Chapter 1 Introduction

lmagirie a life without words! Trappïst monks opi tor .it. But most. of us wouid nat give .up words for anything.. Every day we utter
thousands and thousands of words. Communicaiing our joys, fears, opinions. fantasies, wishes, requests, iemands, feelin d the
occasionai threal .or insult—is. a very important aspect of being human. The air ïs always thick w'ith our verbai emissïons. There are
so many things we want to tell ihe..warld. Same öf them are important, some of them are not. But we taik anyway ven when we
know that whai we are saying is totally unïmportant. We love chi1ch»i md red silent encounters
*wkward, or even öppressive. A life without. wortls wouid be a horrendous privation.
II ïs a cliché to say that werds and language are probably humankia4’s most valuable single jx›zsession. It is language thai seis us apart
frum our biologically close relalives, the greai primales. {1 ivould imägine that many a chïmp or gorilla would give an arm and a leg for a
few wordsMut- we will pmbably never Snow because:.they cannot tell us.) Yet, surprisinply, most of. us take words tand. mo're
generally.IanguageJ for granied. We cannoi discuss words with anything like the .comprtence with which we can discuss fashion.
filnxs or football.
We should not. take words for granled. They. are. too imporuuii. This bmk is intended to make enplicii some of the things hut we
know subconsciously about words. It ïs z linguistic introduction to the nature and structure of English words. It ad‹lresses the question
‘whai sons of things rio pe'ople need to know about English words in order to use them in speech?’ It Lh intended io ïnciease the
üepree of sophisiication with which you ihink abaut wonis. It is .desibned to give. you a theoreticai gmsp of English word-forinaiion,
the saurces of English vocabülary and the way in which we slore.and reirieve words fmm the mind.
I hope a desirable side effect of working through En$msh Words will be the ewichment of your vocabulary. This bmk will help to
inerease, in a very pmcticai way, your awareness of the relationship between words. You.will be equippe‹l wiïÏt.the tools you. neel to
work out the meonings of unfarailiar words and ïo see.in. a new light the iinderlyinj struciwäi patterns in many familiar words which you
have not previously stopped io mink about analytically.
Fer the student of language, wordi are a very rewarding object of siudy. An.understanding of the nature of wonls provides us with a
key thst opens the door to an uixlersisn‹lin of important aspects of the nature of language in general. Wonls give us a panoramic view
of the entire field of linguisiics because they impin e on every aspect of language structure. This bmk scesses the ramifications of ihe
fact thai wonls are complex and niultï-fuceted enüties whose stnycluie and use inleracls with the other m‹xlules of the giammar

2 ENGUiSH WORDS

such as PHONOLOGY, the suidy or how rounds are used lo represent words..in speech, SYNTAK, the study of sentence structure, and
SEMANTICS, the study of meaning in language.
In ortler to use. even .a very simple w such .as Jug, we need to access various types of information from the word-store which we
all carry around with us in the MENTAL LEXICON or DICTIONARY thal is niPled away in the mind We need to know:

iü shapc, i.e. its PHONOLOGICAL REPRESENTATION/frgt wbicb enables us ip proiiouncé ir:bud.its ORTHOCrRAPHIC
REPRESENTACION/rio, if we 'gre Jitwaie and know how io s'pell it ísee the Key to Symbols usod on pags.xix):.
(ii) its grérrimaticaJ propcrues, e.g. it Is a n¢xzn aod it couatab! so you cazi have one/rug aod twg/rugs.
its meaning.

£iut wñrds tend not io.wear their meaning on their sleeve. Normally„there is nothing aboui. the,form of woids that would ennble
anyorie lo work out iheir meaning. Thus, the fact that/rep iefers io one of these simply has tu be lLsted in ihr lenicón and commilted to
memory by brute fdrce..Fór the relaiionship bèlween. a LINGUIÉTIC SIGN lïke thïs w'ord and' its iiv•nning is ARBITRARY. Oiher
langunges use differenl wor4s to refer ló this sniaiI.tailless.amphibian. In.French it is calied f!•? r •":»ui//e. In Malày they call it iimaà
and in Sw'ahili llum. None of these words is more suited than ihe others to the job of referring lo this smàll repule.
And of couise, within a panicular. language. any panicular pmniinciation can be assocïaled wïth any meaning. So long. as speakers
accept thai sound-ineaaing .association. .they hàve a kosher worcL Por ‘inslanoe, cnrwenience originalJy me•ani. ‘suitabiliiy’ or
‘commodiousness’ but in the middle of the.nine1eenih century a new meaning of ‘toilel’ was assigned id ïi nnd people began to talk of ‘a.
public convenience’. In the.early I96Os the word acquïied the:addïtional new mè•anine of ‘easy to use. fiesipned for hassle-free use’

We are the masters..Wórds are our servants. We can.make them mean whatever we want them io mesn Humpty Oumpiy had all
this worked out. The only thing missín from his analysis is the smiai dimension. Any .arbitrary meaning znsigned to a. word needs. to
b•e. acoepleó by the speech community which uses the language. Obviously, language would noi be much use es a means of
communícation if each individual language user assigned a pñvate mesnini: lo each word which olher users of ihe lenguaje did noi
recognise.. Apart frnm that, iL is insirtictive to listen in on ihe lesson on the nauire of language ihai Humpty Dumpiy gave to Alice
(see overleaf}.
Let us now consider one funher example. All competent speakern of En lish know 1hn1 you can add .-s to a noun to indicara thai it
refers to more thnn onu entiiy. Sa, you sai ent when referring io oye and cms if there is more than one. If.yi›u encouniered in the blank
in [1.ej an »r,›n:i:.;‹ ú»o:i:ke xpiei:(which I have: just made up), you weuId *utomaticaliy know frnm the content thärii must have the
pluiai form sp/gïs in this position iince il is specified as plural by ch. Further, you would know that flic plurai of splel muse be:splèts
trather than splrtreii by. analogy to ‹/iiïdren or spleti by analogy io stiniulit. You knnw that ike majorité of nouns form. their pliirai by
adding the regular pIuraI iuffix or ending -s: You aiways add -s unless express instructions wc given lo do otheru'ise. There is no need
to memoñse separately the plural form of most nöuns. AII we neœl is io know the rule that says ‘add -s for plural’. So, without any
besitaiïon, you suffit -s lo obtain the p[uml form sp/e/s in [1.2b]:

We put all Irte bi oo the table.

'We put all the big spk'zz on fhe table.

The - study of w' rd-formation and word-stricture is. called MORPHOLOi Y: Morphological theory pm.vides a general theory of
word-structure ifi all the languages of the w.orld. Its task is to channels rise.the. kinds of things that speakers. need to know aboui the
structure of the w'ords of their language in order to he ablé to use Chem to produce ad lo understaad speecb.
We wiU see that in order to use language, speakers need to have two types of morphological knowledge.. First, they need to be.able to
analyse eiisting words (e.g.. they.. muri. be ablè tó tell tha1Jo@s co'ntains.frog plus -s fnr plurai). Usually, if we knnw the meanings of the.
elements thai a word contnins; it is possible to deiermine.the:seanìng of the entire w'nrd once' we hnve worked.ont how the various
eIements.relaie to each other: Fór irlstance, if we. exemine a word like nutcróc'ke'r we find.that it is made, up of twó worils, nèmely the.
noun cui:and. the' nóiin crnc£er: F.unliermoie, we see that the laiter xv.ord, rrnJlir is divisible. into the. verb cra'ck and anotiser. meaningful
eiement gr (ron hly meanin@. ‘an insirumènt used to do X’ J, whicli, how'ever, is net. a word iii its own righi. Numerous other words are
formed using this paitern of comhiriin Words (and Smaller.meaningful eleo›ents).as seen in [ I:3}::

£iiven the frame [[ )Noun—jer]] Noun, we can fill in different. words wiiii the appmpñate. pmpertíes and gel anodnr cnmpound
wnrd ii.e. a wood containing at least two words). Try this frame. out yourself. Find two more. sim'1*: exámples of cnmpound. words forméd
using this patiern.
Second, speakers need to be able.io work oui the meanings of novel words constnicied usi'ng the cord-
building elements and standard úórd stniction rules df the Language. Pmbably. we all knox and use mode wnrds than are listed in
diétionaries. We can cnnstruct and analyse the structure and meaníng of old' words as well as. ned unes. Se, although many words musi be
lísted in the dictionary and memnrised, listing every word in the dicñonary is nnt necessary. If a word is formed followíng general
principios, ii may .he mode effieienl to reconsiimt :1 from its constítuent eÍements .as'the need aríses rather than permanently cnmmit il lo
memory. When people make up new wórds using exisiinp words and ordforming elemenis, we ttaderstaod.rhenz.wittt e viding
we'knów wkat..the elements they use to €onn.rftose wonls meao and providing the word-fòrazlag rtiles that tÏtey employ are
faoúliar.:TltJs.abi1ity is' aae uf flze ttziogs exptoted in rnorphological investigaiions.
In an average week.. we. are likely to encoiinter...a. couple of wnf»m lisr words. We in ehi reach. for a diétionary and look them up.
Some of them muy be Esied but others might be two new ortiio ephemeial to have found theïr way into any dictionary., In such art event, we
rely on our rnorphological knowledge. tó teaïe òut their mean les. If you heard sna›eone describe their partner as ‘à great lisi aisker
and a fickermfF,.ymi would instantly know’ whal. sort..of person: ihe partncr w though you almòsi certainly bave never eticountered the
xvord fic£rr- befote: And ii is certainly not lisled in any dictionary. The -er ending here has

4. ENGMSH BORDS

the.meaning of ”someone wko dces. whatever. the verb.meaas*.,Given the verb ricAo@ a ticAe -o@muvt be a’. person who.links n@
Similarly, if you know wkat established words like liandfñt, cup I and'sje ez¿/’7r/ mean. you are.4lso able to figure out the meaaiogs
of.novel worrh.liké/auo/din-/zei/i / (as in a/ozoi/oin- n / o/ irut) or Jini'ercrnjfr;'id {as in .Jroverrrnj )' I nder fiovzrcro / nf fizgfisfi
shoppers returned from Calaic feeded dome u init ‹-igare;tes, cheese end piety). Virtually 'any .noun .denoting a container can hnve -
ji‹/ added to it in order to indicaie thai it is ‘full .of something'.
To take another example, a number of wurds ending in -isr, many of which have. crime into use in recent.
years; Geler iö people who discrùoioare a@ainsr, or hold negative views about,. certain less powërfui subgrnops in söcieiy, e.g.
mères, se.v*si. Anynne who knows what roc*sr md serisr mean, siveù the right content should have no difficulté in understanding the
naiure ñf. discrimination perpetrate4 by peaple who tre. described using the novel words ngeisi, siÿsï and sprecfiîsr. Ageixni ü
discrimination cin groiinds of toldJ age fur instance, denying employo›ent to peuple över ihe age of. 60: sic‹sm is discriminaiion
(usually against
far people) on grnunds of size and sprrr/iisni is discrîminafion against people with speech impediments like
stuttèring.
Did ynu notice how I explnited your tacit know led e of the fact thai words ending in -isi -and -isis cnmplernent eacb other? You
were. glad io accept..agei ni. si;-isiii and sptechiriii because. jon know ibat: cnrresponding lo an adjective ending in -îsr. there will
normally be a norm ending. in -isct. This is important.. lt shows thàt .ynu .know thai certain w,ord-forming bits go. together—and
others:du noi. I suspect that yöu would reject pulative woods like. *meiiieur, *si/eznrnt and *ipreclinrrni. tAn.asterisk is used
convenlionally in indicaie iliai a form ii disatiowed.J In w'ord-formation it is not a case of anything gœs.
A challenging quëstion which morphology addresxes is, 'how' du speakers know which non-orcurring nr non-establisfîed words are
permissible anñ which once are not?’ Why are the woods Jounrainyn@/, IÏoi'rrcrn f and xprecùîsr allÖwed wbile *zz;grmrzrf,
*sp'recIiinrzir and *cix-e'nien't are:no4?
Morphologicai theory pruvides a general iheory. of wordformation applicable to any. language but, as mentioneil earlier, this book
fociines ön. word-formation in English. Its öbjective is to prnvide a description of English wnrils. designed io make explicit the cartons
things speakers know, atbefi in an unconscious manner, about English words. The ëmphasis nil be. ön the descrijuion ñf English
wniös rather than thê. elaborañon of mœphological theory. Éo, data and facts ahout English woids are brought to the fore and the
theoretical and roethodological issuds are kept îa Iîte backgroiutd Par rhe rnost pazt. The uce. où NonnaJ notation has alsn been kepi in
a minimum in order to. keep the account simple.

OVERVfEW OF COMtNG CHAPTKRS

At the very outset we need to establish the nature of the subject we :are going to be examining. So,. Chapter 3 discusses. the
nature of words. Then the next thrée:chapters delve deep inside. ..and 'i*vextignte their internal
structure. In the pmcess, traditional morphological concepts: of structural linguistics air.intnx1uced and extensively exemplified.
Morphology ix not a stand-alone module. After-the introductory chapters, in Chapter 6 you are introduced to a thenry where
morphology is an. integral part of the LEXICON or DICTIONARY. This chapter focosei on the interaction of phonology and morphology
in word-formation.
Chapter 7 explores the relationship between Fords'in speech and in writing.. Wha1 is ihe relafionship
betw n sâying words..and wriñnp them down? 1s writing simply i mirror of speech—and .an apparently distorting one-in the case of
English?
 The following chnpter émntinues the disciission of the role of the lexicoe. II ;tnempts to answtr quesiions: like ‘whai. is ihe festem for?’
‘Whai itenis need to be lisled iri the dictionary?’ ‘What is. the difference between idioms Mike io not/ inne 's cofours re iAr user) und
xyritactic phrases t like in void o noiice re ihr door}?’ The next two chapters hishlight the faci that the English word-siore is vast and
in1”initely expandable. First, in Chapier 9 we cnnsider the ways in which, using the internal resnurces of the langoage, speakers are able
tö pruduce an indefinitely large numbet of werds. In Cha'pier IO aiiention sbifts to the eip;insiön of English vnabulary thmugh the
importution of counlless words irom other languages. The story of importefi wnrds is.'in eiany ways also ihe itovy of ibe cc iacis M
spcakers öf English: have hart with speakers of oiher languagex os'er the ceniuries.
Mosi of the space in ihre hook is devoted to an e amiiation of ihe siniciure of Esg[ish w'ords. But the
anaiJsis of word-suuciure is.seen not és an end in itielf, but ratlser és a mens'.io an. end. And lfm end is io iirukrstand wlut it means in 1-
now' a Word..Whai sorts of information itbour u'örds do ymi neml. to have in order io use ihein in commiinitaüon?. So the final chapler
is devoie4 io the MENTAL LEXICON. lt Asses ihe quesiio», ‘hnw is i1 that penple. are able io more a vasi number nf wnr‹ls in ihe
wind and io retrieve the right one xo fast in communication?’ We will see thai wools are not piled in a muiltlle in the. mind. Rather, the
mental lexicon is very highiy nrganisal. This conclinling chapter will msg pull together the various strnrnls developed in ihe earlier
chapters.
I have already. stressed ihe poor ihai murphoiogy. is not a- selfcontained module of language. Any discussion of word-fmmaiion
ranches on order areas of linguiiiics, notably phonology and syniax, en I linve provided a key to die list or pronunciation symbols at the
beginning of the book..I nave also included at the: end a glossary of lingiiisic terms tuiany of them from nther htanches of
linguistiCs) which might be unfnmiliar. Bet siiII I muy have missed oui some terme:. If you encouoter any unfiamiliar iechnical tenns that
are not explained in .this book, I suggest thai ymi cnnsult a good dictionary of linguisiics like Crystal t 1991s. Snmetimes it is
iisefui ie preseni data usin!g. phonetic notation. A key io the phonetic. symbok used

Afier lhis introduciory cbapter, all ehnpters cnniain..exetc es:. Several of ihe yiicN exercisis. mcguire.. you to ir up wnrds and paris
oí words in a geod diciionary like the 0.lord Engtish Dicti-as: Access to such a: dictionary is essential when you .study this book. This is
a practical way of learning a1›out .the stricture of English words (and may.also be a useful way of enriching.your vocabuIaryJ.

Chapter 2 What is a word?

ii
INFRODUCWON
Often we find it very difficult to give a clear and systematic couat of everyday things, i‹kas,. utters aid events that swmund m. We
just take them for grmtnl. We rarely need to state ie an curaie and articulait manner what they are realty tile. Fur instance. we all
knnw what a game is. Yet, as the philosopher Wiïtgenstein xkowed, we:finJ it very difficuli tre state expiicitly wkat.the simple woc‹i
eomc meaor.
Tlit same i-s true.or the term n'ord..We: use words all the time. We iniaiiively know whai the won› in ooi tmguage are. Ntvertbetesx
most of us would be haul pushed to explain tn anyone wbai kind of object a Print is. If a couple of Martian explorers f with a
rudiaieniaiy ur›dm.tandieg .of Eaglisbi come' off ihtir space-*hip and stopped you in the smart to erujuire what earililirips me: t by the
term WORD. whai would
.Too tell them? I suspect ynu might be 'some*hai vigue and evasive. Although .you know s'eiy melt what words m, yoo might find ii
difficult in express explici6y md saccincily whai it is that you know ai›out them.
The put of this chupier is to try to find on mswer to the qotstiñn: what is a work? tt is not only Martim explorers envious about
the way earthliegs live who miphi wani io kanw what wuNs at. We ioo hu\'t till i8teFest ifl tttttlentandiRg Wilds becaose tlley
play *Och an iMQOrtdftt rifle. in OOt li'ves. AS We.saW ifl the lasi chapter, ii is inn ible in inugine boman socitiy. wiiboat
laopingt. And equally, it is inn sible tn imagine a human language that has no wnrds nf any kind. It is impos:sible io ua‹lerstand ihe
aatuñ of language.without gluing snme uadeMiandiag of the naiure of words. So, in thit chapter we will clarify whai we •»• chen we
use ihe term '«'orJ". This ctarification is.esscntial if our investigatiunr.are zo ''^f ' amy headway fur, m you will for presently, we mean
qaite a few very ‹liffereot thiwgx wherrwe tnM of words.
A itundwd defiaiiioa of the word is foond in * paper written in 192fi by thi.American liagoL4t LeonnnJ Bt ernfietd, one of ttie
greatest lingoi›ts of the twentieth cz*tury.. Accosting to Bloomfield, ‘a •n »m»•n. free fcwi is.a word”. By ihis A r«eo«t thxt hoc
ww‹i ix the s«icltest r: • gf«t lixguisüc unit tha c« be used..un iis öw n. lt is a farm that ca«oot be dit'iJ«d into any sma\Ier unics th«t cac
be used indepc»dcntly io convey meming. For example child is a wofti. We ewinut divide ii up into .smelter onits that vm convey
mewing when they stand *lose.
Contrast ihix smith the word cliildi.•h which cm be analysed into child- Grid -i.i6. While the ‹laird bit of
‹/iiidii6 is meaningful when and on its own tand henct Ls a wnrdi, the same. is not iiue of -ink. Although arcurdiag io the D fnrd
Eiiylisli Di‹-ii‹uiar; i LED -i.•h means something like ’having the tnhjcctinnablei quaiitie4 ofi tas in ir«iuiisJi, voirrari*fi, devilish.
Repisii, «pivii etc.), there is rio way we can uxe it on its.

o•'n. If mms shouted tn you in the street, Hey, «re you -i.‹hh’ you might smile hcmas¢dly and think to yourself. 'Isn"t lie wcirJ!" In ihc o xt
chapter we will tcLc up the question tel what t‹› Ju with piccer ‹›f wands that caxnot I+e u'•ed meaningfully on their own. But for lkc
moment we .will hus exclusively on

HORDS GRE LtKE.LIQUOR2CE ALLSORTS

When we tulk of words we tJo nol alwuys meun exactly the samu thing. Like liquoricc allxort , wurds cumc in all sorts of 'varieties. We will Platt
our dLs‹•›••›‹• • by dLstinguirhing the difftrent •••• in which we uxe the tcmi *wurd".

12.t

Let us. use 1e temi WORD-FORM to describe lhc physicat form which realize or represents a w‹›rd in speech cr writing. C'onsider ihe wclxl-•
in it ft›l1owing cxtmci fmm T.S.Eliol’'• pm:

In writtcn E«glisfi, wonls xrs hey to rccugoi . Th«y an' peceJc‹J I y « spscc. wJ futloweJ by a spxct. Using thir criterion, we carl say that
thcr2 afc thirty-one wnrdx fi.e. wnrd-farrrLs) in the cxtract fmm Rhays0dy’. We will call wonJ-funn liIu• the e which we find in wrifing
ORTHOGRAPHIC WORDS. ff yuu look again »l the cxtia«t, you might wond¢r iI sonte oI ihc hyyhcnateJ u+thogr«phi* 'onJs arc 'rc«t1y"
inJividual wunJs. M»ny pcople wr*uld hypkenaic lmlf-poci a» Eliot Jocs but xot .rfreer-/o›ny›. They woulJ write sfre‹'r lamfi *x t'wo scparafc wonJr,
with o spucč bctw¢en tkem. What would you do?
The u'•e uf hyphcns to iriclicate that xomcthing it• a cwmplex w‹›l”tJ containing mure thux ‹me wt›nJ-like unit ir variahlc, I4rgely dcperlrling on
huw trunxparcnt ihe cumpound nukirc of a wtmJ ix. Skakcspcw wmte f‹x/cty es ro-dol' icid fzwrnnz›u* as ro-n/oriY›u*

a TO-mOFrOw. CMsiir.
I shaI1 bc fumishcd ln inform you rigft Iy...
tAnf‹ws” rin‹4 č‘/er› fru I, iv)

e WHAT IS A Wt›Rtr!
 Hyphcnsiing. f»-‹nay and in-nN›rroit' is less.cummun now, probably hetousc most speakers are unaware of the cumpounJ nature of thcxe
wurds. 7A‹n' cumss from Old English 7 dv *It›+day“ axJ r‹worn›o is from Middle English ii› ni rtil» é Ji.e. to (thcl morrowl —fn- em br
traced back ultimately to a firm thai meant
*this“ in Ind‹›-European. Nose in passing that tkrcc major psriudr are distinguished in lhe history of th« E«glixh la«¿uegc: Old Enjjli h
(c‹›nventiooally abh«iaic‹i a‹ OE› wax spoken ‹ .45U- I USL klidJlc English
‹conventionally abbreviated is MED w*s spoken r. I 100-1500 NJ Mo‹lerx English from lñfX) to the

Gsncnslly, the use of the hyphen in such wuW that arc nn longer '›ccit us cumpnunA is in ‹incline. The hyphen teruls in be mostly used in
eompuuntfs tlxit arc regarded as fairly acw @xxLs. iYlmy well-esitiblished wools that are transparently corxpnunJcd, c.g. s‹°hu‹›lb‹›)', are
normally written without a hyphen. Of ccurxc, judgements. as lo whai ix art csiahlixhed w‘urd ozy greatly. There are few firm rules hers.
For instance, in the f!Eñ. fx›th ecix*\' and sru-u!n\' are rhuwn tn kc accepted »'uys t›f writing 1c wunl pn›no+uiced as / si:well. Similarly,
the compiler'. of d+c ITEM shove’ vuriasion in l?ie way they enter bold hyphenuted/ira7-nife ancearm rare written as two wa4« 'separated
by a spau2.
Interestingly. hyphcrlation is ulso ucrd creulively \o indicate that un i‹lea thut woulJ normally he expre+sed fry u phn is keing tzeatcJ
as a ringlc word fur con»nunicasive pui}xz'x•s because it has crystallised in the writer's mind into a firm. single concept. Tiius, for
example, the expression .simple to srn'e is normally a phraw, just like rosy rat ‹ ›nwoi. But it can Jso be used zs a hyphcnatcd wiwd
in xinpie-rri-.ien'e rr‹ipe diriies iMd S ñfnpo.tire l9P2:9). Similarly, on pugt 48 uf the same magar.rue, the writer uf sn advertising feature
nos the phrase frc*h from the farm’ as a hypheriated word in ’licxh-friwn- the-hurm eggs'. But for creative hypticnatirm ynu arc unlikely
to !inJ anything mnre striking shun thix:

2.3|

On Phcaim thrre’ is lilllc c'ri¢lcncc uf ¢hc n /vif-no-7/u\<-n rJir*7c/: /m-sr‹rzz’in4rr, IAc tJrirrn”s-pi‹”fvzz'-ill-ri r¿”-mm uin of attitude.

Wkat xc huvc cxtuhlished is that as a rule, twthoyrayhic •'or‹b huvt a puce un either side uf Diem. Bul there are casct where tliir simylc
rule uf thumb is nN tolLwed. There is s deprcc of flexihilily in the way in which wurds use written doo'x being. or not hcing,
scparatuJ by a space is in itself not u '•ure sign of word status. Some orthogmphic w‹›+cLs which are uncontcovcr'.iully written as nnc unit
cunutin twn wonJs wiittin them. They arc cnmpouxd wor‹L+ like air.smells, aeuu'c/t. »'/tecf5rirr‹›x' axd 7eu/+r›f. Furthermore. thcru arc
forms like 7Jtrt''n’, /zuc/zt'f und /“m which url joined together in writing ytl vhich.ure nr*t conspuunJ wor4s. When yc›u scratch the skin.
you see immediately ih«t r5ej*' r, halo"t.and /'m are really versions of the pairs of wnrdx 7/rci' cn¥', had' fh'+/ ml / ofx. Our thuuty needs
tn .say xumething stout awkward caslomcrs like thesr. Since II+c issues lhey misc are curxplex, we will pnrlpnnc discussion of lhcm until
sections {4.3) and
{S.1›. Finally. them 4rc worJs which arc eump‹›unJcJ (und mxybc hypkenatcd as in [2.ñ]) of » oric«›ff t¢* cryrtal1i'•c a particular meaning.

NSGUSHWOADS.9

So far we have only considered orthographic wordn, i.e. recogriisable physical wrltlen. word-forms. Ob'viöusly, ,üords .as.physical objëcts
mist not oniv in writing, bür aiso in speech. We will:noxv bri'efiy tuin tn 'word-forms in ipoken language. We will. refer to them as
PHONÖkO€iICAL WORD8.
The chailenge of word recogniti'on arises 'Ïn an even more obvious way when we coniider speech. Words
are not separatèd distirictly from lach other. We do not leave' a pause betwëën: wprds .that éoidd be.equated tö. a space iii.writin . (If we did
that,. conversation:wouid be.painfu1lJ./slow.! lust Dy' ipeaking. to orre öf your friends today leasing a two-secnnd gap berween words. See how
they react'.) ln normal speech words come out in. a torrt•nt. They overlap. Just .as droplets of wate'r.caiuiat be seen flowin{i down a river;
individual woids do not Btand nu1 di'screte1 :in'the flow' of üoiiversation. Se they are z uch.hardeI to isolate than words i'n writing. Norm the less,
we are:able tu isolale diem. If you heard an urteran :iike:

cket skpt in bod/

fNoic: “'. show's iJiat ibc falJow n syllxblc ›s:soessed. phononic uaru«1ptüian wnttco betwccn s)ant I nos.)
 yóu xvould .be able io reõogyiie the six phonõlogical 'iiords: that have been writien. iii PHOHEMIC TRANSCRIPTION (which shows Lhe
PHONEMES, i:e. the snunds . that: are used to distingñisli the meanings of wqrds) aithough what.. you heär iS oke cogtinuoui ineam .of sound.
For. purëly' practical reasonsi :Lhroughoui the boot unless otherw.ise. stated, phoiiemic tianscriptioiis and refe'rences.. tö pronunciation will
be ba.uxl on RECElVED PRONUNCfATIÖN {RP), the prestige . accent of standard Briösh English variety'.popularlJ'.known as the Queen’S
English or B 8C English.
An intriguing qtïeition Lhat linguists and psychologisis..ha*'e lrièd to. änswer'ii: hnw do peoplg recognige words in speech* We will address
this questinn in detail in section ( 11.2.1.) bclpw.:' For now let iïs simply assut e ihat phonnlngiéal words can be identified. Our prexent task
Wifi. simply be to outlinú some of their key propcrlies. Tõ do: this ii. uiill. be. useful tó diitinguish beíuigen twõ types õf: words• the smcalled
COMr£NT WORDS and F'Ut4fiTlON WORDS. Content words are toe. n‹iuiu, * , adjecdvei and ada.ei-hx which .éontain mÖst' of the
REFERENTIAL. (or COCiNlTlV.E MEANIISG) .of ä ientenüg. This roughly rn'e:inx. thai: they iume individuals and predicate: öf them certain
pmperties. They. full us; for instance, what happened. or whn did whnt to whom, and in :wha1 'circurxistañces. An example will make the poini
clcar. In. the oÍd :daJS, 'when piople sent telegrama, it was. cõntént' words that wei mainÍy (or exclusively) used. A proud parent could send
a° message. like ffobv gif ar.rived xesierdox. which conuñneA lc'o'. rrouns, a verb -and an adverb. Obyinusly. this is not a viell-formed,
grammatical si•iiterrce: But its.' meaning would be clear enough. .

Function words are. the rest—prepositions, pronouns, conjunctions, articles' and so..or They have: a predõminanily grammatic'a1 role. A telegram

containing only dre üords She'ii and fôr ns 'iioüld üonvey littlê idea pt whit the intended interpretationwas. .This is not .to say' that fuiic;tion Words are

fiupcrfluñux'. Withnut them. sentences are:usually ungrammatical. A sentence: like •Nellv 'cmi.town which lucks the preposition Io is rint perrnitted. Wú

have lo say Nélh’ werir'ro rou'*. .

In Eni;1iih, content woids haS'é thii properiy’. ope '.of theíi sJliables.; ii rpiire piomir it than the .rest because 'it receives MAIN 8TRESS. This is
seen in the words below where the syllable with main stress is precmled by '” :

Iu WHAT JS A WUkD?

Main stress can l'alt on only one syllable in a word. The localinn of main' stress is part of the make-up of a Word nnd i'x not Changed éapriciousli by
individual speakeifi.. You carmot deeide tu *trets Jiii/ïo/ofr›ó on tfie penulúmate syÏlabÏe on a Moriday ¡huRa”balnn), on the.
antepenultimate' éyllable on .a Tuesday (/rr‹ “f/o6n/ooj, on the initial syllable on. a Wedr›isday (’hu/fob‹i/n and on the final syllable for the
re*t of the week lliiill'obo‘loo}. ”
Howevor, in sorra cases, i£ we wish to cootmst two. related words, we can rhift stress from its normal positius tela new position. This can' be
scen in ’vynJur and nen !dee which notm'aily are stre.ssed un the:firsl and sevond sillable iespectiveli. But il’ the speaker wanis to contrast
thcse:two words both words mipht b< stresad nn Lhe final syÏlable as I hecrd an estate agent do in a radio interview.

It is v'r i ‘di›r, not Ihc t'cri ‘Amr' who pays that ixx.

This example- illustrates well the point that a word is allowed ju.st one sLress. Stress ccn be shifted from one' syllable .to, another, but a word
cannot have two"main stresses.. We could not. bave ”‘ven“dor and
*'ve'n'dae wfiere the ttvo syllables received equal soess. Stress has th do with relative prominence. syllable that m¢eives main strecs is so
hat mare prominent than the' rest, some of which mcy be unstressed or weakly stressed.
By'cÓntrast, functioo wórds ar normally unstrexsed. We can say Nell+' u siri re roun with no slJess on In unless we wish to highlight: ro For
contmstive purpasos, e.g. /\ie//\ ii errr ro ro+i'n arrd nnt}âr riwu\”/?wirr roost).
It ix easy.to. see hów stre.ss can funetion:esa valuable clue in determining wliethe two content worcLs are
a single compound word of two separate words.. 3Tte nouns street and longs are both stressed when they
occur in isolation. Bill if they appcar in lhe compriund ’sIreet•laiiip, only the first i.s stressed. The stress on
SLresr is not the only phunologicai cJue. In addition to. stress‹ there are lules reguÏatinp fhe positions in which .various sounds may' occar in a:
word and the c-onibinatioas of sounds that are permissible. Three niles are cailed PHONOTAC.TIC RULES. Th!e y can help nx tö know wilether
we are. al the. begiritiing. in the middle or at the end of a word. A phoriological word musl satisfy the requirements for words of the spoken
 <s••sg. F.or instance, while any vowel can begin: a w.oid, and most consnnan can appe ar alone at the lx'ginning oFa Worrl. the consonant ( )
ir Subject to ceftain restrictions. (This consonant i.h s@lled ni as in fi‹znfi tsee lhe Xey to.símbols used on p. xix). In Únglish words | ] is:nut aJloweó
to occur initialJy altïinugh
it”can ociur in other positiÓns. Thus. | ] is allowed internaily and al .íhe en'd of a'word as 'in (I I ] lo'iigriu
anó (I ge) los yer. But.you couid not frase an Eitglish word Life. ityrtlier, ’| eej with ( ) as its.first sound.
However, in other langziages this waund may be I°ound word-initialÏy as in the Chinese name Ng• I •I
the Zimbabwean name Nkorno [ iiomo].
Tï+Are are also pftonotactic restrictiorrs on the combination of soósonauts. in xaúous po«ióons in a wad in the spoken language. A* everyone l
¢uo\ve, lglish npeÏliny is not always a perfect minor of pronunciation: So when considczing wozcls in the ;poken laat,'uage it is important.to
separats: spciling from pronunciation (çf. .Gapter 7t. Yom inow that the is inici- irrd is prono need 'hI iz nk ni:df and. not *fhe 1s M hi:df. A
pazticuÏar combination of Letters ccn be ass¢ciated with very different pronuitciaúorts in different wards or

t2 W£IAT IS A h'GRO?

hyphenaied word-form sfreef-iniiip veure three times. So it‘ we were úouriiing different word-forms; we would count rrezf-lomp three
limes. Howeveii if we were covriting diuirict wordx: in the sensg öf distinri VOCABULARY lTEldS 'we hemd only.c.ouni ii once
The. distinction between word-formx and vocabulary items is important. Very oftêri, fi'iien v e iaik af›oui woids what' we have in
mind ix not word-fnrtns, but iometbing more abxtmct—whai w'e jill refer lo here o LEXliMES li.e. vocabulary. ifims). .Anyone
.conipilirig a diciionary lixis worös in this sense. So, altbough ihe word-f*r na m of ihe colamns in {2.ä] below are
dÏtfereni, we do not find each one of them given a separate enuy iw an English dictioriary, The fir t w'ord in each column is listed under a
leading of iix owrt. Thu rext may. be inentioned under that binding, if iney do not töllöw a regular paifem öf the länguage .g.
•'rite, w'riilen tpasi parüciple1, «'role tpast tense). tïut if ibey do follow the general patiem te.g. on/res.
»'«s/iirg, u'múed; smiie, iiiii/irip.. imife‹fI ihey will be lela out of the dictioiary tdiogethcr. Liitea4, the gramruar will be .expccted to
prnvide S general staiemeni to the. effect tfiai verbs take an -riu suffi». whiCh marks prngressive aspect, and an -rd xuffix that marks
both ric pau iense and the past partíc ple, and eo on.

büngs

lty {2:d] each lexeme li.e. vocabuttu-y. iteinl thii 'wduld be entered in a dicii'onary ii ahown in t;¡tpital ferrets and ail the diftémru
vi'ord-formx belonging.to it are ahowu in lower-case leiterx.
The example.s in 2.8J Are ail verb.s. 8ui. of course, leierues can be noons, adjectives cir adverbs as well. In {2.9] you will find exaciplex
frnin three other word clasxex.

}2.9.]

VATCH

b. GOOSG

KIR O

kinócr
SOON

WELL

bcucr

In {2.9] ñe have three jiaiis ot‘ lesemes: the nouns,. mar ?i and giiniJ: the .adjeciivex tiiid and bad: and adver6x .rout and w'rll. ln. each
case the word-form.s. belonging to each lexeme in {2.9a] follow a general patiem for u'ords óf. ilieir tipe and necd not be liswd .in the
dictionwy: Bui ail..the oúes in (2-96] are irregular and musi be lisied ie the dictinnary,
The lexeme is am abstract entiiy thai is found ic the dictionary and that has. a certairi meaning. Word- förms are the cöncrete öbjects
that we pin dnwn on papir lorihographic words1 or mier (phonological Gordel w err wi use language.. The relationship between a
lexeme and the word-forms öelr'ngirig ip ii ix orde

of REALISATION or REPRESEQUATION.or MAl'fIFESTAT1ON. If u'e take the lexeme irri/é uihich is entered in. the dictionary, for i•xamp1e,
we é% see' that it rn*y be realised by any oñe of the' wo'ord-forms. vt'riir, writes, wriii'ng, wrore and entice whiéh belong to it. .These are the
actual forms that - are .used in speech or appear ian piper. When you see the qrtbographic wnrds u'r(frézf and wrorñ on die pyge, von know that
althiaugh they are spelt'differéntly they an manifg'siationi of the same *'pcabulari' iterñ WRITE.
The distinction between word-forms and lexemes which 1 have just made. is not absmise. It is a distinction that 'we .ire .inniiÓvely avia;e

.of from afi .early' abe..lt is the disúnction on z4iich word-play in puns. and in intentó al ambiguity in. evéryday life de'pemls. .Át-a certairi periód
in our childhood ive were fascinated by words. We loved joke ven awful ones like [2.101
The. burnour, of cnurse, lies iii récognising Chat the word-form chriNp can bclong to twñ úeparaie'1qyemes
whose very diffeient and iinrelated rriean'ings are óone th'e less pertinnnt her'e. lt can.mean éither .“En edible, long, slerrder crustacean* or “a tiny
person* (in colloquiai Fuiglish). Aiso, the word seii'e hns two possible interpretatioris. lt can rp'ean “lo wait Upon. a persan:al:table* ór “to. dish
pp f‹xid’. Thus,.wnrd-play ezploitf the 1e'xíca1 ambigfiity arising. from. the lact that the. same word-form.represents twn distinct lexenies with
very disónct meanings.
In real-life cop;;nuoicaóon...w.here potential ambiguiiy. oé urs we.. generally rpanye tn copie to just one interpretation without too:.;nuch
difficulty by selgcting the most appropriale and RELEVA•4T interpretation in the situation. Suppose a 2Ó-stone super heavyweight bozer wéril to
lot'.s Vegetariap Restaurant aqd asked the waiter fpr a niée. shrimp çurry.'Unid die waítér iaid in reply, 'We don*t serye shrimps', it ivould be
obvious that it was.shrimps. in the serisg of crñstaceans that w.as-inlended. If; on the other. hand, a .littie ;nan, barely 5 feet tai1.aiid weighing a
mete 7.stone, went io a fish ré'staurant and saw almosl everyone at the tables around him tucking into” a plateful of sue;:ulgnt shrimbúi apd
thoughi that hg would quite fa cy some hirnself, he would be rightiy .offended if the wailer said “We do not serve shrimps.’ lt is. obvious in this
.situation. that shrimps a;e”on tÍie. menu and are dished up for consumption. What is not done is serve up food to peóple

Puns are not restricted to joker J\4any advertisements like: that fór.Siandens rely.on puns for tlieir. effect. Given the content, it i's obvious thai
spiyiid is meant lo be'read in more .than one sçnse herí:.
Serious literatura'. also uses this .device.. FoK in5lance, .the First W,orld War jxiet. siegfried Sassoon: gives
the titre “Base details’ to the poem in which he pamdies cowardly generals who stay away at the base, at a safe distance from the acdon', ,an'd
gladly speed young soldiers to their death at. the frnnt'. The word-form Eye in die' title.represerits two. distinct lexemes here whose 'meanings
ye both:relevant: (i) Biuié &•iails e details. nf what. ix happening at. the hues ,p g, (meaning. “military encampment' ), and (ii} Bace details are
particulais of 'semething:that ii f óre ,yj „, {.meaning 'rt:prehentibly cowardiy,.mean etc.' )..
The term HÓMÓN YJVí is used to derrotó word-forms bélonging to .distinct lexeme's tbal, are 'written 'yd pronounced in the same way. There
are separate di'ctionary entries for such words. Shrimp and base. are Examples of homnnyms. But. perhaps they art not.so obvious. Better
i•xamples of homouymi are shown in [2.1 I ].

i- mi ,t' , 9 •m»li flyití mainmal’

mi ,¢ , 'g u'oodni implemcnt.fór.hilting a ball iq crictét'
 b. Our.: but ,pip, 'the profession..of barrister'
bar ,z , 'a *'‹ntical..liM across a sJavt used io noirI meiriraJ xcce91. in musty'
bâr t$'tb, 'to fibsiruii'

By conirast, word-forme may have the sams prmuncia\ion but diffeient spehingr atid meanings. Such
£ùrmr are called HOMOPHONES. Ses this e»ample from a joke. book:

Why docs thv puny cough'' 8rc u‹isr èc'x z ir/r/è /i‹›*rrr,
(Yuung und Yuung l'/tl 1:57)

The jgke is a pun on fit:s/, the pronunciation of the two leieines represented in wriiinj by Durie and Ji‹inrse. Other examples rif homophones
inc lude iui/ - fo/e, suit - som, weather.- »-fiefiier, see - .run, reed - reed, rrrl - real, serii - srrrté, néed - Iuie‹id.
Coñversely, it is at so possible to. hat'e several.closely related meanings ihal are realised by tire same word- form. The name for tltis is
POLYSEXi Y. Often ,you find several senses listed mutter a single heading in a dictionary. Frir. iauarice, under ihe entry. for the nritiñ /orre: the
tED li sts over ten 'arises. I hin'e reproduced

1. Physical slrcngLh. RarcJÿ itt y/. ï= Fr. Éorr‹'s—III I S.)

* Suengih: impc\us. \’in1cnce, or intcttsi\y. of cffect ME.

ñ. Pow¢r ur mighl: asp. n ili\ary power ME. b. In curly use, the sizcnglh (gd x dsfensivc wr›rk etc.). Suliscq., rhc fighting siñnigth uf;a ship. J577.

A body'N’armodmen, at umiy. !n jv/. tic:truops ur soMic composing \h0 figs t°8 s ^•s* ul’ a kingdom or u commuter ME. b. A hody of {›olice;
of\en ubsot: r/u'fr n-r=policemen cold tively. 1SS1.

Physical sircngtli or piper exerted on an *bjcci: esp. violence or phjsicai.coercion. ME.

6 i\4cnial or moral sirtneth. Now only power of elTccti*c action, or nd u*crcoming resistance ME

The line that xejtaràtés polyrèmy fmm liomonymy is somewfiat blurred bccause it ir not 9ïtogether clèar how far meanings need to diverge
befoig .we should treat words representing them as belonging io distinct lexemes. In | 2.13], i‹ is not entirely clear that Lhe sixth sense of the
nnun /r›rre is not sufficiently removed frñin the other nieaninjs io merit an iniiy öf ice ou. The other meaningx all show. a reasonably. strong
family resemblance. But mental or moritl sLrength shows a somewhat wecker Relationship.
In the C/û7J, there is a scparaïe entry for the lexeme/‹zzz-e, the herb. lt is considereda different iexerr e
becadse it has a different iueaning end belongs to :a dilTérent word-clàss,. being :a erh ari‹t nol a noun. beÏonging to different word-classes is .an
important consiüeraüon in detemlining wfi<ther separate dicüonary ëntries aze needed.
In real-liée communiÿatiôn, fhc Ixc1. où. a one-to-onè match taf een lexemes and word-fônns dues not
nece.ssarijy causr -bisuity. In content,.the relevant meaning is normally easy to dctermine. Bud lhere ale
cases where it is non. For instance..thë homonymy ni ùo/ in 12.14j can cause semantic ùoùfusion:

|2.J 41

saw a At/ under the une.

 b. You eut it.

As the parâphrases show, the word-fnrm Air belonging in. the lexepie lii cañ rep'rcsent either the present tense or the. past .tense form of the
verb. In other words, there is xyncretism. We ha*'e twn different
grammatical words bit „ ,j 'and lift j„.y,, bui a single wnrd-form. The' same analysis aiso
applies to car. It can represeni either the préseiit or ext tense of the verb but.
Syricretism is not iirnited to verhx. li caw apply to nther word classes te:g..nouns). as well:

ïa) Tbe'.c'o1t Li1tcd a shcep aad,ön¢ leo-.

‹b) The roll‘ ki1lcd iwo shecp and three Jocr.

In th e tw'o sentcnéñs; ailiiough. the wnrd-form Shei•f belongs ter the”same lexeme. and is unchanged in form; we know that its grammatical
vaiue i's not the xame. In [2. lda] 'it reaiisex the- word with the gràmmatical 'properlies nf noiin' and xingular, but in [2.18b]..il rejresents'a.
plurai form..Likewise, the' same word-form dier represenLs: a xinguiar.nouù in |2.J 8a] and a piural noun' in [2.18b].
What can .wc say aboul the word' as an entity that furictions as a .grammatical unit in the- syntax nf a language? A's riienöoned already; the
tgrammatical) wnrd is n'ormally defined .' the' MlbflMAL FREE FORM. thal is used in the gmminar' of a Language. .Let us.now put.,somn fienh
on dùs te:rse and s mewiial crypiic statemenl.
By free fnrm wc:rnean ;in' entity thai:can stand ön its own and 'act as i free ;igeni; it is an .clement whose positiön in a iehtenée is, not.tötally
dictaled by' other items:.In order' tö explain 'rvhaf “frredom' mearis. in tbii context, we need In lake on board: two ancillary ideasi
POSITIONAL MOBILITY and STABILITY. Although wnrds are not the smalieit g'ramntatical iiniLs nd to, énristrñét sente'nces (see the
discusx'ion nf m'orphemes in .the nent chapter), at the level .of se'ntence organisation the..niles of sentence .formation txeat. words. as unanaiysnble
units. Often it is possible tn change the order in .which words appear in a sentence and xtili .pied née.:.a vieil-fnrmed ienleiice.. Wörds enjöy
considerahle pnsitiñnal mobility.. .Höweïer, the elemënts inside a word do uni enjoy suéh mohiiity.: W hile syntaclic ru1es.can üarispoit wöids tö
new plage in a sentence, they carinot ihifi in the same way elements that aré'found inside words. Moving-words around in the 'following
pr‹xiuées' gran;maticai .xeniertcgi; ivith basiéally the samé meaning', but with : ine'u'hat diffe'reni emphasis.

This ord induxn-tariff rtv.ixJicd Lane ter, fñriunstely,

b. Fortiinainly, hits ord indiisuialcii reviewed Lane:isicr, L:incaster, ihis old

indubtriwiist rcv}siicd, f;ortunatclt.

d. Fortiinainly, Lancaster was revisited by iliis old mdustrialisi..

Evidgntly, the position .pf words in a senteø e is not rigidly fixed. They can, and often do, get moved around if the communicative rreeds or
Lhe speaker ör writer require it. However, the interior of a word is a no-go area For syntacóc mle5. They are stričtl¿' barred from. manipulaóng
elpmeptx Found'inMde ä worcL As íär as. syntax- is concerned, words are ,indivisible ,units that:.can'noł be split and whose. inte'rnal. units are
inaccessible {cf. Bauer 1988. Matthews 1991, Lyons 19ßS, Hi Sciulln and Williifms 1987).
The. word a grammaücal unit ih ws xtabiliły {oi ØiTERFiAL COHE8IO.N . The order öf glemenii
inside a word is rigidly fized. If the elements. of a sentence. are xhilłed, certain meaningful units {in this case re-visit-ed. and ]òrtun-are-l›•) all
move øn blue, and their order always remains unchanged. The internal structure. of the Word cannot be. támpered wiłh: We are not allowed to
perform operaÓonx íhaÏ would field xvords like *êd-i'isit-re: ••öte-/ørfunø-lv etc. We.will return to thi's point on p. 33 below.
The definition of the wõrd includes the te;iți 'mițiimai' for a. good reñson. .This is' intended to separate words' fròm phrases' like J/iis old
iпdusrr..iałisi. Liie .words, phrases Can occur in. isolation and they 'can. be moved from.one posiñon to: another (as we have seen in (2: l9]).
But the expression ifiis old iùdurfrizifisf is not a minimal.(erm; since ił c'pntains smaller foțms c;ipablc of ocCurring;indepcndenłlF rnimely; r/iiśi old
atrd iiidusréio/!s'f. Furthermorc„ the sequence :fbíș old iiidusJriałrsr dms..not hav'e' the kiral qf hieternal. coh'egiöri found in words. It can:.be
interrupted by other words e:g. f/iis u'zn/rfii' old iiidusrriofisr,- fits i'em' wzo/r/ii', ołd, bè'nevolr*it indu.sfriofisf.
TÏ+e assumpÚon tbat the grammaúcaÏ word i'ș 'a minúnuni.he form' workš: well aș a ruJe of thumb: But it encounlers diFficuJties
vheo:confronted by a COMPOUND'WORD like x'hee/barrorr which contains the words ir-ù'ee/ and bgi.-row whiyh cän strand aÌnræ. In. 5uch ‹ seà;
it: is élear that. the z'oid is noł,.tłie wnalleśł meaningful.unit that.can. be used on.its owii. lł is for this reason diat the definition qf the worò as the unit
cm which purely syntacÓc.operations can .be performed: is preferable. In the case of..compounds thi's. defiiiiti'on works;. The interior of a .cpmpound:
ii a.'syntactié 'nœgy äreã. Syntactic rules are' not alloyed to apply separately to words that.make up a compound. Thus, för example although the
nouns H'fiee/ and barrow can be modified by-the adjective big. I fbig bar.-row], fbig »-hee//), and although we can taik of /brg wiirr/ònrrou'J, n
which.cäs'e b.ig modifies the .entire compound. there.:is no posiibiliły qf saying w'fiee/ Jbig. bori w], with the adjectivg only modifying the
.second..element:of.the compound xvord. .

33
SUMDMARY
in thiș chapter we. havü estabiishéd iiiăi:.noñîia1ty; the term 'word' ix uied ambigøo*xly.. Tø ävoid thè ambiguity, we need to distinguish betu'een
three differ;mt: types. of: word: ti) a wörd-fonri {i.e. a particular physical manifestation. nf one ör more lexemes in spench or writing); (ii) a
.vnabulary item {i.e. lexeme): and {iii} a unit of:giàmmaticai structUre that had Certain rñorphological arid SyùÏaCtic:properties.
We will revisit the disönction between lexemes. grammatical words and word-forms mainly. in Chapters 7 and 11. In Chapter' 7 our main
concern will be.' the realization of words in speech and .in writing. In Chapter 11 we will sh'Ow' that Lhii. disinCtìon. is'.not an
arletack of the linguist's analys'iy. Rathey, it ii. a disönction that:is well supported by studies in the way i'n which we. store. words in the:
mind arul retrieve thwn for use.in cömmuniyatiön in Areal.lifë.
In the coming. ćhapłyh. in Cpsçs where..the. re1ev.ant: sense of .the term “word’ is. clear from the..Coqtexł 1 will not spell out whether ił is the.
xvord as a: vocabulary item; grammatical word, phonological or młhognphic form that ii being dcait w.itli. Byf where it is ncit ćl tr; I will
ițìdicale the sense in which I am usinp this íerriï: We. are nõw..in.a posiöon tn. consider in detail:the internal structure of words. That is the:task of
the next chapter.

EXEftCtS€S

• Comment on the problems you encounter fn determining the number of words in the following nu ry rhyme. Relate your unswcr to the
différe nt hmm in which the term “word’ is tired.

The gmnd old Duke of York He had \en IJzousand men.

He inxrche‹J them up It› the toy of”the hill, When he marched them duwn
ayain:
When fhey were up, they were up.
And when lliey were down. Lhey were duwrt, And 'lien l7iey were nnly half way up
3I cy were neither up nor.duwn.

• Find. and analyse at least thrce cxample of advertisementr that cxploit rhe fiumonymy, polysemy or homi@hnny of wnids.
• Which t›nes.of the it:sliciscd vord-forms in wie fnllowing sen\ences belnng tu the.*ame lexeme‘* What üifficül\ies, i£ any, have you cume
across in üetemining whether worJ-Jörms belony to.thesai e lexei e’'

• . She su» him .com' through that planL of worxl.

• gif/ will pay rhe hill.
• I saw Fumtvr near ynur {uxnt ayain this mnming.
• Jan lookecJ / e/e when she walLod towotds tTie psi/.
• I am ¿i‹-1 of ¿our claiming to be sick all Use time.
• I was iouking at \he /›ook when Ake Br Ord the tick:e\.

• Usiñg ai least two fresh exam plcs, show how syncreiism yan be nd to support the distinction between word-forms. and grammatical
wor‹ls.
• This is the beginning of’ W.H.Auden’s poem “Music des Beaux Arts’.

RM Old Masters...
 These lines. cxn bc paraphrxsed a‹ ‘je.O1d Masters were neuer wrong ahou\ suff”erirtg."
Referring to the definition uf \he wt rd given in this chapter, explain why ii is correct to regcrd
,uiQeri«g as o woid but incoriecl io ireat uboiii sii@rriiig also üs a word.

Chapter 3
Close encounters of a morphemic kind

THE QUEST FOh YERBAL AT.OMS

We saw in the fast cbupter that the w'nr4 i the «mallext mraniriglïil unit of language. ibat can funélión inilepen‹kntly iv the grammar. A word
can be usod on ira oún, wiibout appending it mi .Anne nthei unir. Thus: in the vard rltildi.‹h we can isolate c’/ti/d and u.' • i\ on it.s ov’n bc¢aucc it
is a wnrd in its own right. But we c.annal use -i.r/t a a -slucd-al¢ine unit, for -i?/i ir not a u'ard.
\Vfiilc re.cogrtisinj: Itiat '•'or Is are the s'='•!!•*t meaningful uniLs which £un«fion indcpcndentty in W grammar, w¢ ulso nocd tn recognise
that wnnJs can be. decomposed into smaller units 0ia\ itre oJso meaningful. Our ursk in this vhaplcr in to explore llie in\cmat structure of
words. in order \o gain some wdcruanding pf jTie basic. units which.arc used to tonn words.

Originally 'morphology' meant the x\udy e£ biélogical formx. But.nineicentfi-century students of lazigiiuge borrowed the tern.and applied it.la
the.study of word-swucfure:. In linguistics. MORPH¢\LOCiY is the study oT\he farmution arxt internal orjanisation o£ wwds.
Let us begin our morphologicat analysis by «onsidering half a dozen wolds (not allege\her iendrim+Iy chosen):

hupc soun mcmd bu ) s8fc Ical’woril elephant

Ctbviously all Oie words in }3.I ] have u meaning, hut ink internal slruclufe. We cannot identify any smaller units thiH arc thdmwlves
meaningful which occur '=•'8• rherri. !f u Martian stopped you in a street near the lo¢al zoo an0 enquifod what pram in el«Rñaiit or ho in
hope means: you would thiñk she was asking a mos bizarre question that did nof merit u answer. Or you might c‹vide.aending1y explain ihut. ut”
cc’nJ,ine hcAmtho%hok.+orü-#n? methsug,buO1xpivcanmmbe%d(o ••••anyfhiugcmthek owri. Though somewhat
puzxlcd, rhe Ma+tian mighï accept your explanúiion.
But. 6eing the persisteni typr, let u.s .suppase she enquiied türit›er whether the words in 3,2} were also indivixible into smaller
méaningful iïni€s:

cbildish hopcless sooner rncnded clepbants rc•boil unsafe ex-wifé

Yöu uiould have in live a different answy.'You wyuld weed tö tell vorm. interrogator, who by now would be ketting increasi'nglv bewildered, that the

wnrds in [3.2] can be divided into smaller units of meaning as shown in [3.3.1

[3.3]

chiJd-Isi bope-tess x‹x›n-er nicnd-ect eJephani-i rr-boi1 uri-safc ez-wiiñ

The' part.qf. thi word that is not;ita1icixed cap function. as an indepcndeñt word in' the grammar. Indeed; each nf.the: nonitalicisml.chunks is a
word {i.e. vocabulary item) that is listed as such in the dictionary..B.y ct›nnâst, the italicised bits, thnugh meaningful (and their meanings can. be
indicated as shnwxl' in [3.4].), cannot funcfiqn on their nivn iñ the grammar.

-ish ’having the tobjnctiunablci quaüties of ehi1d•ish-- 'having the qualiiies-of a cbild'

clcphants= 'fru›re th/an ooe clüphant'

: 'again' re-bo I=. bo›l agua'

mi 'hoi X! iJnságfc= 'not sale'

cx 'former' ex-wife= 'Ióno¢r wife'

What 'we have dfirre in the words in [3.4] ean'bc done to tboüsands of. olher words iii Englisk They can be' decomposed intia smaller units
of:meaning (e.g..re- 'again' ) o'r grimmatic'a1 function-(e.g..-ed “past’).
The lerm M€IRPHEME ii 'used tq infer tn the smallest unit that has meaning .or sen'es a grammatical fiiñciion in a language.
Morphemes are4lie atnins with which words are built. It is not possible to find suh- morphemic units that are themselves meaningful or have a
grammaiicâl function. Thus,.given -lecc or on-, ii would make no sefise tn try lo âssign sñuiñ identifiable iiieaning to any pair of these forms. Of
course, it is. s'ible tfi isolate. thé individual Sounds d-1-s/ or / -n/, but those, sounds in' th Ives do not rpéy
We havñ'.nóiv established thixt viordi are.madc up of morphe'iñes. But hoii do we récognisñ'a morphe'me 'whén we see, pipit? .Our definition of
the, morpheme as die '«nsllesi unii of meaning (nr grammatical function) will be the:.guiding principle. Any .chunk of a wnrd with a: particular
meaning will' be said to represent a morpheme. .That is how ive proceided in [3.3] and [3.4') ahavc.
Mórphemes tend io have ñ fairly stabÍe rrieaning' xvhich they brin io any lord in hich diev appcar. lf we take re- and en-,'for example,.they
mean “agaiú' and “not’ respectis'ely—not just in:the wmds we have listed.abevé; biit.alt in thpusands. óÍ other w.ords.. Visua1fy mnrphe'mes:are uxnd
ag'ain and .again' lo form diffeÜnt wñrds. Thus Jr- meaniug “re-do 'wha1'ever the veib orleans' be atiaühed befóre .rnost verhs Io yield a new. word
with a predictahle mennmg. te:g. re-rum, n•-rod, rr-build etc.). In like rnnniv•r, un- 'meâñ'i'ng 'not X' where X ttonds for whate*'e'r thñ adjective -
é:i'ns) can be attached tn varioui adjectives' (e.g: yii-reo/, ml-elicit wi-i'appx élc.) tD yield a:new wnrd widi a predictable ne ad?e meaning.

The.'segmen1atioD of words into iiiorphemgs ih not a trii4al and arcane pastime indulggd in by linguists to while away the time on a .wet
Bank Holiday afternoan: lt is somethinq that is important for ail users of laripuage. During your lifetime, you will probably' eripounter
hundreds of thöusands nl different words. Many of these. words will be new to jöu. För no maiter höw ezlensive. your vogabulary is, you will
iievitably cnme acrnss words thal are unfamiliar. It Ïs impossible for anyone to know all the,.words .that arg fourid in English.
So, ivhat do you 'do when'faced with an unfarniliaF woid? Rcach for a goö'd dictionary!? Perhaps: Bm thix ix not always feasible,..Nor is il
always necessary. Vgry ofien .you just figure ont what the xtrange. word rru•ans iysing.the cnntgxt, together:with your knowledge of the
meaning of Lhe morphemes which the word conlairLi. You norma/ly dq Lhis subccinsciou.sly. Whal we are doing hey is making explicit your
tBcit knowledge of word-structure.
limi9irre this Ferraris:..hi 1992, a newspaper report on the war in the Bosnian republic states thai what w.e are witngssing'iS the Mhm
iiiiozinn .of 8osDia.. Supjx›se you have nol encountered the' word Lrbe inisnfiori before. Would y'ou understaud what the writer is say'ing'?
Probatily you would—williout Ïooking it up in any dictionary. How would yöu do it? The answer is simple. Bj .using -your knowledge of the'
world-in paztisu)ar history f6a/fanisotiony az/d ysur knoWledge of current afFairs (Lhe. civil. wer' in Lebanori) plus your knowledge of lhe
principles of woid:-Formation you are able to: work out the rneaning of

Lot yn.docuSpriziciples of word-Èormatioru You knöw fhat —i;'c/-ire is used When talking about natiof5 to rnean “lo maÏce X". e.g. from Aniei
ito we get Americaitice, hmm Xörea we get Xöreanise, fïom Aerr\o we ger Rer›’iznise cli. By attaching m-f«•ee we tum a noun into. a verb. So,
given the noun Lebgnoii we !c!m
/orm Liie †erb £eöa orrire. Nsxt, froro' th'e verb Wbgnoziise, we cq creBte .a new noun by addinB. -alien
( Wchfomuoounsofacüon)
If 'you kriow that iatiöus w.ar1ords Created warrirtg fiefdoms that:destroyed the Lcba ese xtale .during the

engaged in the Bosnian conflict risk doinp the same to the. Bosnian stale in the l09(is. Leborionisation is the act of 'turning a .counny into
anpfher Lehanon*. Thus, ‹iur knowledge of word-structure .contributes to our understanding of the. meaning of unfamiliar words.
 We have demonstrated that wor‹ls can be decomposed info morphemes. Now we are going to see thai
words have INTERNAL STRUCTURE. A simple way of showing Lhix is lo'analyse words' lii.e iiriraiinj' and ririfiappY. from these words we cail
derive uncannirr and unhappier. If you analyse unfiapprer, you will see thal extracting the correct meaning 'more' (not happy)’ ti.e. sadder)
rather than the incorrect one 'rrol {more kap{iy]' (i.e.. not.h@pier) depend's on the way we gröup togeLher thz• marphernes. Ïn. the fint ariuÏysis
wkere unhappier is interpreted as sodder the meaning 'not* conveyed by .iwi- is bracketed together or th happy (unhappy] as one unit and this
is intensified hy the -rr suLfs. ln Lhe u1'LemzLive second and ysiz, happy'. und - er ark brackeieü to@elher as ä unit hapfiiler) (i:e..more happy)
which then is negated by (un-] to gi e ’not more happy'. whiüh is incnrrect. Wheii someorig is uzifinppizr, it does not mean they are .simply lexs
happy, it meanx' rather that they are not happy at all. They are sad. This shows thal .morpherries in a word with several morphemes m'ay be
.grouped together in different waJS for semantic jurposes. The Way in which this is dor›e has semxntic consequences. Conceivably,
morphernes could be Lhrown together higglcdy-

piggledy to fomai

word. So long as you had the right morphemes, a well-fomed word would pop oul. But

thal is definitely nqt the eye. Word's have internal structuiai .grnupings, we have'seerl.
Furthermore, the sequencing. of morphemes in a word may be subject to restrictions. Take a wnrd litre
ungo 'e mbili which conlaifis four morpherues, pamely mr-, gn\ ern, a6i/, in'. E eryonc who knp s this

' o. b.

•ahii -uri-ity-t,uvem
•ua-govern-itv-ah1l eic.

Clearly, knt›wing is ward means ncM jur\ kno»'ing the morphemes i\ cnnlains, but also the rigid urder in which thty xre xIi‹›wed \o eppear.
We wiiI ie\um \n this pnint in soctinn f4.4).
Tn sum up the discussion to Iñr, wc›tdx are buii\ uring morphemes. If we kn‹\w how morphemes aie used tu furm wurds, we dn nto need \o
Its unduly Jluslored when we mme across a sttanye wnrd. Usually if is possible to work our the meaning of a strange word if ii contains
fumiiiar morphemes.

The idcntificatitm pJ’ morphemes ir rx›t altngc\her srraightfurward. Thir i» becâusc thurc is no simplo tone-\o- one cunespondcncc be\ween n\
nrphcrues .and \hc speech wundr thai rcpwent \hcm. In this scciinn we wiiI attempt to unravel tfi¢ ccimpIexi\i¢s ol‘ the re1a\ionxhip between
morphemes and the actual f‹›rms (+ound.‹ of t}rouys tel xnundr j fry which they aze mani£ex\cm in speech.

Altomorpks: morph famMiee

Any physical fc›nn that represents a m‹\rj›hemc is called a MORPH. The Farms -i.Ui, -/e. .’, -er, -ed, -.Y, re-, un- afiâ rx- in | 1.4j on p. 1 \
arc all mnrphs. Morphological analysis bogins wi\h the idcn\ifica\inn cf morphs, i.c. forms tha\. carry xoiae meaning nr are zssucia\ed with
some ymmniaticaJ func\ion. in «s/to/ugri.I iherc is juu one morph bui in all tfir wordx in [3.4J there arc two.
It is im}xinan\ not In cont”u e morpha wi\h SYLLABLES. When we \alk of morphs we luivc in mind suunds ‹ha can be related tn a
paniculcr meaning nr gran\ma\ico) f”unc\inn \e.g. plural ur past tenaJ. How.ever, when we talk of syllables all we have in mind.one chunks
into H'hich werJs cun be divided fur the purpo s. of pmnuncia\ion.
Ellis is nut en obstrusc distinctiui\. W.e arc nut. hcing pedantic. it is a distinc\inn tkat matters In.urdincry people becuuvr human langungex
are organised in such a way ihJi ilx construe iion of unit.x that are nicmingful is n‹›rmaIly in principle cepara\c Drum the cont\nic\iun of «brings
tAat arc prunnunceablc. Thus, for rhythmical effect, nursery. rhymes often use. non»tnm syllables like ’Krrdlr, drrdle’ in ’llcrdlr d edlr
duutpliay m5' .i‹iii mint’ which ññ no represent anything meaningful.
 Alternatively, a sound represontiny a mcuyheme may nof be a syllable in its nwn rigfi\, c.g. fry itself’, the -
.r which represenlx \hc plural morpheme is not a syllable. The.wnrJ ‹ cH firs Iwo morphemes, ‹uf and -.‹, hut it is all just one syllable. The
single syllable c o/ mali.ar two nx›rpfiemcs. The cnnv¢rse situu\iou. »'hero

mveral xyllables realise a .single morpfieme: is equally possi6le. Thus: tlie trixyllabic. anó quadrixyllabic
ward-forms e/epitome aud az/zorogur both realise just a’single morpheme:
The nauic of.fT›¢. rajatio.nskip bctu'csn founds and n›arphcrries is intriguing. At first *ight, it might )pok reasonable to ass.urns thai marpherncs
are made up oF-PHONfiMfiS. \Ve mgfit be temple.d.:to Ihtnk rMc caf, the English morpfi4nU w'iLh rhe meaning is made up oF th¢ phonemes /ka•f/.
hut we ha\’e scveraJ Linds of evidente -sbowing thai Ns ik tot the casc.
First, i£ morphemes les.èxode tqs of phoncrtws, a given morpheme wouid be umquelj associatcd fiiih.a gi\•en phnnoJogical zcpzcsentaúon. in
mality, the same morpkw can be.zcaJi.scd by different marphs (i.e. munds or writleri focrns), Maq 0i which realise the same rnorplieme w
referred: to es ALROMORPHS nf

Tt›e INDCFINITE ARTICLE is a good exemple nl’ a morpherce wiih. more lbatt nne alÍomorpb. li i's rcalixed by the iwó for ns e and l«. The
sound at the begiiu6ng of the followirig woid. òeierrnincx the allomorph that is xelècied. If the v ord iollówing liii. indeliniie anicle beginx with
.e corisonani, the allomorpb u is xelected. but il‘ it begiex w'ith a vowel the allomorpb un ix need inste'ad:

b Inn ixllmd

del t9cning' avi upinion

Hence ihe incoriecmess of the sentence rnarked with an asteriak in {3.7]:

]3.7]

:1 spcru em evcning with tbcm.

b.

24 CLOSE ENCOUNTERS OF A MORPHEMIC KIND

that allomorphs are forms that are phonologically distinguishable which, none the less, are not functionally distinct. In other words, although they
are physically distinct morphs with different pronunciations, allomorphs do share the same function in the language.
An analogy might help to clarify this point. Let us compare allomorphs to workers who share the same
job. Imaginea jobshare situation where Mrs Jones teaches maths to form 2DY on Monday afternoons, Mr Kato on Thursday mornings and Ms Smith on
Tuesdays and Fridays. Obviously, these teachers are different individuals. But they all share the role of ‘maths teacher’ for the class and each teacher only performs

that
role on particular days. Likewise, all allomorphs share the same function but one allomorph cannot occupy
a position that is already occupied by another allomorph of the same morpheme. To summarise. we say that allomorphs ofa morpheme are in
complementary distribution. This means that they cannot substitute for each other. Hence. we cannot replace one allomorph ofa morpheme
by another allomorph of that morpheme and change meaning.
For our next example of allomorphs we will turn to the plural morpheme. The idea of ‘more than one” is
expressed by the plural morpheme usinga variety of allomorphs including the following:

[3.8]

Singular Plural

a rad-ius radi-1

cact-i

b. dat-um dat-a

strat-um strat-a
. analys-is analys-es

ax-is ax-es

d skirt skirt-s

road

branch branch-es

Going by the orthography. we can identify the allomorphs -i, -a, -es and -s. The last is by far the commonest: see section (7.3).

Try and say the batch of words in [3.8d] aloud. You will observe that the pronunciation of the plural allomorph in these words is variable. It is [s] in

skirts, [z] in roads and [lz] (orforsome speakers [ez]) in branches. What is interesting about these words is that the selection of the allomorph that

represents the
plural is determined by the last sound in the noun to which the plural morpheme is appended. We will return
to this in more depth in section (5.2).

We have already seen. that because allomorphs cannot substitute for each other, we never have two sentences with different meanings which solely
differ in that one sentence has allomorphX in a slot where another sentence has allomorph Y. Compare the two sentences in [3.9]:

[3.9]
a. They have two cats b. They have two dogs

[el heev w: ket-s5] [el haev w: dg-z]

*lel haev w: ket-z] *[el haev w: dg-s]

TGL tSH WOitDS 25

We caunol find two otherwise idefiticcf”.seriten¢:es which differ in manning simply because tJte word rrir.r s pronounced +La [W-.s] and
’(ka•i-2j respectiveiy. Likewise, it is not possibie to have two otherwise identical sentences with different meanings wh<ze the word dam is
pronounoed as |dgz] and °]dg.sj. In otter-word.s, the difference between the allomarphs I â] and |z] of I1ie plural morpheme cannot lx• u'eJ to
distinguish meanings.

Diffgrent morphemes CONTRAST meanings btit different uJloninrphs do not. If' a difference in m'eañing is attributable to the l'act that one
minimal meaningful unit hns born replaced by another; we identify thr morphs involved m man›fésta1 onr of disúnct mórphemes. So, in (3.7]
on p. 36 the:indefiniLe article realisc‹l cby or un i.s a distinét morpheme fmtn the dcfi»iie article rea1i'snd by im. sinée a seinantic difference is
detectable when n neon is replaced with the.
A further example 'of contr'ast is giverl in [3.10]:
 a. I ualixkcd \be door. b. She is untidy.

The two senlelices in (3.1 l - •• ' d'fferent things. Sirice they are identicai except for the facl that wh<cc onu has «*- Ible other has re-, fie
difFerence in meaninj between these two senlences ir due to the difference ii rt›e•rii»y. hciween the morphemen reaiised b$ rr- (meanieg ’do
again’) and yor- (meaning ’reverse the action'T.
Now, sontrcsI the min- oF u tforV wilh the ziir- of r‹nrizfs. In both ccses we. have the same morph riri- (which i spelt arid pronounced in
exactly the ;same way). .But it is obvious Lhat rin- rupresezits diffsient morphemes in these two word-form.s. In I iizi/ncied rhe drier the
morph un- fnund in uuloi-krd red ises a rrv.eisive morpherng which is aitaéhed ig verbs-it reverses die.artion of locking: But in mime it:
realises.a negative rinrpheme attached lo adjective unfidv.means 'trot tidy’. IIf a perspn is rir/idi, .it does not menu thai at some earlier point
they were tidy xnd someone ha.s reversed or undone their tidiness.j
GF morphemes were made up of phonemes a simple correlation of'marphs with morphemes is what we would find. But, in .facC. it .is qtiite
common for the same phonological form (i.e. inorph). io represent more than one morpheme. It is from the-context that we can tell which
morpinme it represents. This is the xecorrd piece sf evidence' against:the qsxumption fhat morphemes are comp<r‹ecI of phonemes:
The complex reJationship between morphemes and the aliomorphs Lhat represent them giyes us a window thmtigh which we can. plimpse
one of the musi fascinating aspecLs of langiiage. tlte relaiionship belween FORM and FUNCTION. In linguisticx we explnre tli'e form of varióus
eleiiients óf languagé xtnicture, e.g. wonls Azid sentences, becau¿e it is important to know how fhey are constructed. However. form is not
everything. We are alxo interested in knowing what linguistic elertients are used for, what function they serve.
Just consider for a moment tlii's non-linguistic analogy. Imagine a:liiend returns frnm a foreign vacation with two beautiful ornamental glass
contiiiners with a globular shapr and gives orre to yoiL as a present and keeps the”other for 1ic'rself. She does not t'ell Jon what' yi›ur present i5
tised for. She uses hers'as a vessel for containing wine at the table—she got the idea of buying thexe containers when she was served wine in a similar
cmitainer in a fancy restatlranl. Yo* do not.know'thi's. You loak at your pie'séñt:and decide lo put il on the

New Yorli - Nvw Yorker HightauM - HiyhlanJor

Ttie • • form, -er, iepre»ents threr dit‘ferent meanings and hence be» io be unsigned to three distinct morphemes. In | 3.1 I a] ii form.s on «genlix'e
noun frum a verb, with the m•• › .g ’someone who does X’ (i.e. wiuievei the vcrb meuniJ. In {3.1 I b] the same -er tonne an instrumental noon from a
verb.. with the meaning“ ni cfhing uxcJ \o X' (i.e. \n dn whale.ver lhc verts means). Finally, in | 3.1 I ñ] llic samc -or ft›rm ir wracked \n u noun referztny \o
a place \o mean “an infiahitan\ of'.
Clcarly, fhc same 1brm does wrve different 1”unrtions Axle. So, it realm.set diflfimn\ morphemes. This is further evidence rhat shou\J quickly
disahusc us ni‘Ille ussun\p\ion thou morphemes are mude up ot‘ innrphs. Not only cau u single. m‹xphcme kavc several ulIt›inurphs (as in the case
t›£ the plural murphciau). the sxme rnorph (e.g. -er) can represent different morphcm¢x. 3Tx•rd ix no simple one-tu-and matching oT” inorpfiemas

FRREDOMANDBONDACE
Whcn ñ¢ clarify rnorphemcr.in tcrfr uf.where thcy arc allo cd \u appear, w¢ finJ lhxt ft+cy fa4l inin tw‹› major groupings. Somc morphemes itre
cupiihle of.ncr urring on Iheir nwn u-s wnrds, while othci rnnrphdmer arc only allowed \o occur in cemhinatinn wi\h sem¢ u\her n\nrph¢n\e\ sT hu\
they ciinnnt he ured fry thcinsclv.cs ur. indcpend¢zit words.

Those .morphemes thai are. ailñwed to occur on tlieir own iii se'ntences as wordx are called FRED MORPHEMES while those morpbeines that
muit occur in the company of some oilier triorpbemes are called BO.UND MORPHEMES. tn [3. i2] ihe bound morpbejnex are italiéised.

pcslt i)•c-'ide

child

ehi1d• isfi

prr•puc¥•rd

\sugh

The. free morphemes in [3.12] can all be marùpulated b'y syntactic .rulex; they can .stand öii their öwri as
words. By contrast, it is impossible to use the forms -cidc, posr-,: -ier. -i li, pue-; -ed or -img, independenil$.
So far; ail theexamples nf fïeë inorphemex thai furictiön asroöis.tha1we have ëncountsred havñ beëii cnntent words (see: p, 14), Hnwever, not all free
morptiemes are content wnrds. Some are employed to indicate grammätical functions 'and logical relalionsbip rather tlian to convey lcxica{ or'
cognitive meaning in a sentenée. Heuée xuüh wörds are üalled FUNCT1O1'ï WORDS, They inilu& words such as the fñllowing:

3.i .1

In ordinary language use such words are extremely common. But on their nwn they would not convey a lot oil information'. If you cx:eived ;i
telegram like.Bal ir niS• oii to Ohr 'iy you might xuxpect lhat the. iender either had a strange sense of.humour or was not meiiiaily sound.

SOUND SYMBOLISM: PHONA 8THEMFS ACD ONOMATOPOEIA

In the 'vast rfiajority of wards,:the relaÖonsftip betwe¢'n söund and mtIming is arbitrary (sec p. 2). There it no reason why. a partictilar xound, pr
group of soimds, xhoulÖ be used to represeni a partictiiar word, with 'a particular meaning. 11‘ someone asked you what | b] in:bed. or [xtr] in straii
ge meant, you would think they were 'asking ;i very'odd question. As a nile, sounds quo'sounds do not ixu•an anything.
However, the genéial pritiéip]ñ that s'ayi thai aé lia between sound and.meaning in wnrds li.arbinwy }:s c*ccasionally dented. This happens
in lwo scre of circumstances. First, certain individual sounds, or groups of soiindxj which do nor represent a specific enough meaning to be
called morphs. iieverthelgxs appear 1ci'& vaguely.assi›ciated with some kind of meaning. Such.snunds are Cailed PHONAESTHEMES,
As our firsi example of a phon esiheme, let. us iakr the RP vowei { ] twhich is historically descended from (H], the x'owe{ that is still used in
word.s like duf/ and liar in thr. north of lingland}...Thii pfronae'sthcme ix found in wordx.associated with varie'us.kinds nf itnllness or
indistinctn'eix, e:g. dull. tied, tinker dusk,

28 CLOSE SNCQUNTfi BS .ÖF A MOkPH£MïC KM D,

bİunc, mud,: slash, sludge, shrimp ełc: Obviously, the vowel ț ] per sr.does not mean 'dull'. If' it did, dim
which.çnntø'ins the vowel {1] 'would n'ol be .a vuiual 'syn'oñym foI:dull.
 Manv wortŁs which,mean “in talk indistinctly* contain .nne or more œxurreøcës of the labÏai Consoiiant.(m], which is made with. tłie lips
firmiy closed, preventing. clear articulation. That. way, the very act of pronouncing the:word.iünriícaily mimics a kcj' dipect qf its meaning, You
é'an 'see this if.}o.u wałch.yourself in .mirroț saying wprds,.like mumble, reum»yr, em/rgr, ziiezed, gmmb'le.etC. It i* probably not . ¡lccidgRt
that tłiexe words also contain the phonaestheme [ ). Similarly, the sourul: [ mp] (spehed -mпp I zs in clutrip, duøtp, fuiinp', loтpMaud hamp is
often foun'd at the And óf words vihich .ãre .ãssociated with. heavirrëss and clurnisinesś although. no one Would wi'sh to sugg'ext that -ump. in itself
represents the ideas of heavine'.ss 'and clumxiness. Interestingly, heir agai'n we have the vowel [ ) followed by die labial consonants [mp].
Observe aim that whereas [ j tends tñ have /ãsiocîation's of.beavineis öé dii1lness,.the high:.irønt vowels li:] and [1] frequently .occur as
phonaesthemes in words associated with smallness, as in wre, røøn):-w'eeziv, lean, mragre, тiвi, thin and little. (The fact that bi g has Ltte opposite
meaning just. goes to. show that phoñaestbemes ñrilÿ rëprexent a ten'dency.)
Second, and more importantly, in additio'n to phonaestbem'es, there äre onomatõpoeic wõi‹ls-.in which a dirrct associaóon' is made between the
soimds of a word-form and the rneinirig that it repres; ;ts. In cases of OHOMATÖPOfilA, die”sounds (run sounds and not ãs.rnmphs)
symboliie:or'refiecł soгne: ñspcüł of the
aniù öf the wõrd .that tfrey ripresent. So,. if speakers-of any language want an .onomatopœic weird for the noise a cat makes, they ivi1l not
choose a noise like biinboban xcept, pe'rhaps, in the land öf the fling Nańg Wrong.
The. worAs :for. souridi made by variofis .animals e.g... wei;gh, nlioow: sion ełc. are the most obfious examples of onomatopœiœ But Were
others siich as roor„ rrack, rlaпg, bang, șpłash, rø'ish, whoosh, box .İi(s.s:' șh'eep, blèép,. ğufiglé, plop and plod. In the base of onomatopøeic words,
íhe relaÓonship betwèeu soimd and meaning is to some extent lfiONlC. The sounds mimir- an aspect of'tłie meaning. of the linguistic sign much
in tłie sämë way thnt this iconic' sign for a re'słaurant represents, more ör lexs directly; the meaning 'resłauțant.:: This Symbol is. Still
yoDveñtional lo some degree. To. peoplg who eat witłi Cbopstiéks, it might not be anirøeHiately .ohi''ious why this sign represents .a restaurant
{rather tłian a cutlery shop), but onct ìt is pointed out the link üatr be seen i}uite'ëasily.
Onoma1o}xieic Wqrds are ićonić. in so far aS th'ey directly reflył somè aspecł of th'e reaping of what'theÿ'
stand for. So;.conventionally in English cows go. 'moo' and horses get 'neigh' and bees go 'buzz'. That is why Spik'e. Milligan’s nonsi•rise'.poem.“On
the Hing Nang firing’ is bizarre.
Tn be' onomatopoeic, the 'sõund must ipyitate to. some degree ari aspect of the noise madg. by the bird 'nr an mat. But exacdy wbat is imitated
wúl var}' from language to language. Aa English cock will say c fzdood/ćdon, :a Russian ‹xrck kHurikn and iri Uganda it maÿ' sãy
›cfn/i/oośpp. (These differences are not Attributable' to dia1ecti'cal variation agiong the males of the hollv.F doønrdriciis species:) Onnmaiopoeic
words.are rrot puxely and simply formed by mimicking precisely the meanings that they convey. To snme ëxłent.:on'omatopoeic ø'qrds'are aisp
mnuldgd.by linguistiç.ćonventioñ. That is whÿ' irrdiffert•nt p1ać'es in the world different onomatiipoeic viords may be.used for the,sarne'.arii'innl oi
bird rioise.

VERBAL BLUEPRŒTS.

Linguistic. theory incorporates' tb'e hy}xiłliesi'i tha't there'are:.universal principles .of grammar that:rëgulałe the ammint. of variatiön in liujuistic
structure aèrõs's' languages: In the lasi. sectiön we saw .the marginal role played by soimd symbolism in word-formation. This doe.s rroł obscure.the
fact that normally languages form
 On the Ning Nang Nong
On the Ning Nang Neng
Where the mws go Bong!

Trees go Ping! Nong Ning Nang! Tbs mice go Claag

VI €’RI1SE IN.Cl1tJNJ’ERS HP A MORPffEM It KtNEi

words by using sounds in a now-imiative wey. There is an overriding ierxkncy for the relationship between founds and meetings to be
arbitrary. innally there i nn n why a part.iculai morpheme is rc,ali•ed..6y any particular soueds. The
choice of the allonioipb or allomorphs .that represent a particular morpheme is

t)bviouxIy, a . every.one knows„.ßl1 Iitnguages do not Ïia've rhe..Hume worth, Sinće virtually пzty arbirrвry mułi:h of cunó and
meaning cвn pmduce a word, i\ is not surprising thitï wпrds vary grsaUy ii\ rheir structure acroxx lunguajes. But Ihis does noï ïneaa
rhat chaos rcigrr•. Tlw w*ys in which morphs use used te farm words is. reguïвİc0 fry gcncmI pńnciples. So. rhe umoufiï nf
ürossliñguist\č vãrìatiun in work-formation lulls wilhin rcrtвin brnad parameters. l\ ir m if there ir u menu at”blueprints for worJ-
toririatiun.from wfiÏch кli I tiguuges md¢ü their xeïecłîons:

ii/ łS0LAT/NG lw ¿ut/ bag tc

iii› AUCLUTIHATTNG lmguages
liiiî++ ltNGlw øn+«1c› I guøc

l•io language. makes all res choices fr•m just one pørt of the menu. To varying degrees ail languages make mixed choices. M idea of
t1ii.s ménu i:s to redicate the predominant word-formation tendencies, if they exist. In the su6seclionx 6elow we xhall consider in Um
examples of the different morphological tÿpcs,

Iri an archełypical isolating language the Word is virtually indisúnguishiißle:l‘ixiin itie. mi›ipbe.me, for e*'ery word conisies just one
morpheme. Every morpheme is a free niorphenœ. There am no łwiund morphemes. Vietnaniem comes clme to this ideal:

a. Tôi quu” bóng vè hn 4 dà \ôi

b. Cbúup . i6i mua à gu

 Typically, llič Hurds arc short.aod cnnlain jusï ooe morphemc eвcfi.. Almnst every concept is expressed by a seyвru\e word. Look
ugain. for example, a\ ihe Izcalment oł” pasł ïense in vcrôs {e.g. pizncńcd. I nıiphi› md Ihe plurality of x”e (plural plus Nrst person).