfies all the criteria for a certain word class, it is by def-
inition a member of that class. However, even a superficial study of
the English word classes reveals that words differ in this respect. For
every word class that we consider, there are words that meet all the
criteria demanded for inclusion in that class, there are others that
meet some but not all of them, and there are words that meet just
one of these criteria.
Word class membership is thus not an all-or-nothing affair but a
matter of degree: some words are much more typical representa-
tives of a certain word class than others. The words that meet all the
criteria for inclusion in a given word class are central members of
the class, while those meeting only one or a few criteria are marginal
members of that class.5
As the table shows, all nine words meet the subject function crite-
rion: they can all replace X in constructions like ‘X is interesting’,
‘X impressed us’, etc. However, the first five words have to be
accompanied by determiners like the, their: we cannot say e.g.
*Teacher (city) impressed us but have to say The teacher (city etc) or
Their teacher (city etc.) (Here and in the rest of the book the asterisk
sign * will be used to identify constructions that are impossible).
Anger and salt may occur both with and without determiners, but
cannot express number. Only Susan and Sydney refuse determiners
altogether and also lack the ability to express number.
The distribution of the different nouns in the table above shows
that the class of nouns is made up of several sub-categories. The first
column is not distinctive in this respect, since all nouns may have
subject function. However, if we move to the second column we
find a clear difference between teacher, city, book, idea, departure,
anger on the one hand, and Susan, Sydney on the other. All the
members of the first group take determiners but neither of the
members of the second group do.
That distinction is the most fundamental one for the noun class.
It divides it into common nouns—which all take determiners—and
proper nouns which do not. Proper nouns, whose special status is
marked in writing by the initial capital, are used as names and nor-
mally take neither articles nor the plural. On the other hand, proper
nouns are certainly used to identify people and places, respectively,
so they would seem to satisfy the semantic criterion in the same
way as the common nouns.
However, on closer inspection, the way proper nouns like Susan
and Sydney identify people and places turns out to be different from
the way common nouns like teacher, city, etc. identify the people,
places and things they are used to refer to. Proper nouns like Susan
or Sydney are used as names: they have no sense or denotation as
those notions were explained in Chapter 1, but simply refer directly
to a specific individual person, place or thing. A personal name like
Susan, for instance, identifies a single individual known to both
speaker and hearer. Accordingly, there is no need to further define
the referent of Susan by means of the definite article the, nor is it
possible to use the indefinite article or the plural.
Within the category of common nouns, a further distinction
needs to be drawn between on the one hand teacher, city, book, idea,
departure, which may occur in both the singular and the plural, and
on the other salt and anger, to which the number category doesn’t
apply. The members of the first group are called count(able) nouns or
countables for short. As the name indicates, these nouns denote
phenomena that can be counted—they denote entities. But salt and
anger are uncountable: salt denotes a type of substance or concrete
mass—nouns like this are often called mass nouns—while anger
denotes a state of mind.
We can sum up our discussion above of the different kinds of
noun as in the following figure:
NOUNS
Recategorisation
The classification above needs qualifying in two respects. To begin
with, it is possible to find constructions in which anger and salt
have plurals or are preceded by the indefinite article, for instance:
On those occasions I felt quite different angers.
The mine produces two different salts.
I experienced an anger I had never felt before.
Such a salt has the following properties.
The first and third lines above contain the word duck referring to an
animal; in the second and fourth lines the interpretation is ‘food’.
Most adjectives meet the first two criteria. Among the exceptions
we find words like e.g. former, naval, utmost, sheer, which are used
only attributively, and words with initial a- like afraid, aghast,
aware, which are used only predicatively. The majority of the adjec-
tives also meet the third criterion, i.e. they have gradable meaning
and can accordingly be modified by intensifiers (cf. next section) like
e.g. very, incredibly, extremely, and be compared by means of the suf-
fixes -er and -est, or the adverbs more and most. Certain other adjec-
tives are non-gradable and normally take neither intensifiers nor
comparison, for example the colour adjectives and words like
atomic, linear, lunar, naval, nuclear.
The bulk of the adjectives may also be turned into adverbs in -ly,
for instance quick-quickly, incredible-incredibly, surprising-surprisingly.
In addition, quite a number of adjectives are morphologically
marked: it can be fairly safely assumed that words ending in e.g.
-able, -ed, -ful, -ic, -ish, -ive, -less, -y will also meet the other criteria
for adjectivehood. (However, note the comments on suffixation
and conversion in Chapter 6).
ENGLISH ADVERBS
INTENSIFIERS NON-INTENSIFIERS
(Modifying function) (No modifying function)
ity with certain aspects of the closed classes. The following brief
comments will hopefully facilitate the reader’s understanding of
those passages.
Auxiliaries
Formally, the auxiliaries as a group differ from lexical verbs in three
important respects:
(1) They don’t take the do-construction in negative or interrogative
sentences, and as a consequence they are moved to the begin-
ning of the sentence in direct questions
(2) They contract with a following not (as in shouldn’t, can’t)
(3) They cannot stand alone as predicate verbs in a sentence
(except in cases like e.g. [Can she come?]—Yes, she can/No she
can’t, where the lexical verb come has been left out).
Determiners10
As we saw in the previous discussion of the noun, one of the char-
acteristics of nouns in English is the fact that they may be preceded
by determiners. The determiners give information about definiteness
and indefiniteness, quantity and proportion. The basic determiners
are the definite and indefinite articles a/an and the, but certain
other words and constructions may also function as determiners,
chiefly possessive and demonstrative pronouns like my, your, this,
those, and quantifiers like some and any. Nouns with an apostrophe
genitive also function as determiners, as in Jessica’s coat, London’s
underground.
Prepositions
Prepositions are words that typically occur before a noun, like beside
in beside Henry, beside the building, beside the new building. Certain
prepositions are part of prepositional verbs, like e.g. on in decide on
something, while others are not, for example on in stay on the ground.
Prepositions are simple or complex. Common simple prepositions
are about, after, against, before, beside, between, by, in, near, on, to,
with. Certain simple prepositions have homonyms that are con-
junctions, for instance after and before which are prepositions in
after/before her departure but conjunctions in after/before she had left.
Many simple prepositions also have homonyms that are adverbs,
for example around and in: in We walked around the house, and They
dropped it in the water, around and in are prepositions, but in We
walked around, They just dropped in, they are adverbs.
Complex prepositions are multi-word constructions that have
fused to form units with prepositional function by means of a proc-
ess known as grammaticalization11, for instance according to, in addi-
tion to, in case of, out of, with regard to, in front of, (in) back of.
Occasionally new prepositions are also formed from single parti-
ciples, for example considering and given, both of which have prepo-
Conjunctions
Conjunctions are of two kinds: co-ordinating and subordinating. Co-
ordinators are used to link any two (occasionally more) units that
have the same syntactic status. There are three main co-ordinators:
and, or and but.
The subordinating conjunctions introduce subordinate clauses
with different syntactic functions. Many conjunctions consist of a
single word, but many are made up of several words, for instance as
if, so that, in order that. These complex conjunctions obviously have
their origin in the same grammaticalization process that was
responsible for the formation of complex prepositions.
Grammaticalization is like meaning extension (cf. p. 156) in
being a slow process that may take hundreds of years to be com-
pleted. As a result, speakers are usually unaware of the origin of com-
plex units and of the fact that on-going change is forever present in
the language of today. However, sometimes we can see the begin-
nings of the formation of a complex unit. A fairly recent change of
this type in English concerns the prepositional phrases on the basis
(of) and in terms of in sentences like He’s asked for special treatment on
the basis he’s been with the firm over twenty years and They’re a general
nuisance in terms of they harass people trying to enjoy the park12
Exercises
1 What is a word class? What kinds of criteria are used when
words are assigned to different word classes? Why is the seman-
tic—or ‘notional’- criterion a problem?
5 In the table on p. 36, the nouns anger and salt both have plusses
in the column ‘Combines with determiners’. However, there is
one determiner they cannot combine with—which one?
8 Given the criteria for adjectives given here, how do you suggest
that we handle items like upper class (upper-class), Sydney and
1930s in phrases like e.g. a very upper-class accent, a Sydney street,
a 1930s atmosphere. How many of the criteria for adjectives do
they meet?
3 Outline of English
Word-formation
Defining word-formation
As we noted in Chapter 1, many English words are simple, i.e. they
consist of a single base morpheme: dog, door, smile, mahogany.
Others contain two or more morphemes, for instance, uncertain,
rider, teapot. Of these three, the first two words are complex. Such
words are formed by the addition of an affix—a prefix or a suffix—
to a stem which is itself a word. In the first word, the prefix un- has
been added to the stem certain, and in the second the suffix -er has
been added to the stem ride. The third word—teapot—consists of
two words combined to form a third—a so-called compound.
The three words above have been formed in accordance with
present-day English word-formation rules, principles for the pro-
duction of new words. They represent three types of regular English
word-formation, i.e. prefixation, suffixation and compounding. There
are also a number of less regular types of English word-formation,
among them initialisms (FBI, asap), clippings (para, demo), blends
(Bollywood, edutainment), back-formations (to backpack, from back-
packer, to laze from lazy), rhyming slang and reduplicative forma-
tions (argy-bargy, higgedly-piggedly). These latter will be dealt with in
the final chapter of the book.
Although both prefixes and suffixes belong to the family of
affixes, they have different functions and produce different results.
Prefixes modify the meaning of a word from a certain word class, but
don’t normally change its word class membership: we may add un-
to the adjective certain to create the new word uncertain, but the
new word is still an adjective1. Suffixes, on the other hand, usually
change the word class of the word they are added to: ride is a verb,
but rider is a noun.
Suffixes and prefixes that can be used to form new words are
called derivational affixes, the words produced in this manner are
known as derived words, and the process of forming new words by
the addition of prefixes and suffixes is called derivation. Deriva-
tional suffixes should be distinguished from the inflectional suffixes
discussed in Chapter 1 (p. 22), for instance the -s and the -ed in She
lives here and She lived here two years ago and the -s in The dogs
barked. The inflectional suffixes don’t form new words, but merely
add grammatical categories like tense, number, person to existing
words2.
The inflectional suffixes also differ from the derivational ones in
being used with great regularity: there are few exceptions to the rule
that the past tense of verbs is formed by adding the suffix -ed, that
third person singular forms end in -s, and that plural nouns also
end in -s. The majority of the derivational suffixes do not show the
same regularity, although individual derivational suffixes may
show almost the same regularity as the inflectional ones, in partic-
ular the adverb-forming suffix -ly (cf. Chapter 7).
Compounding differs from affixation in involving the combina-
tion of (usually two) words to form a new word. However, com-
pounding has one thing in common with prefixation: the first ele-
ment is a modifier: it modifies the second element, also known as
the head. In a word like teapot, for example, tea modifies the head
pot, just as e.g. re- in rewrite modifies the meaning of write. In neither
case is there a change in the word-class of the head/stem: like pot,
teapot is still a noun, and like write, rewrite is still a verb.
Prefixation, suffixation and compounding can be characterised as
‘additive’ types of word-formation, since their output is the result of
the addition of something to an already existing word. English also
has another highly productive type of regular word-formation, one
that forms new words without formal change. This type of word-
formation is known as conversion (alternative names are functional
shift and zero derivation).
Conversion is very common in English, especially in the case of
the noun and verb categories; according to one source3, ‘there is no
English noun that can’t be verbed’, i.e. in principle all nouns can be
transformed into verbs. Thus we can carpet a room, doctor a drink and
cash a cheque. We can even steamroller others into accepting what
they don’t want, and we can audition for a role in a play.
ADDITIVE NON-ADDITIVE
PREFIX SUFFIX
fying the ‘doer’ of a verbal action or process. (The -er suffix may also
be used to form words denoting instruments like e.g. washer ‘wash-
ing machine’ and silencer).
On the other hand, the noun rider is obviously not new in the
sense ‘never produced before’. It has been in the vocabulary of Eng-
lish for hundreds of years. What the rule does is to provide a
description of the internal structure of derived words that may or
may not have been formed already. If the word resulting from the
application of the rule has been stored in the dictionary, then the
rule can be used to analyse that word. Thus word-formation has
both a productive and an analytic function.
The productive and analytic functions of word-formation are of
equal importance; it might even be argued that we have occasion to
analyse existing complex and compound words more often than we
have occasion to create new words. At the same time analytic
knowledge helps us both to create and to understand new forma-
tions created on a familiar pattern. Once we know that e.g. refusal,
bewilderment and readable are formed by the addition of the suffixes
-al, -ment, and -able to the verbs refuse, bewilder and read, we also
know how to analyse other words containing these suffixes.
Most agent nouns in -er (and other agentive suffixes) can be used
both in labelling and in syntactic repackaging, for instance driver,
which is found both in He works as a driver (labelling) and The driver
of the car was Harry Smith (syntactic repackaging). However, some of
the agentive (instrumental) nouns can only be used in syntactic
repackaging, presumably because there is no natural category that
they could be used to label.
It is, for example, perfectly possible to turn the verbs assert, nod
and pat into the agentive nouns asserter, nodder and patter, but they
will never appear in the dictionary for the simple reason that we
have no use for them as category lables. There are no categories of
people that need to be labelled by such nouns: that is why the fol-
lowing exchange sounds extremely strange:
A: What does your sister do?
B: She is an asserter/a nodder/a patter
But as the three examples below indicate, these nouns work per-
fectly well in syntactic repackaging:
Some people asserted that the problem could easily be solved. The asserters
were all males from the south.
When I asked my question, several people in the audience nodded. The most
enthusiastic nodder was Peter Wright.
There used to be people patting my back after my talks. But this time there
wasn’t a single back-patter.
prised us, are used instead of longer expressions like The fact that she
was stubborn surprised us or The degree to which she was stubborn sur-
prised us.
The syntactic repackaging function is also open to words formed
by means of prefixation, compounding and conversion. Thus
preschool in e.g. preschool teacher refers to a particular stage in the
education of children and is an instance of labelling. But in e.g. pre-
school activities, the likely interpretation is that this is a case of syn-
tactic repackaging meaning for instance ‘activities that take place
before school starts’.
As for compounds, the noun show-stopper is common as a label
and is defined in The New Oxford Dictionary of English as ‘a perform-
ance or item receiving prolonged applause’ (referred to below as
show-stopper (1)). However, if on a given occasion somebody actu-
ally stops an on-going show, the syntactic repackaging show-stopper
(show-stopper (2)) could perfectly well be used with reference to that
person (The show-stopper turned out to be a tall, side-whiskered Austral-
ian youth’).
It is less easy to find convincing examples of the labelling-syntac-
tic repackaging contrast with regard to conversion. However, some-
thing like syntactic repackaging is found with a verb like carpet
(‘provide with a carpet’), for example in an exchange like The room’s
got a new carpet.– Really, who carpeted it? This should be contrasted
with the verb carpet meaning ‘reprimand severely’ as in Sidney was
severely carpeted by his boss, obviously a kind of labelling.
In the rather long discussion above, I have suggested that the
question whether a complex, compound or converted word has to
be memorised and accordingly listed in the dictionary or not, is
mainly decided by the word’s predictability of meaning and its
function (labelling vs. syntactic repackaging). The argument is
summed up in the following figure:
COMPLEX/COMPOUND/CONVERTED
WORDS IN ENGLISH
Exercises
1 Which of the following words are complex and which are com-
pound?: reclassify, unbelievable, water crisis, Londoner, sun-tanned,
(the) Bushies, arms cache, Clintonian, Clintonomics, steamroller.
3 Mention is made here of words like love, hate, fear for which it is
difficult to indicate which word class is the original one and
which the converted. Can you think of other words for which
this is also true?
4 p. 51: The tree diagram indicates that English has two kinds of
affixation: prefixation and suffixation. Logically, there is a third
type, i.e. infixation, the insertion of a morpheme inside a word.
Does English have any infixed words?
4 Prefixation
Characteristics of prefixes
As their name indicates, prefixes are bound affix morphemes occur-
ring word-initially, i.e. at the beginning of words. In addition to
their initial position, prefixes have certain additional defining char-
acteristics. One of these is that they are semantically special,
expressing a fairly narrow range of clear meanings of which the
main ones are: NEGATION, REVERSAL, REMOVAL, MANNER, DEGREE/SIZE,
ATTITUDE, LOCATION/DIRECTION, and TIME/SEQUENCE.
Prefix meanings
With the noticeable exception of DEGREE/SIZE, each of the meaning
categories above is represented by a fairly limited number of pre-
fixes. In addition it is often the case that one or two of the prefixes
linked to each meaning are dominant and occur more or less freely
with stems with certain characteristics, while the remaining ones
are infrequent and of unpredictable occurrence. The following list
shows the distribution of prefixes across the prefix meanings listed
above.
Meanings Prefixes
NEGATION 5
REVERSAL 3
REMOVAL 3
MANNER 2
DEGREE/SIZE 14
ATTITUDE 2
LOCATION/DIRECTION 4
TIME/SEQUENCE 5
The table above does not claim to present the total number of pre-
fixes or prefix meanings in English. What it does contain, arguably,
are those prefixes and prefix meanings that are often found in texts
of a general kind. Texts of a more specialised nature like e.g. scien-
tific/ technical texts use many prefixes not accounted for here.
There are three important cases of homonymy among the Eng-
lish prefixes discussed here, i.e. cases where the same spelling and
pronunciation represents separate prefixes with clearly different
meanings, namely un- expressing the meanings negation, reversal
(of the action of the following verb), and removal, dis- with the
same three meanings, and de- which may express both reversal and
removal. The table below summarises this information and pro-
vides examples of how the prefixes are used.
Prefix function
Prefixes typically have a modifying function: they serve to modify
the following stem rather than change it radically. Unlike the suf-
fixes (cf. Chapter 5), prefixes as a rule don’t change the word class of
the stem they are attached to, nor the general semantic category to
which it belongs. Thus if we add the prefix re- to the verb stem fill,
the resulting combination refill is still a verb, and still represents the
semantic category of filling.
There are certain exceptions to this rule, however, cases where
prefixes do not have purely modifying function, but actually
change the word class of their stems. Among these we find the
removal prefixes de- and un- discussed in the previous section,
which are added to nouns and turn them into verbs meaning
‘remove whatever the noun stem refers to’. Accordingly, defrost and
delouse mean ‘remove frost/lice from’ and unroof and unsaddle
mean, respectively, ‘remove the roof (from a house)’ and ‘remove
the saddle from’(usually a horse). Unsaddle may also mean ‘remove
(the rider) from the saddle’ and is usually used about horses that
throw their rider. Another exception is the prefix out- in combina-
tions like outclass, outdistance, outwit. Here the addition of out- to
the nouns class, distance and wit results in new words that are verbs.
The modifying function of most prefixes makes them similar to
the first element in the so-called neo-classical compounds like e.g.
biography, economy and technology (see Chapter 8 for a presentation
of these). Bio-, eco- and techno- are so-called initial combining forms,
i.e. bound Latin and Greek base morphemes that are unlike prefixes
in that they usually only combine with other bound Latin/Greek
morphemes, i.e. the so-called word-final combining forms like
-graph(y), -nomy and -logy. But in certain cases bio-, eco- and techno-
combine with ordinary English words, as in e.g. bio-terrorist, eco-
crime and technopeasant.
On account of this similarity in both function and combining
habits it is sometimes difficult to distinguish between initial com-
bining forms and prefixes: certain of the forms listed as prefixes
below may arguably also be regarded as initial combining forms.
However, it is normally possible to distinguish between prefixes
and initial combining forms in terms of meaning. As has already
been pointed out, the meanings of prefixes belong to a restricted set
[pr peə], and regret [r ret]. In the first three examples—which con-
tain the English prefixes de-, pre- and re- —the letter e is pronounced
[i] and has secondary stress, but in the last three examples—which
do not contain English prefixes—e is unstressed and has the pro-
nunciation [] or even [ə] (as when regret is pronounced [rə ret]).
The claim made for the highly productive prefixes is that they
may be attached to all or at least the majority of stems with certain
characteristics. However, for the majority of prefixes it is indeed
possible to characterise the stems they prefer, but it is impossible to
claim that they combine with all or even the better part of such
stems. Reversative de- is a good example of this. It combines fairly
freely with verbs from Latin, especially those ending in -ate, -ify, -ise
as for example in decentralise, decontaminate. But it certainly doesn’t
combine with all such verbs; thus we find neither *detranslate,
*deglorify or *desymbolise. Prefixes like reversative de- may be called
‘fairly productive’.
Although it is a distinction difficult if not impossible to maintain,
we may want to establish a further category of weakly productive pre-
fixes that combine on an occasional basis with a certain type of
stem. A likely candidate for this group would be negative in-, which
sometimes combines with ultimately Latin stems as in inaudible,
incomplete (and with variation impossible, irregular but are impossi-
ble in others. Thus only e.g. unable, will do, never *inable, despite
the fact that the corresponding noun is inability.
semi-: ‘partly’
combines fairly freely with adjectives with gradable meaning, for
instance in semi-conscious, semi-naked, semi-open, semi-permanent,
semi-predictable, semi-public, semi-rotted, semi-secure, semi-skilled,
semi-tough. The prefix is usually hyphenated. In scientific language
there is also non-hyphenated semi meaning ‘half’ as in e.g. semi-
vowel, semicircle.
super- ‘more than the norm’; ‘highly, extremely’; ‘very large, prom-
inent’
Super- is a very productive suffix that combines with adjectives. In
words like e.g. super-confident, super-elevated, super-natural, supersen-
sitive it indicates that the quality denoted by the adjectives is
present to a greater extent than is normally the case. In most cases,
however, this prefix simply means ‘highly, extremely’, for instance
in super-bland, super-civilised, super-clean, super-cool, super-fit, super-
strong, super-smart, super-talented.
In combination with nouns, super- has meanings like ‘very large,
unusually prominent, powerful, etc’ as in e.g. super-G, super-clerk,
super-nerd, super-tanker, super-state.(Super may also be used as a word
of its own meaning ‘excellent’, ‘very good’ as in It was a super con-
cert, The concert was super.)
ultra- ‘extreme(ly)’
combines fairly productively with adjectives (and certain nouns)
meaning ‘extreme(ly)’. Examples are ultra-conservative. ultra-fair,
ultra-marginal, ultra-modern, ultra-nationalist, ultra-right-winger, ultra-
slim, ultra-Tory. The use of hyphenation varies.
operate, co-exist, cohabit. The use of hyphens with co- varies. It tends
to take primary stress before nouns.
extra- ‘outside’
The prefix extra- combines freely with Latin-based adjectives, partic-
ularly those ending in -al, -ial, -ic, -ar, like extracontractual, extracur-
ricular, extragalactic, extragovernmental, extralinguistic, extramarital,
extraterrestrial, extraterritorial.
trans- ‘across’
The prefix trans- combines with certain adjectives of Latin origin
derived from geographical nouns: transatlantic, trans-Siberian, trans-
pacific, trans-pennine. The use of hyphens varies.
Exercises
1 In what way are the ‘removal’ prefixes de- and un- different from
most other prefixes?
5 What other adjectives besides big, deep, good, strong have oppo-
sites not formed by the addition of a prefix?
8 According to the rule given here, the prefix co- combines fairly
freely with agent nouns. However, there is obvious a difference
between the common co-pilot, co-driver, co-producer, the less
common but possible co-runner, co-singer, co-believer and the
highly unlikely *co-liker, co-sitter, co-snore. Try to account for
these differences.
5 Suffixation
as stems. Other suffixes are less particular: -able, -er and -ness are
found with both kinds of stems.
Suffix meanings
Derivational suffixes usually have two kinds of meaning. To begin
with, they all signal that the word they are attached to belongs to a
certain word class: words ending in e.g. -hood, -ion or -ness are
bound to be nouns, words ending in -ize and -ify are verbs etc.
(However, as described in Chapter 6, the process of conversion may
sometimes override such word-class assignments, cf. verbs like to
commission, to position, and to station).
In addition to signalling word-class membership, derivational
suffixes have other, more or less clear meanings: the -er used to form
nouns from verbs as in e.g. runner, reader etc. has the meaning
‘agent’ or ‘instrument’, the -y added to nouns to form adjectives as
in e.g. dirty, sandy, snowy carries the meaning ‘full of/covered with’.
There is also a fair amount of homonymy among derivational suf-
fixes i.e. suffixes may be formally identical but have different mean-
ings (and different combining habits). Thus in addition to the
familiar agent/instrument suffix -er mentioned above that can only
be attached to verbs, there is another suffix -er which is added to
nouns and has the meaning ‘person associated with what the stem
denotes’ as for instance in New Yorker, jet-setter, villager, brat packer.
We also have to recognise several different suffixes of the form -y.
The y-suffix meaning ‘full of/covered with’ has already been men-
tioned. Another adjective-forming y-suffix attached to nouns has
the meaning ‘resembling’, ‘having the characteristics of’, for exam-
ple in bossy, catty, powdery. A third carries the meaning ‘character-
ised by’, as in e.g. ballsy, gutsy ‘characterised by courage’, trafficky
‘characterised by traffic’. (In addition there are -y/-ie suffixes used in
nouns like aunty/auntie, lefty/leftie; cf. p. 91 below).
of primary stress in such words, and there have also been many
attempts to establish such rules. This has turned out to be a very
challenging task, however, and none of the attempts so far have
been entirely successful.
Nevertheless, if we restrict our attention to derivational suffixa-
tion, certain general principles for stress position may be found.
These principles operate in terms of the type of suffix added. Basi-
cally there are three types of suffixes in this respect: those that don’t
change the stress pattern at all, those that take primary stress them-
selves, and those that require primary stress to fall on the syllable
immediately preceding the suffix. Examples of these types and how
they are used are given below.2
(1) Suffixes that don’t affect the stress pattern (stress-neutral suf-
fixes)
The most important members of this group are -able, -dom, -er, -ess,
-ie, -ing, -ish, -ism, -ist, -ise/-ize, -like, -ly, -ment, -ness and adjective-
forming -y. Examples: understandable, officialdom, researcher, steward-
ess, auntie, panelling, monkeyish, Americanism, Africanist, computerise,
statesmanlike, evidently, arrangement, unfriendliness, spidery. In certain
words ending in -ess primary stress goes on the suffix. The word
stewardess, for instance, may be pronounced both [ stjυədəs] and
[stjυə des].
(2) Suffixes that are themselves stressed
Representatives of this type include -ation, -ition, -ution, -ee, -eer,
-esque, -ette, -ese, (note that -ation, -ition, -ution and -eer count as sin-
gle suffixes here), -itis. Examples of words with these suffixes are
confirmation, definition, resolution, referee, auctioneer, kitchenette,
Kiplingesque, Japanese, appendicitis (cf. also nonce formations like
creditcarditis). Words ending in -ee like e.g. divorcee sometimes take
primary stress on the syllable immediately before the suffix.
(3) Suffixes requiring primary stress on the preceding syllable
A rather small number of suffixes require that the primary stress go
on the syllable immediately preceding the suffix. In stems that
already have primary stress on their final syllable, the stress remains
where it is (as in e.g. intense: intensity); in other stems it is moved to
the syllable preceding the suffix as in rapid:rapidity.
The most common of the members of the third group are -ial,
-ian, -ic, -ical, -ify, -ity. Examples: atomic, sequential, Clintonian, per-
sonify and readability. In many—but not all—of its occurrences, the
adjectival suffix -al belongs here too, for instance in fundamental
and adjectival. Note that moving the primary stress in words like
these often causes radical pronunciation changes: compare Clinton
[ klntən] and Clintonian [kln təυnən], and note how -able [əbəl]
changes to -abil- [ə bl] before -ity, as in e.g. countability, dependabil-
ity, readability.
Although it is usually attached to bound rather than to free
stems, mention should also be made here of the suffix -ion. English
has many words ending in -ion, all of which place primary stress on
the syllable immediately preceding the suffix, for instance prohibi-
tion, invasion, permission, depletion, derision (For the spelling of the
words in -ion, cf. p. 84).
It should be emphasized again that the above account of the
interplay between suffixes and stress cannot claim to be compre-
hensive and is based on general patterns rather than absolute rules.
By and large, these principles provide valuable and reliable guide-
lines for the stress patterns of suffixed words. However, certain com-
plications have been left unaccounted for. In addition, there are
sometimes individual exceptions to the principles.
(1) Verbs containing -mit change final t to ss before -ible, -ion, -ive,
Examples: admit, admissible, admission, admissive; permit, permissi-
ble, permission, permissive; omit, omissible, omission.
(2) Verbs containing -cede, -ceed, -fend, -hend, -pand, -pend. -spond,
-tend change final -d(e) to -(s)s before -ible, -ion, -ive and -or.
Examples: concede, concessive, concession; succeed, successive, succes-
sion, successor; defend, defensive, defensible; comprehend,
comprehensive, comprehension, expand, expansive, expan-
sion; expend, expensive; respond, responsible, responsive;
extend, extensive, extensible, extension.
(3) Verbs containing -cide, -clude, -lide, -lude, -plode, -ride, -suade,
-vade, -vide change the final -de in the spelling to s before -ion
and -ive.
Examples: decide, decisive, decision; include, inclusive, inclusion; col-
lide, collision; allude, allusive, allusion: explode, explosive,
explosion; deride, derisive, derision; persuade, persuasive,
persuasion; invade, invasion; divide, divisive, division.
Note that the letter sequence -sive in the words in (3) above is pro-
nounced [sv], but that -sion is pronounced [ən]. Decisive, inclusive
are accordingly pronounced [d sasv] and [n klusv], but decision,
inclusion are pronounced [d sən] and [n klυən].
(4) Verbs containing the stems -solve and -volve change these to
solut- and -volut before the suffix -ion. Final -t followed by -ion is
pronounced [ʃ]
Examples: dissolve, dissolution; resolve, resolution; revolve, revolution
(5) Verbs in -ceive change final -ve to -pt before -ible, -ion, -ive, -or.
Final -t followed by -ion is pronounced [ʃ]
Examples: deceive, deceptible, deceptive, deception, ; receive, receptive,
reception; conceive, conception.
(6) Verbs in -duce change -duce to -duct before -ible, -ion, -ive, -or. The
combination -duction is pronounced [ dkʃən]
Examples: reduce, reduction, reductive; produce, production, produc-
tive; conduct, conductor
(7) Verbs in -pel change -pel to -puls before -ive, -ion. The combina-
tion -pulsion is pronounced [ plʃən]
Examples: compel, compulsive, compulsion; repel, repulsive, repulsion
expel, expulsion
Changes also occur before the suffixes -ive and -ion in the words
repeat, compete, expose, inquire, oppose, recognize as in repetition, repet-
itive, competition, competitive, exposition, expositive; opposition; inquis-
itive, inquisition; recognition.
In addition, stems ending in the common verb suffix -ify changes
-ify to -ific- before the suffix -ation, as in classify:classification, iden-
tify:identification, etc.
Noun-forming suffixes
Nouns from nouns
A number of noun suffixes combine with noun stems, thus violat-
ing the principle that suffixation should always be word-class-
changing. At least seven types of meaning may be distinguished:
AMOUNT, COLLECTIVES, ACTIVITY CONNECTED WITH STEM, STATE OF BEING
WHAT THE STEM DENOTES, PERSON CONNECTED WITH STEM, SIZE/SEX/
STATUS, FAMILIARITY.
AMOUNT: -ful
The suffix -ful is added fully productively to nouns to express the
meaning ‘the amount contained in what the stem denotes’. It is
attached to nouns denoting containers as in They poured bucketfuls /
spoonfuls/glassfuls of water onto the table. While the nouns in these
words are all prototypical containers, the notion of what is a con-
tainer may be extended considerably as the following examples
show: fistful, handful, mouthful, plateful, sinkful, shovelful.
THE STATE OF BEING WHAT THE STEM DENOTES: -dom, -hood, -ship
The suffixes -dom, -hood and -ship are all added to nouns to form
new nouns denoting the state of being what the stem noun
denotes. Often the very same nouns may also be used to express
The suffix -er may be added to names of cities and other geographi-
cal names to refer to people who live in them. The stems may con-
sist of more than one word. Examples of such combinations are e.g.
Londoner, Dubliner, New Englander, New Yorker (but only e.g. Bosto-
nian, San Fransiscan, Roman). The same kind of suffixation may be
used with nouns denoting groups of people with certain aims and
beliefs, as for example in bratpacker, jetsetter, free marketer, America-
firster.
The -an/ian suffix is found in combination with many types of
stems forming both nouns and adjectives. In the past, it was fre-
quently used to name persons and things from a certain country,
but this type of formation can hardly be regarded as productive in
today’s English. In many cases, the suffix was not attached directly
to the stem, but alternates with final -a and -ia as in America:Ameri-
can, Algeria:Algerian (but note exceptions like Canada:Canadian).
Compare also the use of the suffix with names of cities as in San
Fransisco:San Fransiscan, Paris:Parisian, Boston:Bostonian. Note that
here and elsewhere, primary stress in -ian words falls on the syllable
immediately preceding the suffix, while in -an words it stays in its
original place.
In today’s English the productive use of -an/ian is largely
restricted to combinations with personal surnames and results in
words that may be both adjectives and nouns. While -an is found in
certain cases like Elisabethan and Lutheran ‘a person who lived at the
same time as Queen Elisabeth I’ and ‘person who believes in the
ideas of Luther’, the -ian suffix is much more common than -an in
combinations with names7.
Typical examples of nouns in -ian are Chomskyan and Darwinian
‘person who believes in the ideas of Chomsky or Darwin’, Faustian
‘person who is like Faust’, Machiavellian ‘a person who is like Mach-
iavelli in cunning and ruthlessness’. The -ian suffix seems to be
associated with a certain pompousness, which probably accounts
for its use for comic effect in e.g. the recent—obviously rule-break-
ing-formation Star Warians ‘people who like the Star Wars films by
G. Lucas’.8
The suffix -ese is found in many adjectives denoting nationality
or other origin such as Japanese, Milanese, etc. It is also a noun suffix
attached to names of countries, regions, and the like, to create other
nouns denoting the language spoken in that country or region: Do
you speak Japanese?, My Vietnamese is a bit rusty. This use of the suffix
-ese is only slightly productive. However, the suffix -ese is fairly
freely used as a noun suffix to denote the jargon of certain writers,
groups or types of text, as in e.g. Johnsonese ‘language typical of Dr
Johnson’, bureaucratese ‘the jargon of bureaucrats’, officialese ‘the
jargon of officials’, teacherese ‘the jargon of teachers’, headlinese,
newspaperese. (Cf. also journalese ‘the language of journalists’,
legalese ‘the language found in legal documents’, and translationese
‘the language found in translations’.)
Nationality is sometimes also expressed by means of the suffix -i,
as in e.g. Israeli, Pakistani, Bengali, Azerbajani, Iraqi, Kashmiri mean-
ing both ‘from Israel, Pakistan’, etc and ‘inhabitant of Israel, Paki-
stan’ etc. Some of these—like e.g. Bengali, Kashmiri—may also be
used with reference to the language spoken in the country. It has
been observed that this type of formation seems to be restricted to
countries in the East or Near-East. Since the emergence of new
nations is a fairly unusual event, it is difficult to have an opinion
about its productivity9.
The suffix -ist is very productive. It is used to denote people who
act in accordance with certain doctrines (Buddhist, socialist, flat-
Earthist) and people who discriminate against others (ableist, racist).
In cases like these, there is commonly a corresponding form in -ism:
a Buddhist adheres to Buddhism, an ableist is guilty of ableism, etc.
However, there are also many -ist nouns with no corresponding
nouns in -ism, for instance those denoting people playing a certain
musical instrument (pianist, violinist) or specialising in certain areas
of study (physicist, psychologist).
As for -ite, it may be found attached to the names of politicians to
denote their followers, as for instance in Blairite, Bushite, Clintonite,
Thatcherite, Reaganite and also to the names of people who are the
originators of a school of thought as in e.g. Chomskyite.
The suffix -y (-ie) is used in informal style and may be added to a
noun stem to denote somebody who is interested in and/or good at
what the noun stem denotes. Thus a foodie is interested in good
food and possibly a good cook, a junkie is a drug addict (junk ‘her-
oin’), a techie/techy is somebody who is an expert on or enthusiastic
about technology’, and a winie a person interested in/knowledgea-
ble about wine(s). The suffix can apparently also be used to refer to
FAMILIARITY: -y/-ie
As shown above, the suffix -y/ie may be added to nouns in informal
style to denote people who are associated with what the stem
denotes; among the examples were foodie and winie. An identical
suffix is attached to nouns as a stylistic marker merely to express
familiarity, informality and close community. Thus unlike the pre-
vious -y/ie suffix, this one has no definite meaning besides indicat-
ing familiarity.
It may be added to noun stems as in e.g. doggy (dog), auntie
(aunt), froggy (frog), drinky (drink), girlie, and also in many first
names like e.g. Charlie, Johnny, Teddy. It often operates in conjunc-
tion with so-called clipping (cf. p. 159) as when breakfast becomes
brekkie/brekky, television becomes telly, goalkeeper becomes goalie, and
nightdress becomes nighty/nightie.
Writer and writing represent the two main types of verb nominaliza-
tions: those involving the ‘doer’ of the action, i.e. the agent or
instrument (writer), and those denoting the activity itself and its
results (writing). There is also a third less common but still active
type, i.e. nominalizations denoting the ‘receiver of the action’—
sometimes referred to as the patient—as in e.g. employee.
The types writer and writing are also very common in combina-
tions like letter-writer ‘person who writes/is writing letters’ and letter-
writing ‘the activity involving the writing of letters’. The noun letter
obviously corresponds to the direct object in the syntactic para-
phrases of these nouns: a letter-writer is someone who writes (pred-
icate verb) letters (direct object), and letter-writing is the activity of
writing (predicate verb) letters (direct object).
Nouns like letter-writer and letter-writing may be analysed in either
of two ways On the one hand it could be argued that the suffixes
-er and -ing belong to the entire noun-verb combination. In such an
analyses we would assume that these words consist of a verb letter-
write and the suffix -er or the suffix -ing. The problem with such an
interpretation is obviously the fact that we have to postulate a non-
existing verb *letter-write.
The alternative analysis—which is the one adopted here—is to
regard letter-writer and letter-writing as noun+noun compounds in
which letter is combined with writer and writing, respectively. This is
a more plausible analysis of letter-writer and letter-writing in that all
the elements involved exist independently of each other: letter,
writer and writing are perfectly normal English words.
However, the compounding approach does not entirely manage
to avoid the difficulties of the first analysis. Thus in words like e.g.
stealer and book-stealer. (See pp. 130–31 for a discussion of the prob-
lems this raises for the analysis of compounds).
Although -er is the dominant agentive suffix in English, there are
cases in which it faces competition from other suffixes with the
same meaning, i.e.-ant (assistant, attendant, defendant), -ent (solvent,
repellent, dependent) and -or (actor, escalator, governor). The compet-
ing suffixes are all from the Latin/French part of the vocabulary
and, as the examples indicate, they are generally only found with
stems of the same origin. In fact, most of the words in -ant and -ent
are early loans from French, and many words in –or are loans from
Latin.
Adjective-forming suffixes
Adjectives from nouns
Among the adjective suffixes, those combining with noun stems are
by far the largest group. The suffixes fall into six relatively clear
semantic groups: POSSESSING, LACKING, RESEMBLING, COVERED WITH/
FULL OF, CHARACTERIZED BY, ASSOCIATED WITH/HAVING THE PROPERTIES OF
There is a certain overlap among these groups; in particular it is
not always easy to distinguish between the last two groups. The first
five groups use Germanic suffixes while the final group uses non-
Germanic suffixes (cf. Chapter 1 for ‘Germanic’ and ‘non-Ger-
manic’).
POSSESSING: -ed16
There are two kinds of possession. In the first kind, what is pos-
sessed is a natural part of the possessor. This is the possession we are
talking about when we say that dogs have legs, humans have heads,
arms and minds, cars have wheels, etc.17 This kind of possession—
called inherent or inalienable possession—underlies a highly regular
and productive type of English word-formation, exemplified by flat-
footed (person), long-haired (girl), blue-eyed (boy), long-eared (donkey),
white-roofed (house), four-wheeled (tractor). The adjectives in this cat-
egory always contain two words joined by a hyphen, the first mod-
ifying the second. When uttered in isolation, such words often
have a distinctive stress pattern in that the primary stress falls on
the first syllable of the second element: flat-fóoted, blue-éyed, etc.)
In the second—non-inherent—type of possession, what is owned is
not part of the ‘owner’. This is the type of possession that is
involved when we say, for instance, that our neighbour has two cars
or that a student has three books. Possession of this kind can never
be expressed by adjectives of the flat-footed, blue-eyed etc. type: it is
simply not possible to speak of a *two-carred neighbour or of a *three-
booked student, simply because we do not regard cars or books as
natural parts of human beings.
The formation of adjectives expressing inherent possession is one
of the most regular and productive word-formation processes in
English. As the additional—authentic—examples below indicate,
the possessors do not have to be humans, and what is possessed
may be anything naturally associated with the possessor18:
two-legged (animal), different-haired, strong-willed, abstract-minded,
blue-shirted, white-gloved, blue-helmeted, bowler-hatted, different-
skinned, five-sided, three-wheeled (car), four-masted (ship), long-keeled
(sailboat), shoebox-sized (flat), duck-shaped (bottle).
LACKING: -less
Just as there are adjectives denoting the possession of something,
there are also adjectives denoting the lack of something. There is a
single suffix used in creating such adjectives from nouns, i.e. -less.
This suffix combines more or less freely with both concrete and
abstract nouns. The following examples19 illustrate some of the
range of the -less suffix: baseless (charge), breathless, boneless (beef),
CHARACTERIZED BY: -y
The adjective-forming suffix -y is very productive. In addition to the
meanings described earlier, it also appears with the somewhat
vague meaning ‘characterized by’ and ‘having the characteristics
of’. Admittedly, it is sometimes difficult to distinguish between
adjectives with ‘characterised by’ meanings and those with ‘full of/
covered with’ meanings. The stems to which -y is added are mostly
nouns, but it also turns up in combination with verbs, and occa-
sionally even with adjectives.
Combinations with noun stems are the most productive type,
and especially in informal spoken English, it is easy to get the
impression that almost any noun stem may be involved. However,
there seems to be a preponderance of short Germanic words Exam-
ples: ballsy ‘possessing balls’ i.e. ‘bold’ courageous’, brainy ‘charac-
The -ian suffix seems to prefer stems of more than one syllable,
although there are examples with monosyllabic stems like e.g.
Bachian [bɑkiən] and Wellsian ‘in the manner/style of Bach/Wells’.
As pointed out on p. 89, the suffix -ese is fairly productive when
used to create nouns denoting the jargon of certain writers, groups
of people or types of text, as in Johnsonese, officialese, etc. The adjec-
tival suffix -ese is chiefly used to form adjectives linked to place-
names as in Japanese, Genoese, Burmese, Vietnamese, Taiwanese, Can-
tonese etc. Although the number of such adjectives is quite high,
this use of -ese can hardly be claimed to be productive.
The suffix -esque means roughly ‘in the manner/style of’ and is
used with a certain amount of freedom mainly together with names
of prominent artists: Kafkaesque, Daliesque, Danteesque, Kiplingesque,
Hornbyesque, but it has competition from many other suffixes,
mainly -ian, cf. e.g. Orwellian, Huxleyan, Wagnerian, Lodgian. The
suffix -esque is less common than its competitors: it is often used for
stylistic effect, as in the e.g.: ‘… something Ludlumesquely titled “The
Vienna Project”21.
The suffix -ic is common in adjectives of mostly Latin origin like
atomic, idyllic, poetic. Sometimes -ic may be replaced by -ical without
any change in meaning. However, in a few cases words in -ic and
-ical differ in meaning: economic ‘to do with the economy’: econom-
ical ‘money-saving’, historic ‘famous’: historical ‘on record in his-
tory’, classic ‘great, memorable’;: classical ‘classical music, classical
languages’, electric ‘powered by electricity’: electrical ‘relating to
electricity’. The -ic suffix is productive in combination with Latin
stems in scientific English.
The suffix -ite is more common as a noun suffix than as an adjec-
tive suffix: words like Blairite, Clintonite, Thatcherite, Trotskyite are
more likely to be nouns meaning ‘follower/supporter of Blair etc’
than adjectives. This suffix is less common than its competitor -an/
ian.
The suffix -ous is found in a number of words like e.g. adventurous,
envious, furious, humorous, venomous. These words look as if they
were combinations of present-day English elements (adventure +
-ous, etc.), but most, if not all, of these were borrowed from French
during the Middle English period. The suffix -ous is also common in
loans featuring bound stems, particularly in words taking -ion as a
noun ending. As a result, there are plenty of pairs like e.g. ambition:
Verb-forming suffixes
Verbs from nouns and adjectives: -ify, -ise/ize
The formation of new verbs in English is mostly handled by means
of conversion (for which see Chapter 6) and back-formation (Chapter
9). However there are two common verb-forming suffixes in
present-day English, i.e. -ise (-ize) and -ify that are attached to both
adjective and noun stems
Verbs formed from adjectives by means of -ise/-ize all have the
meaning ‘turn something into what the stem denotes’, for example
Americanise, centralise, criminalise, equalise, formalise, legalise, mar-
ginalise, modernise, tenderise ‘make (meat) tender’. This is sometimes
also the meaning when the suffix is attached to nouns, for example
in vaporise and victimise. But in many cases the verbs formed from
nouns have less simple and predictable meanings. Thus symbolise
means ‘act as a symbol’, hospitalise mean ‘put into hospital’, and
dieselise means something like ‘convert to diesel-powered power’.
The recent formations computerise and itemise mean ‘convert to a
computer-operated system’ and ‘present data as a list of individual
items’.
The suffix -ise/-ize is also sometimes added to place-names, creat-
ing words meaning ‘make something like that place’ in Balkanise,
Israelise, Londonise, for instance. There are also more unorthodox
uses of -ise/-ize, as for instance in the monarchy has Diana-ised itself 22
recently found in a major British newspaper.
The suffix -ify has the same meanings as -ize, but is less common.
It is found with adjective stems in e.g. amplify, simplify and with
nouns in beautify, gasify, personify. On occasion, -ify is used with a
pejorative meaning, for example in speechify ‘make long and useless
speeches’.
Adverb-forming suffixes
Adverbs from adjectives: -ly
There are few adverb-forming suffixes in English, but there is little
doubt that one of them is the most productive English suffix of all,
i.e. the suffix -ly, which is added to adjectives, typically to form
manner adverbs. Thus adverbs in -ly are formed not only from mor-
phologically simple adjectives like brave, nice and slow, but also
from participial adjectives, for example in The professor was charm-
ingly inaccurate, The film is killingly funny, a jaw-droppingly ambi-
tious project, She was unashamedly frank about it, an outspokenly
honest account.23
There are few constraints on the use of adverb-forming -ly The
best known one is formal: as a rule, -ly is not attached to adjective
stems that themselves end in ly. For this reason it is rare to find
manner adverbs like kindlily (kindly+-ly) or friendlily (friendly + -ly).
Obviously there is also a semantic constraint: manner adverbs are
normally only formed from adjectives used to describe acts, actions
Exercises
1 As Chapter 5 shows, there are many homonymous derivational
suffixes in English. How many are there in Chapter 5? Are cer-
tain forms more prone to be homonymous than others?
2 What are the characteristics of productive derivational suffixes?
On what grounds is e.g. -th declared to be ‘dead’?
3 Where does the main stress fall in the following pairs and tri-
plets of words and why does it fall where it does? Johnson: John-
sonesque, Eton: Etonian, racket: racketeer, author: authorize: author-
ity: authorization, history: historic.
4 What is the difference between the two suffixes -ful discussed in
this chapter? Can you find other examples of words with these
two suffixes?
5 What is the pronunciation of word-final -sion in derision, implo-
sion, propulsion, transmission? Can the pronunciation of -sion be
predicted?
6 Several types of -ese suffix are discussed in this chapter? Which
of them is productive?
7 The function of the -y/ie suffix in e.g. auntie, doggy, nightie is
rather different from the functions/meanings of most other
English derivational suffixes. Can you explain how -y/ie differs
from the others and can you find further examples of this suf-
fix?
8 How productive would you say that the suffixes -er and -ee really
are? Can you find instances of verbs which do not appear to
take these suffixes at all?
6 Conversion
and large this is a less important aspect of conversion. There are also
instances of items from the open word classes that have been con-
verted to closed, for instance the prepositional use of present parti-
ciples like concerning, regarding, as in e.g. Do we have a problem con-
cerning/regarding money?
Conversion takes both simple, derived, and compound words as
input. In the case of derived words, it may overrule the word-class
identification provided by suffixes. Thus the presence of the noun
suffix -ion does not prevent words like e.g. caution and commission
from being used as verbs as in e.g. The prime minister cautioned
against tax increases, The new office building was commissioned yester-
day.
There are many examples of conversion from compounds, for
example the verbs to bottom-line and to cold-shoulder (other spellings
may be found) from the compound nouns bottom line and cold
shoulder discussed on p. 114 in this chapter. Occasionally, conver-
sion operates on parts of grammatical constructions as in He is an
also-ran, Don’t you my good man me.
Productivity in conversion
If the productivity of conversion is measured only in terms of word-
class change, this type of word-formation is close to 100% produc-
tive, in particular when it involves the open word classes. Thus
most nouns may be used as verbs, a great many verbs may be used
as nouns and adjectives can be turned into nouns and into verbs.
But in order to be considered truly productive, conversion must
also produce an output with predictable meaning. There are many
conversion patterns for which this is in fact the case, like e.g. the
conversion to verbs of nouns denoting instruments and tools.
These verbs always have one and the same meaning, i.e. ‘to use the
instrument or tool in typical fashion’: to fax is to use a fax to send
messages and to hammer is to use a hammer (or a hammer-like
instrument) to drive in nails.
In fact, conversion works only because it can be safely assumed
that speakers have a certain knowledge of the world that will help
him/her understand a new converted word. Thus even speakers
who have not previously encountered the verbs to fax or to vat will
PROVIDE WITH
Examples of ‘provide with’ meanings are e.g. carpet (a room), wall-
paper (a room), wall (off) an area, arm the people, fuel (up) a car, plas-
ter a wall, line a coat, floor a room and roof a house. Typically there
is an obvious and expected connection between the thing that is
provided and the space that is provided with it. To carpet or wallpa-
per a room is natural in most English-speaking cultures. It would be
considerably less natural to carpet or wallpaper a car or a safe.
REMOVE FROM
Noun to verb conversion with this meaning is only possible with
nouns denoting entities that are by definition part of a certain type
of ‘owner’ (a relation called inherent or inalienable possession in
Chapter 5). As examples of such inherently possessed nouns we may
mention bark (by definition part of a tree) and shell (by definition
part of e.g. nuts, peas, mussels). In exceptional cases, such inher-
ently possessed entities may be removed from their owners: we may
take away the bark from a tree or the shell from a nut. It is such
cases that the ‘remove from’ verbs are used to describe. If there is no
inherent possession involved, the meaning of the new verb will be
of the ‘provide with’ type.
Further examples of the ‘remove from’ type of conversion is
found in constructions like to core an apple ‘remove the core from an
apple’, bone a fish, skin an animal, top a tree, brain an animal. There is
a limited need for such verbs and the pattern can hardly be claimed
to be very productive. In addition, this conversion type has compe-
tition from prefixes like like un-, de- and dis- discussed in Chapter 4.
TRAVEL BY/SEND BY
The verbs in the TRAVEL BY/GO BY category are closely related to the
instrumental verbs. As their name indicates, they have to do with
means of transportation and include verbs like bicycle, canoe, motor
meaning ‘go by bicycle, etc.’, and also chopper ‘send/go by helicop-
ter’, mail ‘send by mail’. Sometimes the connection with the origi-
nal noun has been lost. A case in point is the verb ship, which
means simply ‘send goods to buyer by whatever means available’,
i.e. no actual ship need be involved. On account of this it is not
uncommon to encounter sentences like The article you ordered was
shipped on Feb. 22 by first class mail. There is also an intransitive verb
as in The new version will ship in a month or two.
ACT AS/LIKE
The meaning ‘act as/like’ is found in both transitive and intransi-
tive converted verbs. Thus you may bully (somebody), pilot (a
plane), host (a meeting)‘act as host for’, chair (a meeting) ‘act as
chairperson for’, police (a meeting), referee (a football match).
Intransitive uses are found in e.g. to lord (it over the others) ‘act in a
superior and domineering manner’ and to swan (around) ‘move
about in a casual and relaxed way typically perceived as irresponsi-
ble or ostentatious by others.5 The notion of ‘acting as/like some-
body or something’ is a vague one and it is sometimes not easy to
understand just what it means, as in the cases of e.g. badger (some-
body) ‘repeatedly and annoyingly ask someone to do something’,
ferret (out something) ‘search tenaciously for something’. Although
not highly productive, the ‘act as/like’ pattern is still available for
new formations.
CHANGE INTO
Straightforward examples of converted verbs with ‘change into’
meaning are easy to find: cash (a cheque), wreck (a car), remainder
(books), bundle (one’s clothes), slice (a melon). Often a figurative
meaning creeps in as an alternative to the literal one. The verb
trash, for instance, may mean ‘change into trash’, ‘wreck’, but it can
also mean ‘criticize severely’ in both British and American English.
The British English verb rubbish, on the other hand, is used only
with the meaning ‘criticize severely’. The ‘change into’ pattern has
limited productivity and faces competition from suffixed verbs in
-ize/-ise like atomize and carbonise.
ACTIVITY/INSTANCE/RESULT
The most productive type of verb to noun conversion is that pro-
ducing nouns denoting the activity described by the verb, a single
instance of the activity, or the result of that activity (it is often
impossible to distinguish between these): a fight may be the activity
of fighting (She was wounded during the fight), a cut may be both an
instance of cutting (He divided the loaf with a single cut of his knife)
and the result of cutting (Her arms had several cuts). Other examples
here are nouns like bite, bow, cut, fling, go, leak, look, pee, spin, try,
walk most often found in phrases like take a bite, give a bow, have a
fling, take a leak, take a look, have a pee, go for a spin (in the car), have
AGENT/INSTRUMENT
As shown previously, the most common way of forming agent and
instrument nouns is by adding the suffix -er (and sometimes other
suffixes like -ant, -ent and -or) to a verb: somebody who writes is a
writer, etc. In a small way, the same meaning may be expressed by
conversion as in a bore (‘a boring person’), a cheat (‘a person who
cheats’), a snitch (‘a person who informs on others’).
Partial conversion
At the beginning of this chapter, conversion was defined as a type
of word-formation involving no formal change in the word it oper-
ates on. Rather contradictorily, linguists sometimes also use the
term ‘partial conversion’ to describe word class change that is
accompanied by formal change. Partial conversion is used in three
such cases:
push up or put down a number of things, but the nouns púsh-up and
pút-down can only be used with reference to an exercise for
strengthening arm muscles and a critical and humiliating remark,
respectively. Likewise the verb melt down could be used about e.g. a
snowman or an ice statue, but the noun melt-down can only be used
about an accident in a nuclear reactor.
Exercises
1 What are the similarities and the differences between conver-
sion and derivational suffixation?
7 Compounding
Compound
Modifier Head
black board
text book
accident prone
its uses, -man is phonologically marked in that it has lost its word
stress and is pronounced [mən].
Semantically the meaning of -man is specialised in that it is used
to refer to people belonging to a certain profession or occupation
(denoted by the stem). For a long time it was taken for granted that
the representatives of these professions etc were all male, but in
recent years a distinction between the sexes has been made possible
by the introduction of forms like policewoman, congresswoman, etc.
A later development is the introduction of sex-neutral forms in
-person, sometimes providing a choice between three different
forms, two of them sex-explicit and one sex-neutral, for example
chairman/chairperson/chairwoman, salesman/salesperson/saleswoman,
spokesman/spokesperson/spokeswoman.
Two recent pseudo-suffixes are -babble and -speak. They are both
used to describe the kind of language typically used by certain
groups of people, but also indicate that the speaker takes a critical
view of both the group members and their language (especially in
the case of -babble). Among the examples of -babble and -speak for-
mations we find academic-babble, counselling-babble, hippie-babble,
agent-speak, business-school-speak, Bush-speak, City-speak, Foreign
Office-speak. In addition to being added to stems that are words,
both -babble and -speak are also often attached to bound classical
stems, as in e.g. eco-babble/eco-speak, Euro-babble/speak, psycho-bab-
ble, techno-babble used with reference to the jargons of ecologists,
Eurocrats, psychologists and experts on technology.
which the final word (the head) is a noun, not when the final ele-
ment is an adjective or a verb2.
Compounds and phrases are similar in that both consist of a
modifier (black) and a head (board). They also differ in significant
ways, however. In the opinion of many linguists, the most obvious
difference has to do with word stress: noun compounds have pri-
mary stress on the modifier (the first element), while noun phrases
have primary stress on the head (the second element)3. Thus not
only blackboard, but also other compounds like blackbird and dark-
room have the stress pattern [ - -], while the phrases black board,
black bird and dark room are stressed [- '-].
The difference in stress patterns has often been claimed to be
accompanied by a difference in meaning. According to this theory,
compounds are like simple words in being ‘labels’ (cf. p. 56) i.e. in
being used to refer only to special categories of ‘things’ in the (real
or imagined) world. Such an explanation is certainly true of many
compounds with noun heads, in particular adjective+noun com-
pounds like blackboard, blackbird, hothouse, etc:.a blackboard is not a
board that happens to be black, but a particular type of teaching aid
and parallel arguments can be made for blackbird and hothouse. In
fact in none of these combinations does the adjective have any
descriptive value: it is for example perfectly possible to speak of
green blackboards, albino blackbirds and cold hothouses.
The meanings of the syntactic phrases are quite different: black
boards, black birds and hot houses do not make up distinct catego-
ries of ‘things’. Furthermore, the adjectives involved have kept all
their descriptive force: black boards, black birds and hothouses
really are black and hot, respectively, which means that it would be
contradictory to speak of *green black boards, *albino black birds and
*cold hot houses.
As the discussion above has indicated, the combined meaning-
stress argument above works well for the adjective+noun combina-
tions: if they have primary stress on the adjective, they also tend
to have labelling meaning and may thus be called compounds. If
they have primary stress on the noun, they tend not to have label-
ling meaning and are accordingly syntactic phrases. Inevitably,
there are exceptions, for example combinations like white paper,
Yellow pages and small beer all of which have clearly labelling
meanings, but in which primary stress falls on the noun. (A white
tight above the ground on which acrobats perform their feats’ and
‘permit allowing non-US citizen to remain and work indefinitely in
the US’.
As could be expected, the phrase interpretations of tight rope and
green card have primary stress on the second noun, while in the
compound interpretations, primary stress falls on the first noun: in
other words, the placement of primary stress tells us which combi-
nations are compounds and which are phrases. But the spelling pro-
vides no guidance whatever in this matter.
In addition, in compounding as elsewhere, English spelling turns
out to be quite inconsistent. Consider for instance the items busi-
nessman, business-man and business man. There are three different
spellings involved, but from the point of view of meaning all three
must count as one and the same word, an interpretation confirmed
by the stress pattern. Other examples of the same phenomenon are
e.g. girlfriend:girl-friend:girl friend and teapot:tea-pot:tea pot. Obvi-
ously we would not want to argue that businessman and business-
man, girlfriend and girl-friend, teapot and tea-pot are compounds, but
that business man, girl friend and tea pot are not.
From these examples we may conclude that if both semantic and
stress-related evidence indicate that a certain word combination is a
compound, then the way it is spelt is irrelevant. On the other hand,
spelling may be helpful when we wonder where to place the stress
in written forms: if a written form is made up of two shorter words
and is spelt as an uninterrupted word, then a good guess is that it is
a compound and should accordingly be stressed on the first ele-
ment.
Exocentric compounds
As we have seen, one of the main characteristics of endocentric
compounds is that the entire compound basically belongs to the
same general category as the rightmost word, the head: a textbook is
a kind of book, to headhunt is to hunt in a certain way, originally to
collect the heads of dead enemies, now to find a suitable person for
a business position.
There is another smaller group of compounds containing com-
pounds like hardback, paperback, for which this is no longer the
added to verbs and since there are no verbs *to tax-pay, *to sheep-
steal, *to dress-make and *to train-spot, it seems wise to abandon this
line of reasoning altogether.
Whatever the solution to this problem may be, it is clear that
words like tax-payer etc constitute a challenge to the definition of a
compound as the combination of two or more words to form a new
word.
Adjective compounds 7
Type 2: far-reaching/hard-working
The type 2 compound adjectives consist of an adverb modifier and
a head that is a present participle. They have syntactic meaning: a
person who is hard-working is a person who works hard. Primary
stress as a rule falls on the modifier (cf. hard-working [ - -] but the
word stress pattern is more variable than in Type 1. Stress on the
head is often found in combinations like far-reaching and outgoing,
and often also in e.g. left-leaning, slow-moving.
Most compound adjectives of this type are gradable and may be
modified by intensifiers like very, extremely, quite, etc. as in e.g. very
far-reaching consequences, extremely hard-working people, incredibly
high-flying plans, highly left-leaning views, very rapidly-changing ideas.
The modifiers in these adjectives may be used in the comparative
and the superlative as the examples the biggest-selling book and a
higher-grossing product show.
Type 4: far-fetched/much-needed
The modifier in this very productive type of compounding is often
one of a rather limited number of adverbs expressing spread, dis-
tance and the like: examples are far-fetched, far-flung, wide-spread,
high-flown, deep-seated. The modifier may also be an adverb of fre-
quency or degree as in e.g. nearly-packed (stadium), almost-finished
(game), often-quoted (expression), understated, overblown, or an adverb
of manner as in neatly-folded (napkins), firmly-held (conviction). Note
that far-fetched, far-flung, high-flown, nearly-packed, almost-finished,
overblown, understated have primary stress on the second word.
Type 1: user-friendly/lead-free/work-shy
These compounds are made up of a noun modifier and an adjective
head. The meaning expressed by all such compounds is that some-
thing or somebody has the qualities of the adjective with regard to
the referent of the noun head: a user-friendly gadget is a gadget that
is ‘friendly’ with regard to users, a work-shy person is shy (reluctant)
when it comes to working.
Examples: accident-prone, lead-free (petrol), environment-conscious,
capital-intensive, childproof, crashworthy, machine readable, user-
friendly, carsick, trigger-happy, street-smart, work-shy. This is a very
productive pattern, and certain adjectives in particular enjoy great
Type 2: sea-green/red-hot
These compounds are made up of a noun modifier and an adjectival
head. Their meaning is basically comparative: sea-green means ‘as
green as the sea’, stone dead ‘as dead as (a) stone’, etc. Further exam-
ples of this type of compound are ash-blonde, sea-green, bottle green,
blood-red, snow white, pitch black, stone-cold, stone-deaf, stone-dead,
grass-green, ice-cold. These compounds all have primary stress on the
second word. There is a small set of informal compound adjectives
with similar form and meaning, but in which the first element
seems to have more of an intensifying force, for instance dirt-cheap,
piss-poor, red-hot (and possibly also a form like shit-scared).
Compound verbs
The term ‘compound verb’ suggests that there exists a compound-
ing process that combines two existing lexemes to form a com-
pound verb. However, the majority of verbs that look as if they were
compounds have not been created by combining two existing lex-
emes to form a third, but have been derived from compound
nouns, either by back-formation or by conversion.
When such verb are created by means of back-formation, the
starting-point is compound nouns ending in -ation, -ion, -ing, -er or
-or. The suffix is removed and the result is a new, back-formed verb:
from back-formation we get the verb to back-form, etc. Further exam-
ples will be found in Chapter 9.
For seemingly compound verbs that are actually cases of conver-
sion (cf. Chapter 6), the origins are compound nouns converted to
verbs. Examples of verbs formed in this way (and already men-
tioned in Chapter 6) are bottom-line, fast-lane, fast-track, keyboard,
mainstream, network, videotape, wallpaper. Other instances of such
Compound adverbs
With the exception of the examples in the section on exocentric
constructions (pp. 127–129), all compounds discussed so far have
been endocentric, i.e. the word class of the head word has also been
the word class of the finished compound. Thus textbook and green
card [ rinkɑd] are noun compounds because book and card are
nouns, and awe-inspiring and lead-free are adjective compounds
because inspiring and free are adjectives. Although there were few
directly formed verb compounds, the word class membership of the
compound verb agreed with that of its final element.
But in the case of adverbs, such correspondance is extremely hard
to find. With the exception of the informal British example double-
quick ‘very quickly’ there don’t seem to be any compound adverbs
whose final members are themselves adverbs.
There is no lack of adverbs that are combinations of words: a very
important type of adverb is the type exemplified by words like
mind-bendingly (boring), heart-breakingly (sad), jaw-droppingly9 (excit-
ing). But these are all derived forms created by adding the suffix -ly
to the adjective compounds mind-bending, heart-breaking and jaw-
dropping.
Another type of adverb created by the combining of words are
those consisting of a locative preposition and a following noun10.
Common prepositions in these cases are up and down as in e.g.
downmarket/down-market, upmarket/up-market, downstairs, upstairs,
mid-market, upmarket/up-market, upstream, downstream, upscale.
The majority of these combinations may be used both as adverbs
and as adjectives, but in most of these cases, the stress patterns are
different. As adverbs these compounds have primary stress on the
Exercises
1 Try to find additional examples of formations in -man/-person/
-woman and -babble and -speak. Can you think of other forms
that should be regarded as pseudo-suffixes?
2 In addition to the forms -man, -babble and -speak used as exam-
ples of ‘pseudo-suffixes’, there is the form -gate, found in e.g.
Dianagate, Irangate, Iraqgate, Saddamgate. Try to find out what
common element of meaning these words have and why they
have this particular meaning (clue: cf. the Watergate scandal of
1972). Are words formed by the addition of -gate of the same
type as those formed by the addition of -man, -babble and
-speak?
3 The following word combinations all qualify as adjective+noun
compounds on semantic grounds, since their meanings are not
predictable from the meanings of the individual words that
make them up: black sheep, blue grass, blue jeans, bluejay, blue-
print, fine print, grey (gray) area, grey (gray) matter, small print,
smart bomb. To what extent is their status as compounds con-
firmed by their stress patterns, and what do they mean?
4 Using the criteria for noun+noun compounds discussed on pp.
123 and 125–126, which of the following noun-noun combina-
tions would you consider to be compounds?: bomb scare, bomb
disposal, chocolate biscuit, chocolate box, fishing-rod, fish farm, fish
finger, salmon-fishing, winning post, winning streak. Motivate your
answers.
5 Indicate how you would analyse compounds consisting of more
than two words, for example war crimes investigation, open class
words, talk-show host murder, jet airline pilot strike, spare parts sup-
ply management system.
General characteristics
In our earlier discussion of word-formation mention has sometimes
been made of a phenomenon called neo-classical compounding and
of the so-called combining forms of which such compounds are com-
posed. In this chapter we will return to this highly productive type
of English word-formation in somewhat greater detail. However,
this is not the place to give a full account the neo-classical com-
pounds, and all the brief discussion below can claim to do is to
draw attention to their main characteristics and to point to some of
the ways in which they are similar to and different from affixation
and ordinary compounding.
The combining forms are bound base morphemes from Latin and
Greek. The ‘combining’ part of the name refers to the fact that such
forms are bound stems combining (in principle only) with each
other to create words like e.g. astronomy, biology, democracy, economy,
astronaut, democrat and others. Such combinations have certain
basic similarities with ordinary English compounds in the way the
elements they consist of are related. Both may be said to consist of
a modifier and a head: astronomy, biology and astronaut, for example,
are basically the same as ‘star science’, ‘life science’ and ‘star sailor’.
However, for all their semantic parallelism, the neo-classical com-
pounds also differ fundamentally from ordinary English com-
pounds in consisting of bound stems: it is true that both biology and
life science are similar in the way the meanings ‘life’ and ‘science’ are
combined, but the English compound consists of two words while
the neo-classical compound is made up of the two bound stems
bio- and -logy.
syllable from the beginning of the word (because that is now the
third syllable from the end): cf e.g. astronomy, psychology, kiddology,
democracy. Predictably, the addition of a fifth syllable moves the
stress one further syllable to the right as for example in aristocracy
literally ‘rule by the best’, a word formed from Greek elements in
Old French.
The workings of the antepenult stress rule are sometimes dis-
turbed by other factors, in particular by the presence of suffixes
affecting stress placement (cf. p. 81). Consider for instance the
adjective democratic formed from the same elements as democracy
and democrat. However, while the last two words are stressed in
accordance with the antepenult stress rule, democratic is not, but
has primary stress on the second syllable from the end of the word.
The reason for this is the presence of the suffix -ic ; like quite a few
other English suffixes (cf. pp. 81–83), -ic requires primary stress to
be placed on the syllable immediately preceding it.
As the discussion above has shown, there are several good reasons
to distinguish between prefixes and initial combining forms. How-
ever, there is no denying that certain linguistic forms seem to
belong to both camps, i.e. to be used both as prefixes and as ICFs,
for example the forms extra- and hyper-.
These forms quite clearly combine with stems that are words in
e.g. extra-territorial, extra-kind, hyper-correct, hyperinflation. Further-
more, they carry at least secondary stress in all combinations in
which they occur, they do not end in -o, and the words in which
they occur do not follow the antepenult stress rule. These are all
excellent reasons for regarding them as—quite productive—pre-
fixes.
However—as we have already observed—extra- and hyper- are just
as clearly ICFs in words like extravagant and hyperbole, where they
combine with bound stems, have somewhat indistinct meanings,
and are stressed in accordance with the antepenult principle
(extravagant is pronounced [k strvəənt] and hyperbole
[ha pəbəl]. The conclusion is that we have to accept the existence
of a fair amount of homonymy between prefixes and initial com-
bining forms.
Like the original initial combining forms, the telescopic forms have
a strong tendency to end in -o, but there is sometimes variation
between -i and -o for example in words containing the stem agr-:
the form used in telescopic combinations is sometimes agri- some-
times agro-, both meaning ‘agricultural’. Thus there is on the one
hand agribusiness, from which we get agriproducts, and on the other
agrobiology, agrochemical, and even agrobiotechnology.
The strength of the ‘final -o principle’ is also obvious in what may
be termed ‘fake telescopic forms’, i.e. bound forms in -o that have
not been created by clipping operations on neo-classical com-
pounds, but are—etymologically incorrect—reductions of other
words like e.g. globo-cop ‘police force of the entire world’ (formed
from global), robo-cop ‘policeman that is a robot’ (from robot) and
enviro-friendly ‘friendly towards the environment’.
Initial forms
Exercises
1 Look up the following neo-classical compounds in a large Eng-
lish-English dictionary that provides both pronunciation and
etymologies. Find the meanings of the elements that make up
the words and determine whether the stress pattern agrees with
the antepenult principle: pleistocene, oxymoron, pentagon, carbo-
hydrate, appendectomy.
Anglophile:Anglophobe:Anglophobia
acid:acidic:acidophilus:acidophilic
5 In Chapter 8 it was found that forms like astro-, bio-, etc. may be
both ordinary initial combining forms and so-called ‘telescopic’
forms. Can you think of final combining forms that may be
used as telescopic forms?
9 Irregular Word-formation
Borrowing
Strictly speaking, borrowing has nothing to do with word-forma-
tion: it is a way to enrich the vocabulary resources of a language by
importing foreign lexical material. However, borrowing is not
entirely unrelated to word-formation: what is borrowed, in certain
cases, is not merely words or phrases, but also the patterns and
affixes underlying word-formation in a foreign language. One
example of pattern borrowing is the construction el --- o, borrowed
from Spanish and used in American English with the meaning ‘the
most …’, as in el cheapo (‘the cheapest’), el creepo (‘the greatest creep,
most like a creep’), el sleazo (‘the sleaziest’)2.
Another loan from Spanish is the informal savvy from Spanish
sabe usted ‘you know’. On its own, savvy is used both as an adjective
meaning ‘competent, knowledgeable’ and as a noun meaning
‘knowledge, know-how’. It is used in word-formation to form adjec-
tive and noun compounds like for example computer-savvy, E-savvy,
street savvy, tech savvy. Thus a person may possess computer-savvy, E-
savvy (E = ‘electronics’) or he/she may be computer-savvy, E-savvy.
Meaning extension
Not everybody would regard meaning extension or semantic
change, as it is also called, as a kind of word-formation. Yet undeni-
ably the extension of the meanings of existing words is a very pro-
ductive way of forming new words. In the present book it has been
regarded as a form of irregular word-formation, i.e. basically as
word-formation for which no principles may be established; others
take a different view9.
There is a certain superficial similarity between meaning exten-
sion and the type of regular word-formation called conversion. Like
conversion, meaning extension brings about a change of word
meaning without an accompanying change in form. But at that
point the similarity ends: meaning extension does not change the
word class of a word and it has none of the regularity and predicta-
bility of meaning found in conversion.
Meaning extension operates in several different ways. In one of
its manifestations, it is a continuous historical process of semantic
change that affects all languages and gradually changes the mean-
ings of their vocabulary items. This kind of change is very slow and
speakers often take a long time realising that the meaning of a word
has gradually changed.
It has often been observed10 that the gradual type of meaning
change mentioned above can be described in terms of two diamet-
rically opposed semantic tendencies. On the one hand, it may lead
to generalization or widening of meaning, as in the case of the Eng-
lish word box, which originally denoted only small containers
made of box-wood, but has gradually acquired its present meaning
of ‘any container with a flat base and sides, typically square and
having a lid’11 On the other hand, it may lead to specialization or
narrowing of meaning. An often used example of this is the word or
meat, which originally denoted food in general and only later took
on the meaning ‘edible flesh’.
But semantic change may take place much more quickly: the
word gay for instance used to mean ‘happy’, ‘carefree’ until the
1960s, but is now exclusively used with the sense of ‘homosexual’,
a change that started in American English in the 1960s12 The new
meaning of gay has now replaced the old one more or less com-
pletely.
Abbreviations
In writing, abbreviations consist of strings of letters, which may or
may not be separated by full stops. Most of them are spelt with cap-
ital letters, although there are those that are not. Many are proper
names of institutions or places, for instance CIA (Central Intelli-
gence Agency), FBI (Federal Bureau of Investigation), NYC (New
York City), UN (the United Nations), ANC (African National Con-
gress), GOP (the ‘Grand Old Party’ = the American Republican
Party), GB (Great Britain). A few abbreviations refer to—usually
famous—people, for example FDR (Franklin Delano Roosewelt), JFK
(John Fitzgerald Kennedy), GWB (George Walker Bush).
Other abbreviations are non-names: DJ/dj (disk jockey), DNA
(deoxyribonucleic acid), TEFL (the teaching of English as a foreign
Acronyms
Turning now to the acronyms, we may begin by noting one differ-
ence between these forms and abbreviations: since acronyms have
to be pronounceable, they can only contain letter sequences also
permitted in ordinary English words (cf. the discussion of phonotac-
tic rules on p. 52).
Like the abbreviations, acronyms often denote institutions and
organizations:
NAFTA (North American Free Trade Association), NATO (North
Atlantic Treaty Organization), UNESCO (United Nations Educa-
tional, Scientific and Cultural Organization), WAC (Women’s Army
Corps). Certain acronyms naming organizations have obviously
been composed as a (more or less obvious) reminder of the purpose
of the group, like e.g. GASP (Group against Smokers’ Pollution) and
AIM (American Indian Movement), PUSH (People United to Serve
Humanity).
Among the acronyms that do not name institutions we find e.g.
AIDS/aids, SAD (Seasonal Affective Disorder)13, soho (small office
home office) RAM (random access memory), ROM (read only mem-
ory). The following denote groups of people with certain character-
istics: yuppie (young upwardly mobile), GLAM/glam (greying, lei-
sured, affluent, married) and WASP/wasp (white, Anglo-Saxon,
Protestant).
Clipping
English has a number of ways to reduce word length. In the previ-
ous section we discussed one of these methods of word shortening,
i.e. the creating of initialisms. This technique makes it possible to
reduce multiword units like the names of organizations or products
to strings containing the initial letters of the words in the original
multiword unit.
But shortening processes may also operate on single words by
removing part of the word, an operation known as clipping. In the
most common type—back-clipping—the final part of the word is
removed, as for instance in lab(oratory), porn(ography) and
prof(essor). Other examples of back-clipping are ad(vertisement),
admin(istration), advert(isement), beaut(y), biog(raphy), bod(y),
celeb(rity), demo(nstration), decaf (‘decaffeinated coffee’), ref(eree),
info(rmation), intro(duction), perk (>perquisite ‘fringe benefit’),
perp(etrator ‘person who has committed a crime’), rep(resentative),
veg(etable), Jag(uar), Merc(edes), dorm(itory), cert(ainty), rev(olution),
tech(nology).
Note that back-clipping usually operates only on certain senses of
a word: the back-clipping rev (plural revs), for instance, can only be
used in the technical sense ‘an instance of revolving’; rev is not a
back-clipping of revolution ‘political upheaval’ and consequently we
cannot say e.g. *the Russian Rev meaning ‘the Russian Revolution’,
only the engine was turning over at 300 revs a minute. (There is also a
verb to rev, which means to make an engine revolve at a faster pace:
She was revving her engine). Cert for certainty is mostly used in the
phrase It’s a (dead) cert!. The clipping Brit is predominantly used to
Back-formation
Back-formation creates new words from old by removing a noun
suffix from words that contain typical noun endings like e.g. -ation,
-er, -or, -ing, -ism and -ist. It is a process that has been at work in Eng-
lish for hundreds of years, the classic example being the verb edit
formed from the borrowed Latin word editor. The first attested
occurrence of the noun is from 1649, while the verb makes its first
appearance only in 179115. However, back-formation is very much
alive today and is responsible for more recent formations like e.g.
the verbs (to) orientate from orientation, (to) obsess from obsession,
and (to) opt (for something) from option.
Back-formation bears a superficial resemblance to clipping in that
both serve to create shorter words from longer ones, but the two
processes differ both with regard to their output, to the elements
that are removed from the original word, and to their function. The
output of back-formation is a new lexeme with a different meaning
and word-class membership than the original. The output from
clipping is a word that is shorter than the original but belongs to
the same word class and has the same basic meaning. The output
words are synonymous with the longer original forms, but occupy a
quite different stylistic niche: clippings like e.g. fab and demo are
entirely appropriate in laid-back communication as when you’re
chatting with your friends, but only the full forms fabulous and
demonstration will do on more formal occasions.
A further difference is that clipping may remove parts of a word
that are not meaning-carrying as in the examples biog and info cited
above. As we have seen, back-formation operates in terms of real or
perceived suffix morphemes: it is a process that changes what is per-
ceived as a complex word (normally a noun) into a shorter and
more basic one, normally a verb.
The examples given of back-formation so far have all involved
single words: from orientation we derived orientate, from obsession
we got obsess. To these we can add further examples, like to liaise ‘to
establish a working relationship with’ from liaison, to laze from the
adjective lazy, to peddle from the noun peddlar and to televise from
television.
However, in present-day English, back-formations are much more
often formed from compounds than from single words, usually by
the removal of the nominal suffixes -er, -ation and -ing from the sec-
ond part of noun compounds. (As nouns in -er normally also are
found as -ing-forms, it is often impossible to determine exactly
which of the two was the original). The great majority of the new
back-formed verbs are hyphenated.
The following list contains examples of fairly recent verbs result-
ing from the application of back-formation to noun compounds:
baby-sit<baby-sitter, back-form<back-formation, bird-watch<bird-
watcher (or from bird-watching), blow-dry<blow-drying/blow-dryer,
bottle-feed<bottle-feeding, carbon-date<carbon-dating, co-drive<co-
driver, gift-wrap<gift-wrapping, headhunt<headhunter, house-
keep<house-keeper, hunger-strike (verb)<hunger strike (noun), knife-
murder<knife-murderer, shoplift<shoplifting, stage-manage<stage-man-
ager.
Sometimes there is phonological evidence that back-formation
has taken place, for instance in the verbs contracept, cohese, intercept
and self-destruct, which have obviously been created from contracep-
tion, interception, cohesion and self-destruction. If that had not been
the case, the forms of these verbs would be *contraceive, (the actu-
ally existing) interceive, cohere, and *self-destroy.
Reduplicative compounds
Reduplicative compounds are sometimes more aptly called ‘echoic
words’17. These constructions are combinations of two words that
somehow echo each other in that the second word repeats the first
with a slight change in pronunciation/spelling. There are two cate-
gories. In one the change affects the vowel in the first word, while
the consonants are kept intact, as for instance in flip-flop, riff-raff,
shilly-shally, tick-tack/tick-tock, wishy-washy, zig-zag.
In a second, equally common type, the vowels stay the same, but
one of the consonants is changed, as for example in hanky-panky,
higgedly-piggedly, hobnob, hoity-toity, nitty-gritty, pitter-patter, roly-poly,
teeny-weeny/teensy-weensy.
Rhyming slang
Rhyming slang is a type of slang that can be traced back to the 18th
century and is particularly associated with Cockney English. The
formation of expressions of this type of slang is a two-stage process.
The first stage involves finding a two-word phrase, the last word
of which rhymes with the target word, i.e. the word whose meaning
we want to express. (In actual fact a lot of such two-word phrases
have already been created and are unlikely to be replaced by other
phrases). The two-word phrase is then used instead of the target
word to create a kind of code in accordance with the description
below:
But the process above may then be taken one step further in that
the final word in the phrase—the one rhyming with the target
word—is dropped:
The coding is now complete and users of rhyming slang can pro-
duce utterances supposedly only understood by their fellow users,
like e.g. Take a butcher’s at that car!, Bring your china!
One of the chief uses of rhyming slang has been to provide
euphemisms for offensive or taboo words. Thus Richard the Third is
used instead of turd (‘shit’), Bristol Cities for titties, Hampton Wick for
prick, Berkshire/Berkley Hunt for cunt18.
The above examples of rhyming slang are of long standing and
you sometimes get the impression that this type of word-formation
is no longer productive. However, as the following examples indi-
cate, it is clearly the case that—to a certain extent at least—rhyming
slang is still productive: Adam and Eve ‘leave’, bangers (and mash)
‘cash’, Britney Spears ‘beers’, do bird ‘do time’ (in prison), from
(bird)lime, which rhymes with time, John Major ‘pager’, Steve
McQueen’s ‘jeans’ and (whiskey and) soda ‘mobile phone’ (soda
rhymes with Voda as in Vodafone)19.
References
Notes
Chapter 1
1 The American Heritage Dictionary of English.
2 The notion of sign and the distinction between motivated and
non-motivated signs was first formulated by the Swiss linguist
Ferdinand de Saussure in the early 1900s (cf C.Bally and Albert
Sechehaye (eds.) Cours de Linguistique Générale, 1st ed. 1916,
Paris: Payot. Translated into English as Course in General Linguis-
tics New York: Duckworth.
3 These definitions are from The New Oxford Dictionary of English.
4 The place of English and the other Germanic languages in the
larger context of the Indo-European language family is tradi-
tionally illustrated by a family tree diagram like the one below
adapted from Robertsson-Cassidy 1954:33.
Indo-European
West East
5 The data about the barking of dogs comes from the list Sounds of
the World’s Animals Copyright Catherine Ball at
www.georgetown.edu/cball/animals/cat.html). Another example
making the same point may be found in Pinker (1995:152), i.e.
Chapter 2
1 Extensive discussions of word classes will be found in Quirk et.
al. 1985 and The Cambridge Grammar of the English Language. Cf.
also Crystal 1997:206–213, Lyons 1968:317–29, Pinker
1995:105–6.
2 Assessments of the total number of words in English (and other
languages) vary wildly, as do estimates of how many words dif-
ferent categories of speakers know. For some of these, see Crys-
tal 1997: 118–125 and Pinker 1995:150–51.
3 Crystal 1997:213.
4 This quote is from Rombauer-Becker Joy of Cooking 1962 Bobbs
Merrill, Indiananpolis, Indiana.
5 For a full and interesting discussion of different types of class
membership cf. Taylor Linguistic Classification 1995 2nd ed
Chapter 2–3
6 These and the other criteria in this chapter are based on the
word-class classification in Greenbaum 1996, but basically
reflect mainstream grammatical thinking on these matters.
7 Cf. also the discussion in Chapter 6, where it is pointed out that
the word-formation process known as conversion may override
suffixation: a noun like e.g. position may be used as a verb as in
e.g. They positioned the table between the chairs.
8 There are also verbs followed by a direct object and a so-called
objective predicate complement, for example call, consider, regard as
in e.g. She called me a fool, We consider him a genius, They regard
this as a terrible mistake.
9 Unlike the other word classes, the adverb class is divided
between the open and the closed type: there are sets of adverbs
that are closed like e.g. those denoting definite time (now, then)
and place (here, there). But there are also those that are clearly
open, in particular the class of manner adverbs formed by the
addition of the suffix -ly.
10 The definition of determiner used here is along traditional lines;
cf. e.g. Greenbaum 1995, Burton-Roberts Analysing Sentences
(3rd ed.) page 154, Quirk et.al.1985:253–7. Huddleston and Pul-
lum 2001 The Cambridge Grammar of the English Language.
11 The standard work on grammaticalization is the book Grammat-
icalization by Paul J. Hopper and Elizabeth ClossTraugott CUP
1993.
12 These examples are from Hopper-Traugott 177–178.
Chapter 3
1 As the discussion of words like defrost, unsaddle on pp. 63 in
Chapter 4 will show, there are also prefixes that change the
word class of the stems they are attached to.
2 Certain suffix forms turn up both in inflection and in deriva-
tion, in particular -ed and -ing, which are found both in inflec-
tional verb forms (They talked, They were talking) and in derived
nouns and adjectives like flat-footed, carpeting, etc. Adjectives
like amused and amusing (as in an amused smile, amusing stories)
and nouns like a painting, the painting of the walls pose a partic-
ular problem in this respect (for discussion of these questions cf.
Chapter 2 p. 41, Chapter 6 p. 116).
3 MacArthur 1992:263
4 All languages have rules—sometimes called phonotactic rules—
determining what sound and letter-sequences are permitted in
the language. For information about the phonotactic rules for
English, see Traugott-Pratt 1980:61–4 and Cruttenden-Gimson
1994:216–23.
5 Cf MacArthur The Oxford Companion to the English Language
p. 876
6 This definition is from The Oxford Dictionary of English.
7 A related but not identical question is under what circum-
stances a potential word may become an actual one. Cf. Kjellmer
2000 for a discussion of this.
8 According to Saeed 1997:71, modern dictionary makers do not
always follow this principle but simply list all derived forms
even if their meaning is predictable.
9 For more information about lexicalization, see e.g. Bauer
1983:48–61.
10 The distinction between syntactic repackaging and labelling
used here has been borrowed from Kastovsky 1986 (and else-
where). Kastovsky uses syntactic recategorization instead of syn-
tactic repackaging.
Chapter 4
1 Most verbs containing ‘removal’ dis- like for instance discourage,
dismantle, dismember were borrowed from French during the
Middle English period.
2 This point is made in Adams (2001:71) who has also supplied
the examples used here. Cf. also Bauer (1983:37–38)
3 For the historical background of English prefixes, see e.g.
Marchand 1969:129–34
4 Stress assignment in English is notoriously difficult to describe.
There are several treatments of English stress in general and
word stress in particular in the literature, for instance Bauer
1983 Chapter 5, Gimson 1994 Chapters 10–12, and Minugh
1991, Chapters 5, 7–9.
5 For extensive discussions of productivity in word-formation, cf.
e.g. Adams 2001 (7–10, 146–53), Bauer 1983, Chapter 4, Bauer
2001, Kastovsky 1986. See also The Cambridge Grammar of the
English Language p. 1630. In this book, productivity in word-for-
mation is regarded as a cline extending from highly productive
to weakly productive processes. No English word-formation rule
is 100% productive, but some, like the rules for adverb-forming
-ly and noun-forming -ness come close.
6 From one of John Cleese’s Monty Python sketches.
7 The principle that negative un- doesn’t combine with simple
nouns is (very occasionally) violated in formations like e.g. un-
person/unperson, which means roughly ‘a (usually politically)
impossible person’ i.e. someone who is not accepted by the
establishment.
8 In George Orwell’s pessimistic vision of a totalitarian Britain in
the novel 1984, one of the prominent features is the simplifica-
tion of English word-formation permitting among other things,
words like ungood and even double-plus-ungood meaning ‘very
bad’. See Crystal 1997:135 for an account of Orwell’s fascinating
‘Newspeak’.
9 Cf. Kastovsky 1986:594, who provides the example ‘The Time-
Patrol also had to unmurder Capistrano’s great-grand-mother,
unmarry him from the pasha’s daughter in 1600, and uncreate
Chapter 5
1 For more detailed accounts of the interplay between suffixation
and word stress, cf. The Cambridge Grammar of the English Lan-
guage (CGEL pp. 1669–72 and Bauer 1983:116–20), Minugh
1991, Lass 1987:204–10.
2 The account given here is based on Bauer 1983:112–22.
3 Cf. Ljung 1974:26.
4 These examples are from Adams (2001:64)
5 The examples creditcarditis and televisionitis are from NODE and
AHD respectively. The rest of the examples have been taken
from The Independent on CDROM from 1992. It turns out that
this use of the -itis suffix goes back to at least the early 1900s:
according to the Supplement to the Oxford English Dictionary,
Asquith used the term fiscalitis as early as 1903, when writing in
The Westminster Gazette.
6 These examples are from Adams 2001:62.
7 Cf The Cambridge Grammar of the English Language p. 1691,
Bauer 1983:268, and Marchand 1969:245–7.
Chapter 6
1 Accounts of conversion will be found in e.g. Marchand 1969,
Quirk et.al. 1985: Clark and Clark 1979, Adams 2001 and the
Cambridge Grammar of Modern English
2 MacArthur 1992:263.
3 In most cases of conversion it is possible to determine its direc-
tion i.e. what is the input word and what is the output. Thus in
word pairs like hammer(n): hammer (v) and fax(n): fax(v) it rea-
sonable to assume that the verb has been formed from the
noun, not the other way round. But there are also many
instances of homonymous word pairs for which such decisions
are impossible, for example hate, love, count.
4 For discussion of such categories (cf. e.g. Adams 2001:22, Bauer
1983, Clark and Clark 1979).
5 Both definitions from The New Oxford Dictionary of English.
6 From le Carré Smiley’s People. 1980. London and Sydney:Pan
Books p. 172.
7 As mentioned in Chapter 2 (p. 41), verb to adjective conversion
is probably the best explanation for adjectives like astonished,
amusing, breath-taking, etc. As noted in The Cambridge Grammar
of the English Language, p.1644 this is an unusual case of conver-
sion using inflected forms as its point of departure.
8 From the Financial Times, March 11, 2003, p. 3.
9 For further examples of such verb:noun pairs, see Minugh
1991:227–8.
10 Note that partial conversion from phrasal verb to nouns with
‘agentive’ or instrumental meaning may take many different
forms. Thus in addition to the type drop-out, drop-in, there are
the types in-drop, dropper-in and dropper-inner (the final example
usually found only in informal spoken English). Cf. Bauer
1983:288–90.
11 This example has been borrowed from Pinker 1995:129.
Chapter 7
1 There is a fair amount of disagreement among linguists con-
cerning the proper definition of ‘compound’in English. For an
early treatment, cf. for example Jespersen 1942, A Modern Eng-
lish Grammar Part VI pp. 134–183 and Marchand’s classic Cate-
gories and Types of Present-Day English Word-Formation. 1969:11–
30. More recent treatments will be found in Bauer (1983:106–9),
Bauer (1988:33–7, 100–104), Quirk et.al. (1985:1330–32, 1567–
70), The Cambridge Grammar of the English Language (2001:449–
51, 1644–66), and Adams (2001: 4–5). Note also that the so-
called ‘reduplicative compounds’ like argy-bargy, riffraff etc. are
dealt with in Chapter 9.
2 For this point cf. The Cambridge Grammar of the English Language
p. 447.
3 There is no total agreement about the role of stress in noun com-
pounds. Thus Bauer 1983:104–105 and The Cambridge Grammar
of the English Language (p. 451) have no faith in the role of stress
placement as a predictor of compound status. The latter argues
that the compound:phrase distinction can only be established
on syntactic grounds (pp. 449–50). Others are in favour of
regarding stress as at least an important tool in the recognition
of noun+noun compounds, for instance Marchand (1969:28–
29), Quirk et. al (1985:1568–69) and Adams (2001:80).
4 All the examples of compounds given here consist of just two
words. However, it is clear that both compounds and phrases
may consist of three or more words, for example weapons inspec-
tion crisis, or small arms factory owner. Such compounds have to
be analysed on several levels, the first into weapons inspection
and crisis and then into weapons and inspection, the second into
small arms factory and owner, then into small arms and factory
and finally into small and arms. (Small arms ‘firearms that can
be carried in the hand’).
5 For syntactic compound (sometimes also called synthetic com-
pound), see Bauer 1988:36, Adams 2001:78–9, The Cambridge
Grammar of the English Language p. 1652n
Chapter 8
1 They are not restricted for linguistic reasons—classical Latin and
Greek had very rich vocabularies—but because latter-day speak-
ers of English cannot be assumed to know all the classical forms
that could in principle be used to form neo-classical com-
pounds.
2 I owe the terms ICF and FCF to Bauer 1983. The vast majority of
the combining forms are initial: of the 550 plus combining
forms listed in The New Oxford Dictionary of English, more than
400 are initial forms.
3 Kleptocracy was found in Newsweek.
4 This is the term used in MacArthur 1992:233
Chapter 9
1 For discussions of back-formation, cf. Adams 2001:136–8,
Hughes 2000:54
2 From the Bloomsbury Dictionary of Contemporary Slang
3 The example of uber has been taken from The Oxford Dictionary
of New Words
4 There is a very extensive discussion of the suffix(es) -nik in
Bauer 1983:259–66.
5 The information about the suffix -bot is from The Oxford Diction-
ary of New Words
6 The information about -(s)ville is from Chapman 1986 (New Dic-
tionary of American Slang: Harper & Row)
7 From The Independent News on CDROM 1992.
8 From NEWSWEEK March 25, 2002, p. 58.
9 For example Beatrice Warren in Warren 1992.
10 in e.g. G.L. Brook A History of the English Language pp. 177–180,
Bloomfield Language p. 426 to give just a few sources)
11 This is the definition of box in The New Oxford English Dictionary.
12 The change in meaning of gay from ‘carefree’ to ‘homosexual’
and ‘homosexual man’ became established in the 1960s accord-
ing to The New Oxford Dictionary, The American Heritage Diction-
ary of the English Language, Crystal 1997:138.
13 See the entry for SAD from in the Oxford Dictionary of New
Words.
14 The clipping blog and its derivatives blogger, blogging was first
found in Newsweek May 27, 2002, p. 79. Later issues of News-
week also had blogoshere and blogspeak.
15 According to the OED
16 Cf. Adams 2001:139.
17 This term is from MacArthur 1992:854-55. Discussions of redu-
plicative compounds will also be found in Bauer 1983:213,
Marchand 1969, $ 8.2.1, Adams 2001 127-29, Katamba 1994:
79, Jespersen 1942, part VI pp 173-183. Cf. also Hughes 2000:61
Discussion of exercises
Chapters 1–8
There is no discussion of exercises of a fact-finding nature. This is
indicated by means of a dash (–).
Chapter 1
1 In motivated signs there is a natural connection between form
and meaning while there is no such connection in non-moti-
vated (arbitrary) signs. Flag signals are usually non-motivated
and so are most human sign languages, although they also have
certain motivated characteristics. Animal territory marking uses
motivated signs.
2 Grammar is concerned with changes in form with predictable
effects on the meaning of an utterance. This makes it possible to
write rules for grammar. Vocabulary is concerned with the rela-
tion between the form and meaning of linguistic signs. In a
vocabulary consisting only of simple words, that relation is arbi-
trary and consequently not predictable, and no rules can be
written about it. For non-simple words see the answer to ques-
tion 4 below.
3 Onomatopoeic words are intended to imitate certain sounds
like e.g. animal calls. Sound symbolic words on the other hand
contain (sequences of) sounds supposed to be associated with
certain meanings.
4 Unlike simple words, words created by word-formation rules
have meanings that are partly predictable: we know, for
re call un truthful
truth ful
the same lexeme BE. In this count, plural suffixes, past tense suf-
fixes, apostrophe s suffixes and n’t as in Don’t and ‘s as in name’s
have been counted as inflectional endings. The numbers would
be different if we counted n’t as not and ‘s in My name’s not Jack
as an instance of be. We might also wish to count at once as a
single complex lexeme.
10 The non-Germanic languages that have had the greatest impact
on English vocabulary are French, Latin, Greek.
Chapter 2
1 A word class is a class of words sharing certain important char-
acteristics having to do with syntactic function, the grammati-
cal categories expressed, and meaning. These characteristics are
also used to define the different word classes. The semantic
(notional) criterion is problematic because it is vague and less
comprehensive than the others.
2 Content words have denotation. They denote phenomena
found in ‘the real world’ or some imagined world, like people,
objects, processes, actions, qualities and others. Function words
do not have denotation, but are mainly used to link content
words in different ways. Most function words are limited in
number and do not readily accept new members: they are said
to belong to the closed word classes. There is no upper limit to
the number of content words and the word classes they belong
to are accordingly known as open classes.
It is possible to leave out certain function words because they
are predictable: leaving them out does not deprive the text of
essential information. Leaving out the content words, on the
other hand, would make the text incomprehensible.
3 It may for instance be argued that count nouns – especially
those denoting concrete phenomena – are more central to the
noun category than uncountable nouns.
4 Proper nouns lack denotation and are used as names, i.e. they
refer directly to different phenomena. Unless they are recatego-
rised, they do not normally take articles or the plural. (Note that
Chapter 3
1 Complex words: reclassify, unbelievable, Londoner, Bushie(s), Clin-
tonian.
Compound words: water crisis, sun-tanned, arms cache, steam-
roller. In this book, the word Clintonomics is analysed as a blend
(cf. Chapter 9) combining Clinton and the segment -onomics
from the word economics. An alternative analysis is to treat Clin-
tonomics as a suffixed word consisting of Clinton and -onomics.
2 The fact that forms ending in noun suffixes like steamroller and
audition may be converted to verbs indicate that when suffixa-
tion and conversion clash, the latter wins.
3 The problem with noun-verb pairs like love(n):love(v),
fear(n):fear(v) is that it is difficult – without turning to the his-
torical evidence – to determine the direction of the conversion,
i.e. what should be regarded as the starting-point for the con-
version process and what as the result. Other words presenting
the same kind of difficulty are e.g. count, play and whistle and
many others. In such cases it is not worth the trouble to try to
establish the direction of the conversion.
4 Strictly speaking, English does not have the type of word-forma-
tion known as infixation, i.e. insertion of bound morphemes
inside words. However, it has often been pointed out that Eng-
lish has a similar construction, making possible the insertion of
‘swearwords’ like bloody and fucking into words with at least
three syllables, as for example in e.g. absobloodilutely, imfucking-
possible. As the examples show, the insertion place is immedi-
ately before the syllable with main stress in the original word.
The effect of this type of insertion is to add emphasis to certain
words.
5 The permitted consonant combinations – known as clusters – at
the beginning of English words are listed below. Note that the
symbols used are phonetic and that sometimes there are several
spellings corresponding to a single cluster. Note also that certain
rare combinations occurring in loanwords have been excluded.
Permitted word-initial two-consonant clusters: [bj], [bl], [br],
[dj], [dr], [dw], [fj], [fl ], [fr], [gl], [gr], [gw], [kj], [kl], [kr], [kw],
[pj], [pl], [pr], [sf], [sj], [sk], [sl], [sm], [sn], [sp], [st], [sw], [ʃr],
[ʃm], [θr], [tʃ]. Examples of written words with these clusters:
beautiful, black, bride, duke, drive, dwell, few, flat, from, gloat,
grade, Gwendolyn, cute, class, croak, quick, puke, pluck, prim,
sphinx, suit, school, slow, smart, snow, speak, stone, swallow, shrink,
schmuck, throw, cheat.
Permitted three-consonant clusters: [skl], [skr], [skw], [spl],
[spr], [str] as in sclerosis, scream, sprawl, squeamish, spleen, street.
6 We use our knowledge of word-formation productively when
we actually create a word that is new to us. We use it analytically
when we analyse an existing complex or compound word.
7 –
8 Chief among the factors that determine whether a derived,
compound or converted word is entered in the dictionaries are
the word’s frequency and the question whether its meaning is
predictable or not. Words that are very infrequent or have
Chapter 4
1 The ‘removal’ prefixes de- and un- untypically change the word
class of the stem they are attached to. This is also true of certain
other prefixes like e.g. out-.
2 A fully productive prefix combines freely with all stems belong-
ing to a certain well-defined class of stems. If new members are
added to that class of stems, a fully productive prefix is able to
combine with all of them. Fairly productive prefixes are less reli-
able in this respect, and weakly productive prefixes even less so.
3 –
4 When negative un- is attached to an adjective it creates a word
expressing the opposite of the original stem meaning, while
words created by the addition of non- merely deny the existence
Chapter 5
1 –
2 The definition of a fully productive derivational suffix is basi-
cally the same as that for prefixes, excepting the type of forma-
Chapter 6
1 Conversion is like (most cases of) suffixation in changing the
word class of the stem it is applied to. Conversion and suffixa-
tion also express much the same meanings, and are often ‘in
competition’ with each other. But – with the exception of ‘par-
tial conversion’ – word-formation by means of conversion does
not involve any formal change in the word to which it is
applied. Suffixation always changes the form of the word it is
applied to.
2 Truly productive conversion results in words with predictable
meanings as when instrument-denoting nouns are converted to
instrumental verbs. Such conversion is obviously based on the
assumptions speakers have about the way(s) different instru-
ments are typically used.
3 Like other types of regular word-formation, conversion is used
to create linguistic labels when its output is no longer predicta-
ble.
4 A noun like carpet is converted to a ‘provide with’ verb in e.g. to
carpet a room, while the noun bark can only be given a ‘remove
from’ meaning when converted to a verb as in e.g. to bark a tree.
These different interpretations have to do with the fact that car-
pet and bark have different relations to their ‘owners’, i.e. rooms
and trees: a room does not necessarily have a carpet, while a tree
must have bark (possesses bark inherently).
5 The formation of instrumental verbs from nouns denoting
instruments and tools exploits our natural expectations about
the way(s) the referents of such nouns are typically used. How-
ever, the meaning of the potential instrumental verb may
already be expressed by another word – a phenomenon known
as blocking. Thus an instrument noun like spade is not normally
converted to a verb to spade, since the verb dig already exists.
(Note however that there is a verb spade but with a different
meaning. Compare also axe, pen).
6 Eats ‘food’ : verb to noun conversion with plural suffix added,
graph ‘show in graph form’: noun to verb conversion, swot ‘per-
son who studies hard’: verb to agent noun conversion, dovetail
‘fit in with, agree with’: compound noun to verb conversion,
flatline ‘die’: compound noun to verb conversion The word
unmicrowaveability is the result of several word-formation proc-
esses: the compound noun microwave oven is abbreviated to
microwave (n), which undergoes conversion to the verb (to)
microwave ‘prepare (food) in a microwave’. To this verb the suf-
fix -able is attached, producing the adjective microwaveable. The
negative prefix un- is added and finally the noun suffix -ity is
attached yielding unmicrowaveability.
7 As in the case of Sydney, 1930s and upper class in question 2 in
Chapter 2, sand and Chicago are best regarded as nouns used
with attributive function, and let us all be friends as a phrase
with the same function. Regarding such cases as instances of
conversion to adjectives would force us to assign adjective sta-
tus to all words and word combinations that may occur attribu-
tively and predicatively, including most nouns.
8 –
9 The only productive use of ‘partial conversion’ is when it is
applied to phrasal verbs like e.g. show off – with main stress on
off – to turn them into nouns like a show-off ‘person who shows
off’ with main stress on show. As an alternative to calling this
Chapter 7
1 –
2 The form -gate comes from the name of the Watergate hotel in
Washington D.C. connected with the 1972 scandal that eventu-
ally toppled President Nixon. The -gate element then came to be
used in names for other political scandals. In such names, the
-gate ‘suffix’ is added to the name of a person, place, etc.
involved in the scandal, as in e.g. Dianagate, Camillagate, Sadd-
amgate, Irangate, etc. The words in -gate differ from those in
-babble, -man, -speak in that the element -gate has nothing to do
with the noun gate, while the meanings of -babble, -man and
-speak are clearly related semantically to babble, man, speak.
3 –
4 The combinations bomb scare (‘B is produced by A’ ), bomb dis-
posal, salmon-fishing (‘B is activity involving A’), chocolate box,
fishing-rod, winning post (‘B is used for A’) and fish farm (‘B pro-
duces/causes A’) meet both the stress criterion and the semantic
criterion for noun+noun compounds. Chocolate biscuit and fish
finger belong to the problematic group of ‘B is made of/consists
of A’ and have main stress on the fist element, as does winning
streak.
5 The tree-diagrams below give the analyses for war crimes investi-
gation and talk show host murder:
war crimes investigation talk show host murder
talk show
Chapter 8
1 –
2 Neo-classical compounds normally have primary stress on the
third syllable from the end – also known as the antepenult – and
this is what we find in the majority of the examples given here,
i.e. acidophilus, Anglophile, Anglophobe, automaton, automate,
cryptogram, cryptographer, cryptography, pathological, pathologist,
pathology. In acidic, acidophilic, automatic and cryptographic on
the other hand, the main stress falls on the second syllable from
the end of the word, a position due to the influence of the suffix
-ic (cf. Chapter 5).
3 –
4 Heli- in these words obviously stands for the whole word heli-
copter: heli-skiing is helicopter-assisted skiing, a helipad is a land-
ing pad for helicopters, etc. Heli- in these words must accord-
ingly be counted as a telescopic initial combining form, just like
the form bio- in words like biodegradable, where bio- stands for
the whole word biological(ly).
5 Word-final telescopic forms are much less common than word-
initial ones, but they do exist. One example is the form -naut in
chimp(o)naut ‘astronaut who is a chimpanzee’. In this example
-naut stands for ‘astronaut’, but its normal meaning is ‘sailor’, as
in e.g, astronaut ‘sailor among the stars’. (The example
chimp(o)naut is from Bauer 1983:271).
Index
abbreviation 157–158 endocentric compounds 121
acronym 157–159 exocentric compounds 127–
adjective 129
definition 42 noun compounds 129–134
adjectives in -ed and -ing 41, 116 spelling 126–127
adverb syntactic compound 129–131
definition 42–43 verb compounds 138–139
‘direction’ 105 conversion
intensifier definition 109–110
manner 42–43, 104–105 ‘act as/like’ 113–114
‘with regard to’ 105–106 ‘activity/instance/result’ 115–
affix 15, 49–50 116
antepenult 146–147 agent 116
‘change into’ 114
back-formation 161–162 instrumental verb 112–113
barking 13 location/storage 111–112
blends 163–164 partial conversion 117–119
borrowing 154–155 ‘provide with’ 112
‘remove from’ 112
clipping ‘send/travel by’ 113
definition 159
back-clipping 159–160 denotation 25
fore-clipping 160 derivation(al) 50, 80–81
cognate 11
combining forms etymology 11
definition 143
word-final 144 generalization of meaning 156
word-initial 144 Germanic languages 11
telescopic 148–149 grammaticalization 45–46, 122,
compound 191
definition 121–122
adjective compounds 134–139 head 50, 121–122
adverb compounds 139–140 homonym 22, 45, 62, 70, 80,
compound vs. phrase 123–126 116
idiom 25 prefix
immediate constituent 18 definition/function 63–64
Indo-European 11 ‘attitude’ 74
inflection 22 ‘degree/size’ 70–74
initialisms 157 ‘location/direction’ 74–75
irregular word-formation 153–167 manner 70
‘negation’ 67–68
Jabberwocky 54 productivity 66
‘removal’ 69
label(ing) 55, 56, 59 ‘time/sequence’ 75–76
lexeme 21–24 ‘reversal’ 68–69
lexical density 21 preposition
lexicalization 56, 126 definition 45–46
loans complex prepositions 24
from French 28
from Greek 28–29 recategorisation 38–40
from Latin 27–29 reduplicative compound 164–166
reference 25
Middle English 27 rhyming slang 166–167
Modern English 28
modifier 51, 121–123 sense 25–26
morpheme sign
definition 15 definition 9
base morpheme 16 motivated 10
bound morpheme 16 non-motivated (arbitrary) 10
free morpheme 16 simplified language 33
morphology 17 sound symbolic words 12
multimorphemic word 17 sound-imitating words 12
multiword unit 23–24 stem 16
suffix
neo-classical compound 143–151 definition 49, 79–80
nominalizations 93–95 adjective-forming 96–103
noun adverb-forming 104–106
definition 35 ‘agent/instrument’ 94–95
common noun 37 ‘activity’ 86–87, 95–96
count(able) noun 37 ‘amount’ 85
mass noun 37 ‘associated with’ 100–102
proper noun 37 ‘characterised by’ 99–100
uncountable noun 37 ‘collective’ 85–86
‘covered with/by’ 99
Old English 26 direction 105
onomatepoeic/peia 12 familiarity 91
List of prefixes
sub- degree/size 73
super- degree/size 73
trans- location/direction 75
ultra- degree/size 73
un- negation 67–68
un- reversal 68
un- removal 69
under- degree/size 73
List of suffixes
(WCF stands for ‘word class function of suffix, WCS for word class
of stem)