54 This chapter has looked at how linguists handle sound

'"
structure.' This is the first aspect which confronts anyone
o
c; working on an unwritten language. ln practice. sotting out the
j
a. sound structure overlaps with the analysis of Jarger units, such
'ti as worcls. This Ís the topÍc of the next chapter.
~
3
'" Questions
1 What do you understand by the term phoneme?
2 What is a minimal pair? Suggest at least ten examples tif

o
c.n
minimal pairs in English.
3 What Is an allophone? Glve examples.
4 What are distinctive features?
5 What is a natural class?
6 What is metrical phonology?

This chapter looks at the

problems encountered ·in
identifying and defining the
notion 'word'. It then
discusses the identificàtion
and description of
'morphemes' (piecas of
words). Finally, it looks at the
way ih which words can be
assigned to 'word classes'
(parts 01speech).
56 The wOId appears to he a widespread concept. Even in primitive A flea and a fly in a flue
cultures, informants are often able to identify words. This is Were imprisoned, 50 what could they do? ~
somewhat surprising, because nobody has yet proposed a Said the flea: 'Let us fly'. s.~
:l: .••
satisfactory universal ddinition of the notioo 'word', or Said thefly: 'Let us flee'. 00.
., '"
0. ••
pIOvided a foolproof method of identification. People 50 they flew through a flaw ilt the flue. '" :J
a.
sometimes wrongly assume that a word is recognizable because At the simplest levei, this rhyme Contains thirty-six written "ti
it represents a 'single piece of meaning'. But it can easiIy be words. But some of these are repeated. If we decide to leave out ~.
shown that this view is wrong by looking at the lack of repeats and count the number of different words (in technical
m

correspondence between words from different languages. ln terros, count word types instead of word tokens), we come up
English, the three words cycle repair outfit cortespond to one in against several problems. Should f/y (noun) and f/y (verb) be

ocn German, Fahrradreparaturwerk.zeuge. Or the six words He used

to live in Rome are translated by two in Latin, Romae
habitabat. And even in English, a word such as walked incIudes
counted as the saroe, since they sound the same, or as different,
because they have different meanings? Should fly and flew be
regarded as the same, because they belong to the same verb, or
at least two pieces of meaning, 'walk' and 'past tense'. as different hecause they have different forros? These problems
This chapter will deal with tbis marter. First, it wilI look at the can be solved onIy if we decide what kind of 'word' we are
problems of defining and identifying words. 5econd, it will talking about. Ir is important to distinguish between lexical
consider pieces of words, or morphemes. items, syntactic words and phonological words.
If by 'word' we mean lexical item (the technical term for
'dictionary entry'), then the sound sequence 1flaI! 'fly' represents
Defining words twowords, since most dictionaries have separate entries for fly
(noun, N) and fly (verb, V):
The best-known definition of a word is that proposed by the
American linguist Bloomfield, who defmed ir as a minimum free fly N: ao i1)Scctwith two wings.
forro, that is, the smaIlest form that can occur by itself. This is fly Y: to move through the air in a controIled manner.
fairIy unsatisfactory, because words do not normally occur by This is perhaps the most basic, and roost abstract use of the
themselves in spoken speech. Even if you ask a simple question, word 'word'. However, both of these lexical items have various
a normal-sounding reply often requires more than one word: syntactic forms associated with them. The insect could occur as
Who did that? John did. fly (singular) or flies (plural), and the verb could occur as fly,
What's that? An oak tree. f/ying, flies, flew, flown. 50 if we counted the various syntactic
Furthermore, some apparent words, such as did, the, and, are forros as different words, the overaIl total would he rouch higher
found alone only in exceptional circumstances, such as in (Figure 6.1).
answer to the question: 'What does a-n-d speIl?' Lexlcal Items Syntactic
f1y words
f10wn
f1aw
flias
mas
flying
Bloornfield's definition works best for written English, where we f1yV f1yN
conventionally leave a space on either side. But linguists are
concemed primarily with the spoken word, not the written, and
the two do not necessarily coincide. For example, it seems to he
purely accidental that the name of a certain type of snake, a úoa
constrictor, is written as two words rather than one, or that
seaside appeats as one word, but sea shore as two.
Why have linguists found it so hard to find a satisfactory
definition ôf the nOrion 'word'? The answer seems to be that
there are different types of word. Consider the rhyme: figure 6.1
 A further complication oceurs with a lexical item sueh as flaw. ruptib1e and mobile. These are useful guidelines in rrtany 59
58
This has the two syntactic fôrms flaw (singular) and flaws languages. A sequence such as chickens carmot be interrupted. It
a~ (plural). But the singular fonn flaw then has two different sound is impossible to say .,ehiek"little"ens, or .,ehieken-little"s. In
oao
:e
sequenees assoeiated with it, /f1:J:/ before a eonsonant, and Ifh:rl addition, the sequence chickens can move about. It can occur
alll before a vowel (Figure 6.2): next to different words, and in different parts of the sentence, as

I
1Il~
Q. in: Chiekens lay eggs, faxes eat chiekens, the ehiekens clueked
'O The flue had a flaw Ifbú which allowed the fly to escape. laud/y, and so on.
There was a flaw /fl:Jrr/ in the flue.
flaw
syntactic Iflo:J
/lIo:zJ words
Phonological
WOrd$ To take another example, suppose we had come across the
Lexicalltem sequence greentrousers, and wanted to know whether this was
flawN flaws IIlo:rl one or more words. We would begin by looking for sentences

o
cn
which included any part of the sequence greentrousers. We
might find: o
cn
Green leather trousers, Red trousers, Green shirts.
The fact that greentrousers can be interrupted by the word leather
figure 6.2 indicates that we are probably dealing with at least Mo words,
green and trousers. This suspicion is copÍirmed by noúng that both
green and trousers occur with other words. 'But since green and
These examples show that we must rt'ôt'expectAn exact overlap trouSers seem to be uninterruptible (wedo not find "trous-green-
between different types of word. Ancl in someother lânguages, ers, for example), we surmise that each is a word.
the situation is far more complex than in EngliSh. InLatin, for
example, the lexical item rosa 'rose' has. twelve different At the end of this stage of the analysis, we have a rough list oí
syntactic forms. In Welsh, rhe initial consonam of ~ach word 'words', thoug11 a list in which we are likely to have c1umped
varies systematically, depending mainly ou the precedmg sound: together diffhent lexical items which sound the same
the word for 'father' could be tad, dad, thad, Of nharl. lhe last (homonyms), and to have separated different syntactic forms of
lmes of the choros in a wel1~krtown ~elsh hymn have three the same lexical item.
different forms of the verb meaning 'sing': canu, ganu and For the second stage of the analysis, we need to consider the
ehanu - and there is a fourth possibility, nghanu, which the syntactic behaviour of these possible 'words', that is, their role in
hymn omits. the overall sentence pattem. For example, fly N would show up
as behaving differently from fly V, since each would fit into a
different 'slot' in the sentences:
Identifying words The fly buzzed.
For anyone working on an unkrtown language, it is important Birds fly.
tO identify these various types of word. There are two main On the other hand, fly and flew would tum out to be somewhat
stages in the analysis. First, finding chunks such as fly, flew, similar, in that they would fit into the same general slot:
which recur as self-eontained units. Second, deciding how many
lexical items are covered by each ehunk (as with fly, which They fly home on Sunday.
covers two lexical items), and conversely, deciding how many They flew home Ott Sunday.
different chunks belong to the same lexical item (as with fly, However, the syntactic behaviour of these different forms can be
flew, where different syntactic forms belong to one lexical item). supplemented by an analysis' of their make·up, or, in other
For the flIst stage, finding chunks which behave as self- words, the morphemes out of whieh they are eonstituted. Let us
contained units, we look for sequences which are uninter- therefore go on to consider some basic facets of morphology.
 60 Morphemes ln Turkish, the similarity between adam/ar, 'men', and kadmlar, 61
'women', eoables One to identify a plural suffix -lar, and the
5l.o~ The smallest syntactic unit is the morpheme. M()r~hemes va~y words for 'man', adam, and 'woman', kadm. In Swahili, the
~a in size. Neither syllables nor length are any gmde to the1r overlap between:
~ li
lEI!! identification. The essential criterion is that a morpheme cannat nitasoma I will read

i
ã. be cut up into smaller syntactic segments.
nilisoma I read (past)
'2. The sentence in Figure 6.3 has eleven morphemes: utasoma you wiIJ read
dream 3Ing
2974
8sleep
10
11 56y chanl
albatross
lullaby
a ed
walk ulisoma you read (post)
The

1 allows us tO identify soma, 'read'; ni, '1'; u, 'you'; ta, future

tense; li, past tense.
o
O') figure 6.3
Not alI morphemes are as easily segmentable as these examples.
But the identificanon of morphemes is dane wholIy by means of
o
O')
this one basic technique - the comparison of partially similar
utterances.
The, albatross, a, lullaby, are ali single morphemes because none
of them can be syntactically split up further. Alba- and -tross, for
example, do not have any other role to l?lay in t?e syntax of
English: they exist only as part of the smgle umt, albatross. Types of morpheme
Chanted and dreamy, on the other hand, each con.sist of two Morphemes such as aibatross, chant, lul/aby, which can oceur
morphemes: chant is found in .words. such as .chantmg, chants, by themselves as whole words are known as free morphemes.
and is also a word by itself, wh1le-ed 1Sfound m wanted, batted
and so on. Similarly, sieep-waiking consists of three morphemes, Those such as anti-, -ed, -iy, which must be attached to another,
because sleep, walk and -ing are all found elsewhere. In theory are bound morphemes. Bound morphemes are of two main
types. Considér the sentence:
there is no upper limit to the number of morphemes pe~ wor~:
antidisestablishmentarianism, for example, has at least s1X:ant,- The owllook-ed up at the cloud-y sky.
dis-estabiish-ment-arian-ísm.
SuperficialIy, both looked and cloudy have a similar make-up,
consisting of one free morpheme, folIowed by a bound oue. Yet
the bound morphemes differ in nature. -ed on the end of looked
Recognition of morphemes is an infleetional morpheme, since it provides funher
Linguists iden"tify morphemes by' comparing a wide variety of informá.tion about an existing lexical item look, in this case
utterances. They look for utterances which are partially the indicating that the looking occurred in the pasto Other examples
same (Figure 6.4): of inflectional morphemes are the plural, as in owls, aud the
and
snlfl·ed
for-wards
and
possessive, as in Peter's car. However, -y 00 the end of cloudy
dlnosau'r
dinosaur
back-wards
arroganl-Iy
plodd-ed
edg·ed Ioud-Iy
grunl-ed
The behaves tather differently. It is a derivational morpheme, oue
The which ceeares an entirely uew word. Cloud and cloudy behave
quite differently and fit into different slots in the sentence. Other
examples of derivational morphemes are -ness as in happiness,
figure 6.4 -ish as in greenish, and -tnent as in establishment.
In moSt cases, it is easy to telI the differenee between inflection
The partial similarity between sniffed, grunted, plodded ~d and derivation. Above alI, inflectional endings do not alter the
edged enables us to isolate the segment -ed. And the partial
syntaetie behaviour of an item in any major way. The word still
similarity between arrogantly and loudly, and between
backwards and forwards makes it possible to isolate -iy and - fits iuto the same 'slat' in the senteoce. Derivational endings
wards. create eotirely new words. In addition, itifleetional endings can
 be added ou to derivational ones, but not vice-versa. That is, we 0--' ,I [+ sibilam} - z. e.g./JDlszl--'lh:nslzI 63
62
find words súch as establish-ment-s, but not ••establish-s-ment.
ta -+ I [+ sibilant]-z
English has rclative1y few inflectional morphemes. }hese are on
the whole easy to identify, though they sometlmes present zero changes into J in lhe following between a sibilanl
(i.e.lnsert I) circuli1stances
prob!ems of ana!ysis, as discussed below. and~~
figure 6.5

Allomorphs
Sometimes a morpheme has only one phonological formo But
z --. s/[-voiee l-. e.g. Ik.retzl-+ Ik.retsl

c
O')
úequently it has a number of variants known as allomorphs.
Allomorphs may vary eonsiderably. Totally dissimilar forms [-YoieeJ- o
O')
may be allomorphs of rhe same morpheme. Cats, dogs, horses, z ehanges Inlo 5 in lhe following alter a yoiceless
sheep, oxen, geese alI eontain the English plural morpheme. eircumstances sound

Ao allomorph is said to be phonologically conditioned when i~s

form is dependent on the adjaeent phonemes. Ao allomorph IS figure 6.6
said to be lexically conditioned when its form seems to be a
purely accidental one, linked to a particular vocabulary item.
Note that these 'rules' must be applied in the arder given above.
The English plural morpheme provides excellent examples of If the arder was reversed, we would get forms such as "[dIjsJ
both phonologically and lexically conditioned allomorphs. Let instead of the correet [dIjlz] for the plural of dish.
us look at some of these.
Lexical cor1"ditioning
Phonological conditioning
Words slleh as oxen, sheep, geese present a problem. Although
The study of the ,different phonemic shapes of allomorphs is they function as pluraIs in the same way as cats, dogs, they are
known as morphophonology - sometimes abbreviated to not marked as pluraIs in the same way. 511Ch lexieal1y
morphonology. eonditioned pluraIs do not follow any speeific rule. Eaeh one
l-zll-sll-l71 are ali phonologically conditioned allomorphs of the has to be learnt separately.
English plural morpheme. That is, each allomorph oceurs in a Words such as oxen, sheep, geese can be identified as
predictable Set of environmeuts. syntaetical1y equivalent to the cats and dogs type of plural
l-v occurs atter most voiced phonemes as in dogs, lambs, bees. because they fit into the same 'slot' in a sentenee (Figure 6.7):
(A voiced phoneme is one in which the vocal cords vibra te, as in
The _ _~~_ . are making a 101of noise
/hl, IdI, IgI,/v/, and vowels.)
I-si occurs atter most voiceless phonemes, as in cats, giraffes,
skunks. (A Vóieelessphoneme is one in which the vocal cords do caIs
not vibrare.) dogs
horses
occurs atter sibilants (hissing and hushing sounds), as in
I-rz./
oxen
horses, chee$es, dishes.
sheep
li we take ltias basic, then we can say first, that 1·71turUs into geese
IrI1 aftee sibüanl:S (Figure 6.5), and second, into 1"51 alter
voiceless sounds (Figure 6.6):
figure 6.7
64 Oxen, sheep and geese each contain two morphemes: ways. For example, the word class traditionally known as 'verb'
ox + plural can be recognized as a verb partly because it oceurs after nouns
sheep + plural (or phrases containing a noun), and partly because most verbs
have an inf1ectional ending ·ed to indicate the past:
800se + plural
Arabella detesteJ snails.
But only oxen is easily divisible into two: Marianna smiled.
ox + I-eml (·en)
Careful ana1ysis is needed, because in some cases, items which
Sheep can be divided into two if a zero suffix is assumed. A 'zêro superficialIy appear to fit imo the same slot in a sentence can
suffix' is a convenient linguistic fiction which is sometimes used rum out to be rather different in character. Consider the
in cases of this type. It is normally written 10/: sentences:
o
O')
sheep+ 10/.
There is no obvious way to analyze geese. At one time, linguists
Charlie ate caviare.
Charlie ate well. ocn
suggested that the plural vowe1 li:1 in Igilsl (geese) whi~h At first sight, we might wrongly assume that caviare and well
replaces the IUlI in Igu: si (goose) should be regarded as a speclal
be10ng to the same word class. But a less superficial analysis
type of a1lomorph ealled a replacive. And they analyzed the reveals that they behave somewhat differentIy overall. If we
plural as: tried altering the sentences around, we could say:
Igu:sl + li:! +- (/Ul/). Caviare was eaten by Charlie.
Here the formula li:! +- (lu:!) means '/ill replaces lu:/'. What Charlie ate was caviare.
But this is rather a strained explanarion. These days, most But we could not form the equivalent sentences with well:
linguists simply accept thar rhe form Igi:s1(geese) represents two • Well was eaten by CharLie.
morphemes: ,.What Chaflie ate was well.
goose + plural These dissimilarities indicate that caviare and weLl are
and rhar these two cannor be separared. And a similar syntacticalIy different, and be10ng to different word classes.
explanation is required for forms such as went, took, which
It is not always easy to telI how many word classes a language
represenr:
contains. Many traditional textbooks c1aim that English has
go + past tense eight 'parts of speech'. But this claim turns out to be based
take + pasr tense. largely on old Latin grammars which were in tum translated
from ancient Greek grammars, which mostly divided Greek
words into eight word classes. If we look more c1osely,we find
Word classes several discrepancies. For example, nouns and pronouns are
traditionally c1assified as separate parts of speech, yet they have
In every language, there are a limited number of types of lexieal a large number of similarities:
item. These differenr kinds of word are traditionally known as
Max laughed.
'parts of speech', though in linguistic rerminology the label word
He laughed.
class is more common. Word classes are conventionaIly given
labels, such as noun, verb, adjective. In fact, nouns and pronouns are more alike than the different
Words are classified into word classes partIy on account of their types of word which are traditionaIIy labelIed adverbs. Words
syntactic behaviour, partIy on the basis of their morphological suchas quickLy and vety are both usualIy classified as adverbs,
formo That is words from the same word c1ass are likely to fit but they behave quite differently:
into the sam~ slot in a sentence, and to be inflected in similar He ran quickly.
"'Hera", very.
 66 The number of word classes varies from language to language. NOUh [+ N, - VJ
67
Some word classes, such as noun and verb, may be universal. Verb [- N, + VJ
Rut others vary. Nouns, adjectives and verbs are on a Adjective [+ N, + V]
continuum. At one end are nouns, words which maintain their Preposition [- N, - V]
identÍty over time, such as iree, cai, river. At the other end are 'J!1i~se~~s to be .a useful and economi~al way of capturing the
verbs, words which signify rapid change, as in walk, kick, push.
slmllarltles and dlfferenees between rhe major word classes.
In the middle come properties, such as large, beautiful, old. In
English, these form a separate word class, rhar of adjectives. Rur The major word. classes are known as lexical categories. Lexieal
this is not inevirable. Some languages rreat them as a type of eategones contam content words, those with intrinsic meaning.
verb, so-called stative verbs, ones which denote a state. Where They contrast with funetional categories, which include 'little
English says: wo~ds' whose me~ning is often difficult to specify, as the, a,
WhlCh are determmers (D), or the eomplementizer that in I
Petronella ;s happy. know that Paul is m, often ahbreviated to COMP or C.These
a language such as Chinese might say, as it were: funetion words are important for gluing pieces of sentenees
Petronella happ;es. together into longer syntaetic patterns.
using a verb instead of an adjeetive. English also sometimes flips This chapter ~as diseussed words and morphemes. These fit into
between verbs ànd adjectives. Compare the archaic he ails larger recurrmg patterns, which will be the topie of the next
(stative verb) with the modem day he ;s ill (adjeetive). chapter.

Questions
Major word classes
1 Suggest three different ways ín which the word word might be
English is sometimes considered to have four major word usado ;
classes: nOun (N), adjective (A), verb (V), preposition (P) (Figure 2 What Is a morpheme?
6.8). 3 Dístinguish between Inflectlon and derlvatlon.
Nwater
Vunder
P
5wim
frogs 4 What Is the difference between phonologlcally conditloned
Big
A allomorphs and lexlcally condltloned allomorphs?
5 How mlght one Identlfy word Classes?

figure 6.8

Of these foU! major classes, nouns, verbs and prepositions

behave fairly differently from one another, though adjeetives are
somewhat strange, in that they have some noun-like qualities,
and some verb-like ones. In Blessed are the brave, brave seetns
to have become a noun. And in Mavis is asleep, asleep seems
fairly verb~like, since it fies into the same slot as sleeping in a
sentence such as Mav;s is sleeping.
Ir has becn suggested that we should describe these four word
classes in a manner parallel to the distinctive Jeature
descriptions used for sounds, whieh can show shared
similarities.