MORPHOLOGY,MODERN LING ISTICS SERIES Sere tors Poesor Noel Ruston Roberts Univer of Newcastle pon Tyne ach txtook inthe Mun Linguists src is designed to provide & ‘arefily graded introduction to atop in contemporary linguists ae Sled disciples, presented ia aster that acca and trate 10 "ede with no previous experince of the tpi bat kading them some Imderstading of erent ses, The text are desipned to engage te active print of the reader, fvouring a peobemsolving approach and ‘clang Wr nd varied exerci mater Ties publ the series glh Symes and Argumetaion (acond eiion) Bas Aarts Phenaogy Pip Cart “Lingus and Second Language Agustin Vivian Cook ‘Soelingutes’ Rea and Cowsebook "Nikolas Coupland sn Adam Jaworst Morphology (cond edition) Franc Katamba ap os Stonkam ‘Sonar Kate Keats Symiacte They Geofiey Poole Comact Lanewages: Pid ant Cres" Mark Sebba Further tiles preparaion Sos Dein, lo Dito La gtk Hone Ro NS Ed Morphology Second Edition Francis Katamba and John Stonham StesUle Wace fedrehfesic palgrave macmillan _ gaeonaiyunwynn ei py ain To Janet and Eun-Sook to pega i pinyin pel end ye ug rd ee oy ce ‘Banmhrncasr donate a Te the mtr nd ig en a ato stoner ih ph Dw Pa A 88 Seka naga ‘tn sereetee York wo Cond wpe eae at ac aca eb aang oh Pg ‘cna dune art nl ag hn (estar tee eee ne tes ese ogee {piste ces hg sme et ne ee five Pat ieee nel ‘stooped ange atefrncing daa a iy ‘hg cua are ap ryan fete eeetcoonoms erence ey foe Coenen at "phy anes amt adn Soto. Fees ttopup eae odd, as a) SN -a “Mee, pn ngs ep ta oT rs andbond nc ran by Cana ace copra anteContents Prface Prface 0 the Scond Eton Acknledrmene ‘Aiteratonr end Syboe ‘he Internal Phonetic Alphabet PART 1 BACKGROUND INTRODUCTION LL The Emergence of Morphology 112. Morphology in American Stata Lingus 13 The Coney of Chomshyan Generate Grama BIL The place of morphology early senerative 1.32 The morphology-phonology intriction 133 The mombology-tytaximeracton 134 The morphology-semantics interface 135 The leon 14 Orzanisation of the Book INTRODUCTION TO WORDSTRUCTURE, 21 What ia Word 21 The lewne 212 Word 213 The erammatial word 2.2, Morphemew The Smallest Unis of Meaning 22.2 Morphemes, morphs and allomorphs 223 Grammatical condoning, lxkal conditioning and supletion 2. “The Natre of Morpher 24 Sanmary rte Reading TyPES OF MoRPHEMES ME Ks ties, Stems and Sia Roots S12 Atenwit Conon Contets 3.13 Roots, seme and bases LIA. Stem extenders 32. ifectinal and Dettloal Morphemes 33. Mute finan vy ae 1X7. Morphol Type Ag WP the Conrlit ofthe Word 4 PRODUCTIVITY IN WoRD-FORMATION {AD The Open-Ended of he Lesion 1 What fe productive? 12. Semipredactsty {N13 Prodaciaty and ret rit on Prodctiy 421 Blocking 43 Does Proustity Separate Insti from Derivation? 4 The Natre ofthe Leicon 41 Potential words 422 Kowa of language andthe role ofthe exon Further Read | PART I MORPHOLOGY AND ITS RELATION 10 PHONOLOGY, 5 INTRODUCING LEXICAL MORPHOLOGY 2) The Lesical Phootlogy and Morpology Model 2 Lexa Strata S21 "Derivation nese morphology 2.2 Inston ales! morphology ‘33 Lexa Res 124.ilferccesbtneen Lexa and Pest Lencal Rls Farther Reading Eyerehes 6 INSIGHTS FROM LEXICAL MORI 2 ies GET Stator ordering refecting morpeme sequencing (622 Stratum onering und prodctity aepaeaeens aeecezcausess SeeEiseee 4 nL 1p (623. Suatum ordering and conversion G24. The Sit Cy Condition Farber Reading Exerees LEXICAL MORPHOLOGY: AN APPRAISAL, 71 ntedacton: The Claims Made by Lexical Phonology 72 Creins f Leia! Phonlors ‘TELT" Are les! stata determined by affine aber ‘han rots 122. Doafizsusiguly belong t one stratum? 724 How many sent are neal? 724 Are phonologal rules estited one statu? 735 Are morholosical ales resicted to one stu 78 Conason Exerc TEMPLATIC MORPHOLOGY A ntodscton| 2 Phonolgit Poade: Antosymental Phonology AZT Atosepmentl ptoology: mapping principles 422. The able er 2 Root and Paters Morphlogy S31 "Anbie Biyanir 832, Prosodic morphology and non-oneuterative rmorpbology 433 Thermorpheme er hypothesis 4 Conton Faeries TEMPLATIC AND PROSODIC MORPHOLOGY 91 What Redpleation? | 92 1 Redaction Cominst Copying? 92 Cy Templates an Redapication 531" Undenpecicaton 9532 Redupieaton as pretation 533 Redupention aaton 934 ner edupeation 933 Redupintion and fod sepmentsm 9.4 Prod Maran) Other Prod Phenomena OT” Subactve merphcogy 982 Rng expletive ina us Rh rr ry ba bo 0 1 1st 1 1 155 162 18 1685 m a 0 180 1 18 se has 9 180 192 199x omens omens 26 Cochaien Faeries OPTIMALITY THEORY AND MORPHOLOGY 02 The Bas 03.1 Notation 10222 The interaction of ensues 103 Morpoloy in Opinalty They PART IIL MORPHOLOGY AND LEXICON AND SYNTAX RELATION TO THE, 112 Inflscton ad Desvation 121 Diffeentadng beewen ifeton and derivation 1122 Relevance an geneity 1123. temorpholoeynecenary? 113 Vesa neta Categories USB lnkeent verbal proper 1132 Agreement properties of ves 13.3. Cenfgustionalpropertion of ers Ua Inherent eatporie of nouns 1142 _Agreameat extegorie of nouns 143 Configurations etegores of ans eerie 12 MORPHOLOGICAL MAPPING OF GRAMMATICAL FUNCTIONS 122, Prediaen, Aguments and Lexical tin 123 Theol and Leica Eats 124 Grammatical Relations 125. Grammatical Funcion-ChangngRales 1252 Antipasive 2m am my ma a x 21 mm 76 1253 Apnliative 1254. Camative 126 The Miro Principle 127 incorporation IRI Noun incomoraion| 1272 Verh incorporation 12.73 Prepon nsomportion 128 Conc 13 THE INTERPRETATION OF THE LEXICON, MORPHOLOGY AND SYNTAX 1 nrodctn: The Intertce Between Modules 152 Phonologlal Factors In Compodnding 133 Are Compounds Diferent fom Syatacie 133.1 "Toe notion "word svi 1333, Lene 333, Unoted mompologia objets 1334 Syntace objets and sytacuc atoms 134 The Character of Wort Formation Ras T341 Headenes of compounds 1342 The righthand head rle (RHR) 1343 Lefteaded compounds 1344. Headless compounds 135 Compounding and Derivation Tas Crasberey won 1382. Neos compounds 136 Cc. 137 Conchson goa Tex Sale Ide 2 26 = Bs 296 Fa ou 310 316 08 a ue nl mi mIPreface “This books an eoducson to morphology that pesuppses ile previo Verbs NP) whigh enumerate permisible com Tuts of word in phrases ad Setenss fom russ of word-stuetre Ite the oe in (13 ht ives wale from ul, All ese rs ae banded Inesher bose they are sonserned. nih ceumerating permissible ‘Stubiaton of mores eabove)- Note, however, hat this reatent Sr Sptactialy motivate eration in the or of words controversial. ‘Welhve moey ied the probe Torte present. We pospone detaled “scion wnt Chapter 1 1.34 The Morphology Semants Interface Turning to semantic, the connection between morphology aad the lexcon| mth on had wih sung on the oer obvious ie a major role of the xeon o tomar to nthe meanings of words This because ually the rsaioshipbetwsen word and is meaning is arbiter.Onan ofthe Book 6 ‘Theze sno reason why a word asthe particule meaning that it has. For ‘msn, yo jn veto memorie the fact tha the word alert ind of beaddrex worn nthe seventeenth century. Tete no way tat {you could dacoer ths far fom the sods or he sree ofthe word ‘We will come back to this loi a Seton 13.32. AS The Lexicon 113 es inmeditly obvious hat, nation tinting the msg of| ‘words and morphems, the licon mas ala stor other Kinds of infor tion reevant to te application of ntact and phonclogkal rues. Satan reeds to have acest 10 morpiosyntctic properties (Le, popes tht tre partly morphological and partly synticte) sich as wheter noun iS oumtahe te spades or uncountable like eownment ‘This aes it haviour in phases and senenes. We may sy he paleo tes spe but we can only sy sequen (nt "these eum) ertermore some phonolorial rules apply to worde dren depend ing on thar moron tactic proper. Fer example, sane phonological rly ae sesitve tothe flees tween nan and verbs Thi, the ‘Nord pri the main ses falom te Ba sla the word fenton 2 nw (petit Bue Fanci sa er (mig) in ses fllson the tt sible. Obvios for phonological rules hat sign ese apply correc, acceso such morphonymtatc information resi. ‘This afemation must frm part of he ety ofthe word inte lexicon "Tae study of nopholgy thesore, cannot be el contned The 4 turalist doctrine of teri separation of Knguisti bv skated 1} Unteable Trae, thar are sme sss tate fhe ners concths of ‘morbolgy, but many torptologial problems involve the intracton tetween morphology ad other modules ofthe amma Fer his reason, mck ofthe space the capers that follow devoted othe eration etmen the leicon and merpholog wih the eter modules 14 ORGANISATION OF THE BOOK The book i nears Follows: Pare 1 (Chapters 1-4) ineoduces basic concepts and tains otons hich ae fundamen oll morphologtaldions, Part (Chapters 5-10} explores the relationship between epopboloy. phonology and th lexicon in arent generative theory, examining sever odes of morphology. Part IM (Chapters 11-13) deals withthe eationship betwcen morphology and syotx in eaten generative theory ‘Ove he yar, hee ve been several morphological theories that have teen propose by linguists, One way of introducing ou to morphology Would be to preset sore! and comparative survey. We could Bave ‘hamine varows theorie i arm apd perape compared them. Or lr tative, we ould ave ben poem td peeeyising. trying operands outta a preted theory isthe best thor. Tha snot what me shal Jin this book Tasted, we present yoo, sympathetily, Sata the sme tine eal swith one theosesaly cert approach 0 morphology. namely tether) ‘morphology ncuron maisacam aeerave ganar. Tie decsion ‘Soule nt only tossed ete dominant model nthe Bi oda, bat tk beans ws thik eff he mot promising solos to the prea [woblens in morphological analysis ven 20 the book it intl sete, We have aot attempted 0 ssprvet every shade of opinion wih the gveratv school Rater we fe cas a ides and races that sce to orm part of the emerging “nonin mainsteam generative morphology. Obviously, to some extent ‘his fea matter of abjecive judgement In some cam the jdgements hee ay not be the me thw of se ter ng Th pacclr, we have only brief dicised What now the dominant ‘woud of phonaogy, Opiality Theory in Chaper 10 The main reason {or hiss that the tcatment of morphology in Optimality Thecey remains in “rather cent ste wih many question remaining undid. o ee ‘sepiored. Where curent Optimality Theol analyse ae avalible Peipally with respect to redupication, they will be discussed inthe et (Of couse, morphological theory in. cucresk minsream generative rammar docs ot enjoy a monopoly of sigh. The deb owed to ater “nproaces wil he cede in hs Book xpeclly in the eanychaptes and sth ecommended readings ad iho phy ‘A‘major fre ofthe Book i tht ou wil Be asked fo be an active issn, not passe reader. We have endeavoured to engige 308 sctnely and prety in dong morphology rather than in merely ening shout is hiory apd watching om the al how its done. As yu read tchehaple, you are ssked (o pause a pees and answer i-iet questions exces before proceeding (he questions abd etre re ial by lise sro the pags), Each caper ae ths one) ends wth Tate ‘eres deine with ois ied nthe body of he tex. Thisnesence on felling you to analy dats is du to our feb conviction tha the best ition for nyo who wih to ocome Hingis todo ngs inal ight rom the st rather thant rd abo inthe ett new morphological terns appear in Wad type ad they are expluinad when They te hist ittadced They Wil als appa in he Hossary at thee of this hook whereby fern from other branches of Hnguintc re abo explained. Transco of Standard RP, English arefa Introduction srawn from Jones and colleagues (203). For any other Hngusi temas that ae unimii,apood dictonary of ngustcs, sich as Davi Crystal's 4 Ditonary of Lingus and Phones (2003) or Laure Baers A Gla “ry of Morsolog (200), shoul be cosa 2 Introduction to Word-Structure 2.1 WHAT IS A WORD? “The sumption that Languages contuin words is takon for grated by nowt people Eve ilieratespenter know that thee are word the lingutge. Tru, smetines there te difrenes of opinion ao what ite ue tobe tet as words Fr instance, English speakers gh not agree ‘ether alright tone word oto a at rest dips may are 38 teople can ealyeeogase a word of thls lngusge when they se or tt ‘me! And aormaly their Judgements a6 to what Is oF isnot a word. do side Engsh speakers are, for example that the frm od i the Se Ty ct ihe manta Eh ned 21 The Lexeme owever, loser examination ofthe nate ofthe word’ eval somewhat tore empl ltare than pant above What we mci by "word not nays Gear. Aa we shall fein the next few paragapte, eas fa Taig the ature of the word are largely de tothe fae ha he tr ‘nod used ina variety of senses hat wll ae not early dstagused In faking the existence of words for granted, we tnd (0 oveook the ‘tuple of whit we ae taking fr arated What would you do i you were reading «Book and you encountered the ‘wn poked forthe st ine in this conte? He went to dhe pub for pit an thes peck of ¥ would probably look up chat unui word in a ditonary, ne Ur poker pack Tis Bsn you Know that pckied i Told you hat the words peng un poker wll ho ei. Farherore,i Intron to WordStractone Morphemse The Sales Unt of Meaning » you know that poking. poke, pokes and poked are fre maniestions of he sme absent Vocus em ‘Weakall efr othe won inthe sense of watutvocbulaty em sing the ter lxene, The Tors pocing, poole, pokes and. poled ce iret ealsatos (or representations or manifestations of te exeme oCKLE exmes wile wien seat eters) Tey al share core Inesnng although they ae spel se pronoun aileelly. Less ate the yoabuluy Hes tht are listed the dhconry (ef, Di Slllo und Wis, 1989), xr ‘Which one of he words i [22] Rlow telons to te same lexeme? 2} see catches ler boy caching ses Sleps woman catch saw tallest seeping oye sleep seen tall amped ashe sec jump women Sept Jumps amping We sald all gare that The physi! nondomns se relstions of the eee ep spn. spt Steer tc ctche, etc, cught caret The physi wordforne she exeme Jp, jam, jonped. ning iM ‘al alert ran bo, bye mv 212 Wonkonm [As we bave just seen above, sometimes, when we ue the em “word i fot the abst robo) em with common core of meaning, the KEteme ht we want refer to. Rater we aye the tem wonder to putiodar physical elation of iat acne i spech oc wing that 5,4 parseularwordfrm, Ths, e can refer to see, ese, so and “bea ve dileent words, In this sen, the ferent osrenses ‘one of these word-foms would count a thee words, We ca lo say that the word ir ave hn tice tr ad the word Ton eight And we wot counting dhe member of wor ina psi, me woul daly count Ste ae, ig 0 ane five dierent words belousng Lo Re as ene) 2.13 The Grammatical Word The wor! can also be sen as representation ofa lereme that is associated with eeain morpharytactc properties (partly morphological and uty mtatie proper) such a6 noun, active, verb, tense, ender, ‘mber, ee Wes we he em grammatal word offer tea ‘Poercise Siow why ew should be repre representing two distin! grammatical de the win Ry. Usually Feu the bread onthe abe 1. esterday Tet the Bre nthe sh The same word orm ct, beloging to the verbal lexeme CU, can represt| to diferent rammatial words. In [23a eu reprsents the grammatical Wor lye ep ot pr apa Ra the reset tense, nowt In singular form’ of fe verb Cor. But im [3] st repesets the [mati word eon poy wich eles the pas tate of CUT. Tee he two aie words elo y the word-form cu which te menoned above, teeisa thd one which you can obese in Jane It on fer nger This grammatical wor ya gry BeOS {i sopaate seme CUT he noun, Obviously, eT he nou rated a cians Yo CU the ver. Hlometer, CUT, the noun, is «separate seme Thom eur. the ye, baaueit bloat diferent wordt (ee Section Vsti The natare ofthe grammatical word is mporant inthe discussion of ‘he tclaonship Dewan words and setener and the boundary between 12 MORPHEMES: THE SMALLEST UNITS OF MEANING Morphology isthe say of wordstructre. The clam that words have Airtore might some as arse becouse oemally speakers thik of‘words a invisible units of meaning. This is probably du othe at that many words are morphologkaly simple. For example th fre. lephant, ‘at oon, fs, mou, et, came pment 2, vid up) io Scaler ani that are thet sein is spss to sy what he “tute pc f asp sre at fre tana But very many Engh words are morphological complex. They canbe broken dow into stiller units tat are meaning. This steve of wort, like deter and fon, formance, whore desk reer to one pice {urine ad boo sft to one lem of footwenr, we in bth as th serves the gemma function of iadiating pelt "The feat morpeme is wed to refer tothe sale, indivisible units of semantic content o grammatical Santon Irom which words are mie wp By dfmton, 1 norpheme eunno! be decomposed into smaller ute Wich ‘te cer ncaniagil by females or mak t gamit ncn Uke finglar or pal outber inthe noun, I we dived up the word fo (8) (hich contains just ose mopheme) into, sy, (and ft wuld be Impose to say hat each ofthe sounds] and mene by it, sce Sounds in themes donot have ani. Ho do we how when to cogs a sige sound or «group of sounds as tepesnting morphs? Wheto area sound oe sing of sounds isto be regarded as « manifestation ofa morpheme depends on the word in ‘whch it appears. So, wile wn represents a negative merpbeme ad has 3 ‘meaning th can roughly be plod at Words sh a ut! ad ‘ry, has mo claim fo meepeme status when i ccc in wl o a ier, sine in these later wore it docs not hive any Went gras rnc or seman vale, Beni -le andr on tei no not mesh Snsthine(Morpbemes il be separated with hypben nthe exp) Tego provide auf analogy! morphomes canbe compared to pers of {Lego that ca be wed aga and aguas buling Hoes o frm ieeat ‘words, Resarent prs of werd that have the same meuning ie lata {and recognised as manifestations ofthe se morpheme. Ths the negative ‘orphan oscars ina indefinitely age mune of words, bees hose Tied above. We hi iin sgl, wap. at nce over, eeureoe ina ive number of Words is not ab eset propery’ of morphs. Sometines « morpeme may be retried To ree vey few word. This i tre ofthe morpheme “bm, mssnine ‘conn, Se dig, which oun in words ike marordom, Kingda, chen ‘te (Gloss, bare and elsoubete ia the book, are based on defaitons athe (Oxford Bgl Dietary OED.) Tefus ben aged tha, nan extreme cas, a morpheme may osc in sinale word. Lightner (1975) as cles thatthe mere ric mcaing “Miscou only found in the word Pho. Bu ths clay is dpe ‘by Bauer (193) who agg Instead tht ethape re snot dint Monphemes: The Smal Unt of Meng 2 imorpeme and that ishopie shoul! be titel in the dictionary’ 3s an namapsable word. We mil leave this controversy at that and instead ee Now morpane re ienifed ia less problematic cise. {ere [iste other words that contain ech morpheme represented belo. Ral a aeallomorpbs ofthe fist-pson plral morpho. (For simpli ake for our present purposes we are resurdingst-person pra singe ‘amino some) On heme oa om ‘together as allomorphs of the past tease morpheme in English, — Moree: The Smalest Uni of Meaning 2 “The lationship betwen morphemes allomorphs and morphs can be ‘epresnte sing diagram ite fllowing wa" aM a, Beat Moepeme te "ps ee lor alonorpalooryh smote morph orp ny be tagonds —Momhome ——- Morpurme—-Morpheme ee ssp pra!” “Tutu allogorphaloiorph—allomoph ——alloorph morph morph morph op fat Jey jnabey ‘We can ay tha () il and) ae Engh morphs and Gi we can {owp ll these thee morph topiher at lomarps of the past tone Irorpleme, Likewise in Laganda we can sy Aa () fan abo tte movphs an furthermore; (i) and wae allorphs ofthe same Inorpheme sae they represent the sate superordnat coeep, the mor heme test person pl "Te eal echnigus wad in he etfction of morpemes is Based on tne oton of isbn, tat the teal se of conten Jn whch fur Fnguste or oovtrs, Wels a set of morphs as lomorphs [rik sare morpeme they ae complementary detbton. Merpls are Si ben complementary astibuton i) they represent the sme caning or serve the sume gummi fanction: and (they are never Found tcl contests, So, he tree morphs a which Fepmsent the Engl regular past tense marpee are in somplementary bution. Eachmomph i rosicted to the eontexts specie n [29 ke they are allots of the sme toepheme. The same ansss ipl lo to Uganda and Both morphs mean we! and they arin mplmenary dsibin, Tir css before consonants and before Sowels Te as thtfoeaiomorpi af he s-person plural morph. Morphol hy a invariant form (fe ad oo) ar a C0 havea sng lmnop (ct Mathews, 197)2 tradi to Word Scans Morpheus: The Sas Unit of Meaning » ere {Lets pow examine some English words, fcusiag onthe pronunciation of| the underined pat of ech nor, which represents he nepative morphic fp Tht morpho cin soupy be plored os nat 2], imposible [em'posb impatient frxst) Jobless) intolemble (ilo Ines lesen in [mek iseganee _fwebns) incomplete [inka ioompanble inkamyca) ingrade gue (0) Lent the bomorps ofthis negate morphens {i Wate statement aecountiog forthe dribtion ofeach lomoerh, Hopfily you have nated the following llroeps of the morpheme (a aa “The section ofthe allomorph that is wc in «partic context isnot random. th [212] the nso! consonant in the varus allomorpe of the ‘morphene iti pronounced ina variety of ways, pending 0 the ate ofthe sound that immediatly follows, To predit the allomorph tat Selved in each cas, rue [2.13] eg onsonant (es BB fm) as i [rmjposibe, mitt minor, set efor the velar consonants kj he spelt with) and {a} asin pkomplinc, (merce © soket fin] elsowhere, thai before an alselar consonant ike Th d.s.2 a} asim fnfolerabe, bangle ad fleet oF bores vowel a in nee, eens “The thee allomorpi [ol] and ofthe morpheme bare incom: Pementary dsuibuton. This ans that sletng one pretads sles the others No two of hem sun soc in Went none “This example artes what se 3 very common state of aus. Ia orpome her several allstrps, the choke a allomorph usd ih a ven Context normaly plowlopealy condoms. This means eh the alo- ‘aoe select to reprset te morpteme in panics comet i on6 tose phonologeal properties are sma to thse of sounds Tound i a Iighbourngalomorp of ame other morpheme “The phonslogralrxmblans betwen the nasal nthe pref and the is ‘consonant representing th morpheme Before which its ples i doe 19 ‘sunation, The pronunciation othe naa th pref iat 1 atch the pace of arcunton ofthe st consonant peeing he ext mar phere Ths, n[2-12 he abil consonant ry occurs ia before a aba onan, the shcolar consonant [a in i} oscuts tore alvelar con- ‘nants ad he ver consonant [in] osc efore velar consonant ‘ch es, the to consonant np shang the sme pac of ateulaton “Tis example ao ister nother poi, namely that spling is sey poor guide to pronunciation in Engl and may other ngage ‘Where the pont at issue would otherwise be obsured by the standard ‘ahograph, phonetic or phonemic eansiptin wl be wed Tn the ight of ths neato, let ws eter wo the carer example of the sulomorpis of the Ena sein pst tense monpteme io (23). Cleary, {he dhrhato of sllmorphs is phealogcalycondiiona chosen flr the aol stp fe ad ith Beng ised separate the lielur stop ofthe suis fom the fal eon top of te verb to wc ‘attached: ied /- chosen aller void segments other tha abd ‘ost tr ehoson afer vous consonants tbe ha ‘So fall the examples of soph tht we hae seen hae nvote only ‘vols ad consonants. Bu, a the examples om Luganda in (2.14) show. ‘norphenes may also be sigualled by tne, that the contrasive we of sate pte ef. Hyman, 1975; Kalama, 198; Pie, 1948) ta] a. Sak hunger’ nga smviél “sweeper més "moon tii conked”= bath an! tinpd “hewnes’——Bagayd. ‘envy shi “she frie? “one who ie ‘imi “She ris! stmt “one wh reads ‘pti “She chase’ igo “one who chases [Noe ttrpet the tone dais follows" High tone (N° = Low fone L) and Falling tone).» arucion to Wond-Srasare 1m (214, tonal difereoss are wed to diinguish lei tans. The ‘word-fom are deat i al especs cep tone Th 2-14b] onthe other and, one i sed to signal grammatical distinctions. LHLH cocresponds 1b LAH inthe fs yer, while inthe ast wo, LAH corresponds to TELE: Inch cas he fit patter repost hed person maim lie ‘roel tome form ofthe verb nd the scondpaltern repeats tele She form 122.3 Grammatical Condtioing, Lexical Coeioning and Sappltion We have sen inthe ast seton hat the tribution fllomorphs i uy subset to phonologkal condoning However, somtimes phonologl Factors lay no role nthe sketion f allomorp, Lead the choke of sllomorpi maybe gammatalycodaned, tat it may be dependent ‘a the presence of particular gramimateal element A special allemorph riny be reghtd in given grammatical conte ough here might not be ny sned phonologel reason for selection. For examples (1S the Presence ofthe pat tense orphan in the majoiy of cases has no eet Sn the selection of the allomerph that represents the verb el. But a5 [2155] and [216 show, in certain verb the presence of the past tense Porpene requires the sletion os opal allamorph a he ve pas wkd suped C swept pep) swept sent « shook —fok/ took ok) 1 BLAS, the ehoie of allomorph is grammatically congtioned, The presence a he pst ese morphee determines the choice ofthe wep and Fswep)allomorpis in verbs that belong to this group. For the webs in [2.15 the pan tee tate the ec f the allomarphs ook and slo he vert fate ad shat expect Inter css, he choice ofthe allomorph may be eel conned, ‘hats use ofa partularallomorph may Be egatory fa certain nod present. We can ss this inthe realsation of plural in Enlsh ‘Nona the pal morphume is realised by 2 phosologkally cond ional llomerp whose dstbuton stated in [216 Morphemess The Suaet Unit of Meaning u B16), see allonocph fz) Ifa noun ends in an abeolar or veo Palit thle (ea consonant ith sharp, hissing Sound ee. /825 1083) amples aes acer fier Badger beaches Jv) re) a) badge) rt) sot allomorph a a noun ends ia a nonstident voles fonsonal fe any cao ofthe sounds jp Fh amples: caps leks carts laughs mots ‘isp! ikay rts) ody kt allomorph a) olewbers (le, ifthe noun ends in @ ‘oe non-trden serment,incaling al vowels and /b dQ 0 mantre i) Examples: body mgr rome keys shoes Tose raga) rama) Ska (Can yo expat why the en [2.16] fl to account for the easton of the pliral morphome the word axe? No dou, you to have failed to tnd plausible explanation, ad with ood ean, Thee ae ests whee, fro appre reson, the eau rae {in oesplcably fale o apply. The paral of a i mot "escrbut ser, bough words tht ene with x te he expected i plural llomsoryh (ck fsa faxes and bok) basen) The chic ofthe allomorph “ni Ici contoned. Is dependent on theprseneo th specie noun o= inal hee Tew morphs whose allmorpbs show no pont simian A case example ofthis x provided by the forms sed eter Untick oh contain the leteme goo dept the it that they do not have ‘hen single snd in common. (Nowe hat Bere consists ofthe lereme [food andthe compart er) Whar allomorphs of « poreme are hontai unrelated, we spk o applet, yer Te pr dan Petter snot unique in English, Find one ber example of ‘Other example of spplton in Engin octade bd ~ worse (no hae sos won! hot aol)2 troduction to Word-Srctre 224 Uneyng Representations Abowe we have distinguished between, on the one hand, spur, ru soveredplowloleal alternative (s sition whe the choke beeen Sarat allomorphs is eulted is gute predable ways by the phono” ogi ropes ofthe diflerest morphs tat scce near each othe oe Seaton 2.22) and, onthe ote hand, cases of suplton whee there phonological attrary aeration inthe ralstion of morphs (ce Section 2.23), This is standard in generative phonology (e,Chomsly od Halle 1968, Kenstowice and Kibet, 197% Anderson, 1974), Merely tsingallomorphs does noc allow eitingish between ses alternation ike good ~ ber) abd rele aleratons ke that Shove bythe mgitive pet or by theresa pal afl The aie ‘re gener and wil normaly apply to any form with the rkvant ‘honologa! proper, uns it is speically exempted, Thue the veue ral re ia [216] above wed to attach) at virtual any own that ends isthe appropriate wound. By sot. «leo sappeion Keil condoning only appli if form i expressly marked’ being subject it Ths, for example, oa English ses only goodie sabet othe suppeive rule tat ges Derby othe comparative ad ony at Subst 0 the lexealy eonctoned role that sli on to yo the purl ve, Siilaly «grammatical condoned rls wil nly be tggered i the appropeate grammatal condoning factor is pent. For example the allnorph slp of the merpheme slp only eo-ocurs with the pst tense (or the past priiple) morpbeme. I cannot he sled to co-tcat vith the present ese: sep + (est) elses slept aot pt) while Sap + [preseal] ge lerpalep (00 kp) “To bring out the distinction between zeslar phonolal alteration, hich phonetically motivate, and other kids of morphological alles tion that ack phonetic bas igus posits sage dering repre ‘scatationo ase form fom whic he vrioa llomopis (or aller tleratve phones ranauon) of » morphea died by appiing ‘near more phonologia rus. Thestapes whch a form gs through when its being converted from an underlying representation to pone pe sentation constiut deraton, ora conctes example, et us consider agin the representation of then ‘morphs [2.13} whic rept blow a [21] fox convenience [BATL a select (mn) before a lai consonant (es mB fm) ain [impossibe, fxpatient, [movable > seat i] before the velar consonants] here spelt with) and [alsin ieomplance,bakompatil, estiude Morphemes: The Sma Unit of Meeing 3 (© sket[n ebewhere, that is, Heore an alveolar consonant He (4.2, ah as in molrabl,ffanpbie and feldeent or afore a vowel asin nat, nega, Thc wit point to poe is ha the tice pts of the ren (2.17) are not indspenden of eachother. By msking thee spare stamens we have Ine a genrasation. A siperion sition would eto resiate UT os [BAN The revised statement in wich ne pont snl undeling repre Seration from which the thee alloworph ae derived, capure he Tat thatthe alteration in the reason of ths allo i toast Factor namely stimation, [IN] The nasal seaising the morpheme inf anst appear in the hone representation as asl consonant thet sare the lac of Scution of the inl comsonant of the form f0 whi i stitched Sut how can we be certain hat the base form rater than aor We have sen tht the nasal assimilates to the place of atelton of he consonant hat fallow Theft that whet a ve! follows we sil fod [ind appeatng as in dn] dude and (neisb nal Yr) ‘revealing From a phesetc pin of view vows do a have definite paces ‘fata, oly consonants do. So,a consonant canto asin othe lac of articulation oa vowel. The eccrrenee of more vows st “ue te place asian, Besides the volar asl found ropes of ‘whether the vowel that follows mae inthe font ofthe mouth he. oF inthe back tke]. So, the iauence f the sowed cannot he spon for the choke of br} 'Asinpl solton i o assume that i]s the dat form, that the form sete uns there are expt insracons to do otherwise we ost hs form as the underyng representation, we do ot acd to change {Chetore vowels or before aolar eoeonants We only noe to change before nonalvole consonants. If bowerer we ponte fm) off a the ndeyng representation, we wold ed rue 0 oc them whe they iar nt only before nonabil and notre consonants especie bts before vowels If vo anaes can oth acouat propery for he lies, th ana that provides a simpler solton pelted Thi ase Principle of sence fen refered to as Oecem (or Ockham) Razor) ‘Obviously inthis case isthe analiin 2.18] (thf athe ase frm) ‘Phonnogicaly cons morphological ateriion tend to be very fener Ofen alemorphs representing eilfrent morphencs wil dps) ‘he wine phovologcl alternations If they oar Io sina phonobgisau Intoaction to WondSrucare ‘nvronments, Ths, fr example, the voice assimilation procs diplayed bythe pra stein [216] snot unigue to Wat morphere. The ‘hsd-person singular potent tens fin verbs shows eal he sume sitertins as You ean se 119) afer sibians eg. ash ~ wa wakes) (2) ater voiced Segments othr Than sins (ea) wead~ fri rads) alter voles consonants other than siblans (eg d5ampy sony 005 nes) ‘Te same ele apps gente: 20) 2 afer simi (en) Li ~ ha) Li's) a) ater voiced samen othe than sans een) Jane~ Hg dane) (safer oiclss consonants othr than sibilonts (eg, mk Mises Mk) wemake tee separate statements on forthe plural, nother forth tid Person singular and atid oe a he geiiewe mis the gern hac sian sic ares in voiing with the ast spon of the Frm to ‘which tis tached earls ofthe rpms th nul represen "This generation i captured i we post ist One Underingrepe- sentation (or tase form) or any sian sf, and that ‘underyng Fepresoiation i converted ite ilferent phonetic representations Wy he Gatormad phonological rls how: 221) a. The underlying representation of ay iat sui). eis ele @ (2/ater aol andsveopaatl silts (2 5 5) (G) (2 alter vod segmeets pier than sibians (eg vO tnd volo consonant Ie orm 9) (iy x) ater soils consonants. other than silats Uptktoy “The statement in 22] shows thatthe aerntion in gueton it not te itosgnratie property of any obe morphemne, bu rath 4 genera pono Topealprocn inthe aneuage. The ters morpophoneni i Asericay linguists) and morpo(phowlog (European ngs) ae use to reer to rules of thie ind that acount forthe reaation of phonology onions allamorphs of morpheres. Te ule fr the relation ot) 22.18] another example of morpophonei re The Naar of Merphemes 3 23 THE NATURE OF MORPHEMES ‘Words canbe divided into segment of sound. Thus, th word ook can be vied into the segments 7,0, Ki. Tadd, the division of wor ito phonemes forms the bass faphattic mig sens ke tha of Engl But it sao posible, and satura, to divide mors into syle: Fer insane, Japanese uses 80 distin! symbols to reeset 50 of the sabe ‘pes fouad i he lngung So itis important to avoid confusing morpemes with las. Sables are proupings of sounds for he prposs of ataaton, while morpheme ne the smallst unt of meaning or prompt fanction, A fen eels should euiy the distinction. Om the one han, the words sof oo fy Stor ln conan two sylable cach whl cumere ke mora) and ‘oon hu: gan) contain the sabes cach (lates are separated wlth space) However, all thee words have only one morphs cach On the other hand, the word foots bole) his oe slate, Wat > moremes. They ave the morpbeme book bak andthe fa | which "epresets the plural morpeme although it isnot aya nt om ie In ac. in the phrase ihe Hook” coer, Boos contin tree morphs tk, PLURAL and 3 POSISSVE, We wil act the mets ofthe ‘vo 1) morphs a phevomenon known os Iplalgy, in reer da Section 36 When we divide a word into morphemes, we foes on strings of sound that ae meunigfl gales of whether thy consti labes the phonologal level. question that Turks inthe background concent ‘he prise nature ofthe relaomstipbstwsen stings of sounds and he ‘meanings that they represen. This ithe question to which me now tr The deus that flows draws on Mattos (1979, ‘A ist, might som reasonable fo astute that thefelatonship between ‘womphames and ring of phoneme, which ar ideal as morph one ‘tcomposion In that cst, we could my thatthe moptee Bot, Bok) made wp ofthe phonemes) and ‘As we wil ee ina moment, n aproach which asm that sorghemes are made up of phonemes leads to 8 theorist cules. Is Peter able Io view ores as heing represented of eealied or massed by ‘om His unsound to assupe that orphan re etaly composed of (sequences of) phonemes bocise this wowld suggest ta the mening ht morpheme iy 4 function of i phovenic: smpostion. Ik nak since phonemes in therseles cannot have meanings Fr example the sume ‘ioneme 9 (peter) ean represent ether the comparative degre of ‘jets asin kinder and the o-ring li rt fn worker which is formed fom the verb werk or cum be part fa word thou, crib meaning Fis own, 8 wer, Clay the morphs et "han morphs that are made yp of (equens of) phonemes. Pose36 Inirction to WondSiracore he Naar of Morphemes ED ‘elatonships tetneen morpbemes and (sequences of phonemes can then be ‘mari in hi aon ‘As There may be @ one-to-one cordaton tetwsen moephames nd morph (which sre made up of individual phoneme sine of ‘honemes). For instances French, the word ea (vat) as ome ‘morphome whichis teased by 2 momph tat is composed of jst ‘one phoneme, namely) Thi isthe sips station, [R, The relationship teen sound and meaning in language is for the most part arbitary. tht is To sy there iso. 00d reason why a Pact sound of string of towns ha 9 partclar meaning. Gem this several ileret pugs of sounds wih meaning are posible (As we sw above wits regard 1 9 sale phonoloia form fay be usd to represent deen! morpemes. Now we il ‘atin this pin in moce dtl Conder he phonological fr ‘a which happens fo avs thee onhograph epreentaions, tach one of whch represents diferent mrpheme. Ako think ‘out he poeolgi frm ru wich has four apelings wich ‘sprsent four separate morphemes: 223} a sight ste te 1 daht wre wright ite What we ee he re homophones, hits, Frms which sound the same bet cfr in ther meaning or grammatial function, From ‘orp; aay the fra st) ad eat The two mor ep ‘ent thre and four morpemes epentivly, but writen English tesa ferent form to repent cach morphea exch eae. Homoptiony sete common in anguape-Pans depend on ‘And munyoolts de othe fact that mop ie deck ad end represet mote than ne morph (iy The comers i also common. A singe morpheme may be Sanied by avait of phonological representation. We have trendy seen hi the cae f the purl morpeme, whichis ‘he the allomorp [2 and (os [216 ove). The ame ‘ppl othe negate pred (which Hs the allorph [en and [26 [217] abow) (i) The same sting of sounds may cumulasvely represent several orphans Theo nding in Eas verbs (eas) tals three morphemes simultaneous, namely thie person, pst tense and singular number. If morpemcsconstd of morphs this would not be posible. A separate orp would be need « ‘morphemes ae as oppoed to morphs. Morpheres theres fre aot composed af sous Du tbe) are represented b) moepS ‘hac are made up of sounds. The term portmantens moe i sed to ree 1 ess ke the abone hee eng morph silliness represents bande af several ferent gamma eames ‘Morphemes are to morpis what lxemes are to wor-foms. -Morphemes sa lexmcs ae the ata ents found inthe leioon whe morps and word-forms are the physical ete ound in spec oc writing In aion 0 direst morphemes bung epesnted by the same sory we sin aso havea stuatlon where diferent grammatical trond ar represented by these wor forms, This cl meet dit sa seal of metalation. The sae form i wed to epreset ‘dsinct mophologieal onceps. Ths, regu vers, he sane Word orm reprosents two diet grammatical words for example, wale [pat athe ao walked) vs walk + pas patil] waked Gs nT tv walked) Trl verb Iie se and take exit Doyen, ‘They have disinet past tense and post partcpe forms ee + a (a) an ae + pat] (ok ee (ps arp) (en) and ake + Inn parti) ake. Finally, an approach that asumes 2 oneo-one comespondenee etweesmorphemes and mogpis encounters dices when thee is Simply no match betveen morpheme and morph. Thee are wo so ‘Sicunsancesin which tiv may bape: (© The sumber of morphemes present may excad the numberof morphs svalable to represent than This happess when a frammatcal contrast winch i marked overtly by a morph in foe word: snot avery naked in others, This, fr example, we ow hatin English ithe adver yesterday ota phrase ike fst ‘vs i found ina senienes, the ver it tat Sntece must bea {he est tengebecure hnt the orn fthe er that i rouied Swhonever ayer dante an event ation state oF pres tht Fappened orto the moment of speaking or witng. AS ris sha verb wllend ine 1223) Last wes the farmer sowed the corn psec Yesterday Sane pine the rot Inthe ight o he ast roma, explain how the past tens is mark in he folowing.» Introduction to Word Siactare (2M) Last wek J cut te ass. [pat thos earations nthe vase yesterdny. Yesterday they sh the actors dows ‘The mob hic hin lat wosk ‘We know that ct pu st adit are every bi pst sone and painted in [2.23 beans only verbs in he pst tense cam oeur fogeter wit ested oF ks weok in sentence “The pst ese mores, which represented by in 2.25, is eased by sere lomo in [224 In oer words, we ct Infriom the struc patteras ofthe laguage that the ers In the pas ens alflough nothing about the shape ofthe woed ‘tert how hi Iw allow oust post aloo, ‘he asmmpion that morghemes ‘conse? Of phoneme mst be ‘efeted, Instead, we hall rept morhemes ax sirac entes that are represented’ by mogps. In spsch the moepis are ampose of phones but the motpkemes themes fe at {Sera the dcuon of constr in Sston 35) The comers san acre the nmaber of morphs that cm be inated may exceed the cumber of morphemes repeated Im other words there may be sup word-buldig eet ‘which doesnot realise any morphemne. Such un sleet i ome- "ics al an empty or ee Describe in det ow the acive in [2.25] and [225 te devved rom pas) Noun Adietve 4 icine foedisn) edna desma person /pusan/)—penooal/pusaaly fhe fra) tial feral bese seosv-al jaja fact Feccaal ek In [2.2] the adjectives ave forme ply by adding the a ‘0 pouns. Io [2.250 however, there isan empty morph (iu) that doesnot represeat any morpeme tat sere immed asl before al "Empry mar ap fertinate choice of Termiaelogy abd we wil yt avoid using it where pombe Sumnary » ‘An alternative view ofthe station in [2.20] to consider the tase tocontan the wea aarp inthe case of both ena tnd fecha This would expan the existence of fu, 20 aly i entation ad acu T's snorptompl by vine of repeating some mor sinc a splos wording element that does oe represent ny morpome shold ao be regarded as a morph Hence, the Sppropeiscns of the more neal frm formate fr Flee {any mord-blig clement. Most formats are morphs: hey ‘pret mnombemes And some sre no They are the so-lled “empty morphs 2.4 SUMMARY The chapter opened with x dacusion of tbe nature of the mond (Se vion 21), We disinguhed between eres, word-forms and grammatical trons eens are atsrat Sctonary words ke the verbsing- A leneme ‘steed ty ene or move wordforms, Word forms are contte words that ‘secur in spech and wing. for example, in ngs san and sume. Wealso Sth the word ean be Wewel 36 2 Ltome associat wit 8st of ‘morpho-syntactie properties, Tor example. Sa pn oy si Tn thin case, we ae Woking at a grammatical word “The nent ston ede the sopmentation af words ato he smallest stat unit of meaning or prammaticl funtion. These units ae called thors. We saw tht the ass of Words into morphemnes begins ‘vith the contrasting of pais of uances tht are paral ferent in Shand sod meuning. Woreforms ave spmenied ato morphs, which are Ieee physi, wordsforming shanks. Any morphs That represent he ‘ine meaning ef grouped together as allomesphs of tat morphem. ‘Meaning plays tle inti, bat the mai pine wed tht of di Ibu. Mrs ae lds allomorgh of the sae morpeme hey tren complementary distin, tat i they are eels ofthe “ive orphome in dierent comer (Somaimes a morpeme his. Sng snp) Novy, the ditetuton of allomorphs i phonoogcally conditions The wtnhip between llores a ponete motivation. A singe taney Cs) Frm posse the poet presentation of te Stvo alamoph dered from ting phonological oe, hut sometimes allmorphs ay te grammatical condoned or even lexical eontonaly that prtulralomorh sete either a portico pronation or parce word preset ‘Ocasonaly ther ssppltion, which ean hat an temorph ears mo phenets miuriy to otberallmorpts ofthe sune morphs« Iocton to WordSerucare “The lant section dealt withthe tltonthip between moroboloical and phosolopeal representations. I nas esalhed st the eaonship Tetween morphenes onthe ons han and momhs on the other is one OF representation (or easton) rather than compoiion. FURTHER READING ‘Rue, Laue, 2) famed Lie Mortaoy, 2nd ta (Edinburgh abrh Univer Pre) Mathews F191) Mrpie (Cambede: Cambs Uses Pres EXERCISES |. Dine and sive one sh xample ofeach of the ong fi) exeme; grammatical word; wordorm {iy morphome: morph; allomomh; portmantsu morph; suppltion zero morph. (2) What isthe allomorph ofthe plural morpheme hs sarin ech soup of words flow? (©) Esplia whether the choice of llomorp is phonology, eam rately or leraly condoned (0, agenda data strata media desiderata (i) small radi fag ars 3. Sty he following data atl answer the gustons that lo luke wind report iteast uncover recover ‘omble rewrite lock landless dint rdrmw ‘cemoak dinllow peoniess hippy mpd ceoach (3) What isthe meaning ofthe mosphemes represented in wring hyo, asi, ra es? (b) Are any of te masings you engi only of histor! intrest? 4) Comment on cases of homopton wre tps morph represents ror than one morph 44 (@) Distinguish between phonological condoning and grammatical ontoniag of alemorph (©) Gite one esh empleo each takes from any langue that you “a Deseribe the tonal patterns found in the Reale data tow (om (omca-van Sehendel, 197. (by What ae the Fancions of tne in thse examples? fylx "young mae came! ayieke "young female camel elem ran hater hsp Kelen “bg male gost? KEN “bg female gat ‘ifn anal kept for dfn “asset for slaughtering slaughtering it ‘Gomes’ nif “a domes nimat Nove on fone-marking: "aks High one; ° Low toe; ~ Mid one; NTHigh-Low) Falling tone Ttoteaing your answer with examples of your own, const an fraument for setting wp undeing representations or base Tors. {ead Seton 224 gph before attempting this question)3 Types of Morphemes 3. ROOTS, AFFIXES, STEMS AND BASES In the lst chapter we saw cat words have ternal strstr. This chapter introduces ou to awie range of wor uns ements walt crest tt Structure We will art hy considering roots sd flies RL Roots A oot isthe ireduible core of a word, with absttely noting ee tached tot isthe pst thi salva be pene, posi with ome ‘olson, inthe various mien olan, Fr expe wal {oot aud appear inthe tof wordforns tht lata the ene ‘Wk sich 25 wal, wats, wolking and walked. “The only situation where the aot tre when suppletion takes place (aoe Seton 223) In that ese Word-orms that prot the same ‘morpheme do aot share 3 common root moyphene. Tht, sth bot {he wordfonms good and fee rele the lxeme 0, ony aod i Phonetlly sia to good "Many words conan 100 stnding 0 ts ova. Roots which are cpuble of standing depends are al fre rps, or exazple Bt} Face worptumes fan book tea sweet cook fet" very aardvark pain” alk Single words tik those in [3.1 ae the minimal fie morphs capable of occurring isolation. "The fee merphemes it [3.1] are ako examples of leat merphmes. “Theyre noun, active, verbs prepositions a adverb Such morphemes ary mot of he 'emanticonent outers —louly defied cone tions bike relring to india (eg the wouns Jol, mbes ing properties (eg the adjectives Kn, lve) desing ation, proses for staes fg, the trbs we, res) le, expressing retions fe te ‘poston in on ander) and! describing crustaaces ike manne Tan rey ui ‘Anoiher da of fr morphemes ar fmetlan words. These ier fom ‘esical morphemes ia that wile the lesa morphemes cry moe of the “semantic conten the function words maily (But not excave) sgn grammatical infomation ological relatios ia a seen. Type fasion word fc the allowing Roots Apis, Stems and Bases a 182} Funetion words Mec ade Semonsrtve: thi that thes hose Donowns Kyou we they them: my your is ors who whom onjncion: Sad ye i but however ore Distinguishing etwoun exe! so grammatial morphs & normaly th tefl and stiahforeard. However, there are cases where thi ‘inne is lured. This becuse here se fee morpbemes Ge simp ‘Tord that donot tay ft iter stegory. For example, a conjune Sow tik rng sgl logis eaonship and, tthe sme tine, spp ts'have conslerhly more “aesipive seman coma” has, 9, the srl he ‘Whi oly roots cam te fee mompames, not all oot ae fee, Many soot a incapable of eccating insolation. They always coc wth some whee wor-buling seme stachd to them. Suck toos ae clled bound ‘turphmesExarpis of bound morpho ae given below HO) a. mit asin permit remit commit, adit “eve as in pete eve, consis, desire © prokSein predtor, prostory, predation, depedate foe Sein fete, dea, sedentary, sediment "he bound 00s, ee, pred and se co-ccur wit forms ke dere tat, ment wich reer in namerots other words as pfs or sans None of these oot co! oes an independent Word Roots tnd io hve a cove meaning that i sone way mie by the a, bat termining meaning i soetines wick Perhaps ou ae ale to ‘scoune the meaning “pre tat rans trough the oot predate vaous ‘ondsin [3.3 and pete you ar so able to deny the meaning ie ithe forms in [36] whieh contain se "Tho rool Latina a they came no Engh fro Lai (ten via Front) but uness you have sued Lato, ou ate probaly unable to {Sy that mean sed, dante means lake’ without ooking > ‘ni and see nan efymolopical Stonary. In pretty English, sone sites mensings 8 reogaiable These formauvs cannot be signed & To th lst shaper, the morphine was defined asthe smallest wnt of sncaning or eammatia ncn. In he Hiht of he foresing dacsion, the inten on the egurerent thievery morpene mus have a eae, onlant meaning (or grunaillfunclon) scene too song some Ungun. There are merpheres that ack a cer meaning. Trscad. hey sug fe the wor rather thn the momen that Masta be“ Type of Morpames independently meaingol whence iis we. As we sayin Section 22.1 hove the crac thi about morphemes snot thst he a independent ‘menial Duta he ate eceznsble rational nits Hari 1951, ‘At Aronof (107618) pts tw ea acon noes when we ee ‘morph "which can be connote to lagu ety outs that sng ‘What is import notte meaning. bt arian, ‘The reson for teat those recurring portion of words that appear o lack a clr, constant meaning a morphs representing some mony that they Behave in a phonological comustnt way i the langue that isdfereat fom the behaviour of morpholoialy world bt ponoi- ‘aly sini sequences. Take mi. or example. Aron! (1976) points ut that, notwithstanding the tenuous semantic ink between sntas the Line rot vn hey nevertheless share common ferture which not radial from any properties of the phonetic sequen ft AI aac of Latinate smi have the allomorph fn} o os] before the sie tom, ‘ory te ad abl ea 3 cis 184) fou) ofl befor on ra) tfore ary Pentit. peminaoa pervs Sibmit Sibmision —— Fabminste ‘mit admission admisive ‘By contrast, any oer phonetic form fn] does not andsrgo the same ‘honelopes! modicaton before sich sie. Tn hough forms Ike ‘lotion and rrr have aot) phonetic shape preceding the ait or thy alto undergo the rl tat changes tS) I that rele app ‘woul incorety deliver “drmiioryor"vomusay, sinc these phonetic Sequence fm ts hatin (3) prcads thes sory. Clearly the [ot] SSuence rrr and donor sot amon epost the Latte “nt morphome The rule tha sop the slp fut of eis that Contain tat] ool activated where mnt epee the tna root mi ‘What this discuson shows tha even where the semantic ba for recognising a morpheme i shaky, there may well be dba on siderations that may save the day. Oaly the rot =m a the allomerph Imus} Any word form that pay the ft) ~[on alteration in the contexts in (34) contain the root more mi 2 fines An aM i x morphome that only oscars whe attached to some ote Imoypleme or morphemes rch as oot or seh or Be. (THe ler (wo terms are expel im Saton 3.13 blow } Obvouly, by dion aes Rots, Afien, Stams and Bases 4“ rebound morpbemes. No word ay contain only an afi sanding on Atsown Ike "ro oral Or eons air of aes stung together "Three thee base types of aes We wil coer them intr ) Pretces 1 pre ra fis atashed efor rot, stem or bss, ke rama 5) somake unkind indecent feread untidy iceurte (Sates ‘i sana tached aera ot or tm or bse et r, -s fog and ed 16) kindly waiter books walked quickly plysr mats jampsd (id tes SK max inserted inte root sel. Tbs ar common in some ihnguaps, bower inting i are in Englsh. Seat and Taser (197) ‘he thatthe nl inf that ours in English moepolog i whihs Incr before the consonant of he ot now words Lan cg Sia appear tobe an ary Pan The ne wero place of oan ation Th, the rote meaning non oF PON ‘cers without (tfore the in some words containing tha root, fr ‘Kimple intl, edu, conebin and seeds, Ba (nine before tharume ot inn tr words nem cea. an dane This sna et osm hrc rl from Latin Uf, twas a oem Wont ede tai itn a sr il apes i conto Engh bho tuto spec pe, Conse he xa in 7a eh ae lene om “ata and Pato (1987) aad thse (370 taken Geom Bauer (983 IIL 4. Kalamazoo (lace name) —> Kalana-goddamn-z00 instantiate (ers) infukistantoe bb kangaroo — kang-boody-109 Impossible — infin pore In presenta English, nation, aot ofa ax moepeme but ofan entire ‘word (which mays more thane marpeme, eg. Mov ot Jeti ase to modly word. Cuno, this ination is vials4 Type of Morpher Roos, Afi, Stoms and Bases ” restricted to insering expletives into word in expessve language that one would pray not we in pe compan. Farther examination revel that lee fe aden, prosaic conditions on these of sen, and tht willbe dieused in more deal in Section 93.2 Foracase ft, morphological iaation we mut look beyond Engh to‘ language sucha the Native American language Nuueaaulth, where ‘nation wed to sneat he pra in (38) Below Bs) rors hl feats ‘tilda? tn ih hau “nhs finvagsi “mind” tags and Matas) the other tah the hes 1a this case, the nation of paral semployed onde tha the noua ‘o which the ine aac plural for furtherdncusion fhe mechani of ination, so Chapter 8). Such eases ae not uncommon inthe angsges M13 Roots, Stems a Bases “The ste stat pf word tht i neste before any yectonal aes Ge. those alias whose presence i egured by the mnt sch rarer of singular and plural tuner in noun, ens in verb ete) hove fer added nfstion ir deed in Sstion 3. For the moment fe ctples should sufi: Purl Suis In the won-orm cas the pla infctional suit is tached to the simple ten ca which sa tarot, thatthe izeduible core ofthe Word In workers, the sume satectenal-ysulx comes ar sighs ore complet tem owing of th oa wrk pls the sue, hi wed 0 {oem acti nouns from vere with the meacing someone wh does he action designated by the ere. singe. fer, dane. Here work {he oor bat worker these fo whit tach aly, a base any unit whatsoever co which ies of any kind can be ae The aie tached 0 base misy be aetna alles sete for sytctic reasons or dean aiffes which ale the msuning or srammatical ategory ofthe bus (ee Sections 82 and 12). An unadorbod Footie by en bea base sinceit can ve atached tt infistiona aise Ike o form the purl yor deviation alls hie nt tn the noun ay ato the active oy a otber word, al rots are base. Bee recall storms onlin the cones of infin sorphlogs, [eri the inestionl fies, derivation aes, rots, Sates wad str Inthe following 10) fais frogmarched fail bookshops th rindoneseaners Faithsinese hardships Hops, your oion i ike thi: ra Tntion Stems ses sites a fait ef faint bookshop frogmarch = Sean windom-leaner Bookshop ess hard ap window sp indo frindowleaner atin 1s lar om [3.11] that ts poset form a complex word by adding snes oa form containing ore tia on 00. For stance, te adepend nt word rogund march canbe joined together to for the base (sem, 1 te proc) formarcho which thesufl-dmay Be added 0 el [fe Fnrch} ed). Silty, widow and cea ca be joined to Torm te bse [iwindw-{sen[] to ie the devationa sul ver ean be added to Produce ///windofcen| fer) Aud [window -[cten} er] ean seve bea sem to wich the iafecuonal plz ending “is athe 10 give [indole Jes} word ke his, whieh conais mor than ome "hot ical compound word (se Sesion 3 blow and Chapter 13). fal seelea 73 ee tee af aging eoctccs ne wendy ‘ent that appear devoid content, Such empty formaives are Sometnes refer a somewhat inappropriately as enpty morshs, 1m Engh, empty formtives are itrpond howe the 0, ase oF stem aan aft For stance, while the bghly regula plural allomerph “en atlached del to the sem ox to Kn on, the formation of Shubren ican ony be aed le the en as her extended by tacking“ Types of Morphomes “r= chi, ying the bound fore chi. Hence, the ame stm extender For this typeof formate "The woof sem exteners may not he ents arbitrary Tere maybe 8 008 stra reason To he ae of pac stem extenders before ets tier. Tosomecitent current nord Formation alert the history of he Jungusgs, but ths lationship te opaque and inaccsibe fo eames. “The history of stem extender -- may be istrative A small number of nouns Ok English formed thie pal by ang The word chi was ‘inthe singular tnd cle ae paral afr that survived in some onsrvatve North of England dat ac opel cider). But ate, Jost ts alu and en wah added as ne plural ending, ther temaiing only asa stem extender PLP] & Siguur Pad 8 New Sgr Pt Stem etre re not esta to English Haas (1972) deus « mich ‘ote widely ocouring cae of sem extenders in Novel also known Noo), ‘The mott common typeof stem extender in Nootte is single aon- totaled stop o apranttlonsng to the dora as of ther the no Teac (hg) of abled 8" g" 37°. 2") sas (Hass 197283) Some erample fhe ws stemextenders in Neucabl appesrin. 13H Ay Tatf old, double uy! — anf “colapsad™ Vie white {isk ash whe hitf mina tf “glow witha brig Fie i pinloned dows’ iq jammed, packed tight at “CALS break” Gul” amputae get amputat As you can se, the semantic relationship between the basic and extended the Tors No doubt the vagueses is de To thee Into inves the ‘volition of hse fra, 32 INFLECTIONAL AND DERIVATIONAL -MORPHEMES [Ase have aleady hinted, mospemes canbe dived into 40 major functional eaegors, nary erations! momen and. itn Ieflsiona and Derratonal Morphames s rmorphemes, This reacts recognition of two penspal wor-bulling rocmae:lfcton and deraton While all morpologiss ace hi tisinetion in seme for, es nevertheless one of the most contentious In momholopea theory. We wil bre ntrodace you hee tothe exon "thr dnnction, hot posipone deed dicasion wal Chap 1 Tnfconal and drtainal morybeaes form words in dren Wap. Devtionl marphenes form new words ethe (6 by changing te meaning ofthe Das to which they are atcha, for cumple in vs nd (oth are aetves bat wih opposite meas Ings obey vs vey (bth az verbs bu With oppesit meaning). (i by changing the wor-lass tht base belongs, for example he “nin of 4) tothe astives Kind and simple produces the adver indy and sips A 9 le, it posible to deve an adver by ‘ang the sui =) to an adsl base isercee ‘Stal the following data and answer the qusions below. ota) 1 ducked He was shopish two ducks three daclings He ehumoue You ae ducking the sue (0, soy the sti the underined words. To what words do the words o which the suffnes ne edded belong and what wordt (4) For cack sufls dete whether i is ifstion or deivatonal, ‘rey js your dion opal, your ansner is ery else the folowing HK1S1 —Sltic Input Ostpat Remarks a at Vestn: mars pst ese in fckel NN nfletonat it mars pura number (in (tv) deco ad decles) Icons: portmanea morph ‘nathing thefts, present tense tin gua2 pea Murphoes Iocan Drona Morpher s “ng VV insta mara progeine ae coma eniadtriney Saati: add sy Words may have mali af ether wih iret sulias ppeaiag ina ssuence as in [23 or i he sme pei recueing as bel in [3.24 1824) a. the latest reseremke of Bom Gest 1 the prent-arentgreat-aeat grandson ofthe lst Tsar of Russia, ‘What (324) shows ie tha, with a imted number of morpbemes, ‘mombolgil prefsation rls can apply ects im Em oe Seton 13) However, pertermancedificalie in working Out wat excl gre retreat great grandon of rereremmae means do severe) reste the ‘hunosof ich words bing wed But the point athe grammar cannot ‘cle them tl formed, Recarve ras ae one of the eves that make ‘mombology open-ended. They mike posite the eration of now words ‘tn he mame morphmss eng wed Ove and ovr (Secon 4). TReallaching the samme morpheme agin and again is permite, but ‘unusual. What iscommon ismutipleafsation of diferent affine. is sich ation that we wll encetrate on, Wehave lady cen an example i in eotnederortness in 3) rcs “abe the res bave avon and st it 8 many prefs an sixes as you {am Ate to go thvcugh a least four rounds of asation, Hopf, you have come up with something ke this 251 ation tational Tatioabise ‘denatonaise ‘Senatonnliaon (hut thoe i no Penatonalate) "ntdensonatiaton preatdemtionsation COtsere that where several prefs oF sufies occur in a word, thir flee inthe sequence prmaly fg te. whereas theres usually hme scope for teranging word in dierent odes io snes You 24 Yo can play badminon, —b. What I needa sce cup of, Con you pay dminion? A ice cup af eas what Ind here i virtually no possibilty of arranging morphemes within word, ther than in compounding, be dicaned below So, fr exampt, Che Inorphames in denrion-ae ust appear in that onder. Rearanging the jes proces iormod rings ike erate of “senatonde Themain problem and interest as we wllse in Section 621, determining Inc ont of derivation fests several of them occ ina word 3,4 COMPOUNDING [As we ily sa i Seton 31.3, compound ward contains at east to thos that ae oth won t any ete root mores. Paecbe ‘Analye the following compounds ito ther constiuen element: apn, seeker, Kind ert6 Types of Morphemes ‘Wie expect you to have workout an answer clos othe folowing aM] a fey (pote teapots (reeIsendhy, ——[verkandy 1 Ihe ese, Petes kindy tear ly — hindered, Compounding is 2 very important way of adding tothe word sock of English, os we wil See Sometimes {8 bare tows that ae combined ‘compounds, in [27a and sometimes a input base contains aad Formas in (3.27) Notice hat compounds may conan lmens inert tres, wih consequent dilferencr im meaning, fr example, ama ‘hain. We wil deus compounds bey agua in the nee choper and return to them in greater dea ip Chapter 3. 3.5 CONVERSION We ive sen tht complex words maybe formed either by compounding ‘or ty aftuation, or by 4 combination ofthe wo. We ar poig to 5 ‘ov that thera aleroaive wor-formatin sity whch tscommonly ‘ed in English, Words may be Formed wathout modyng the fore ofthe Input word that serves as the base. Thus, Hea san be a bous r e, This ‘ataled comers How do you know whether Asad it non or ve inthe following? A2M1 8, The fad ofthe village schoo as ative ‘The Head of the illage choos haves She will head the village schoo ‘She Rede tha schol is party the morphological streets, and partly he syntactic positon that the word oecpies tat els you wheter ta noun fa web. From Syntactic pont of view, we know tha in [3289 rhe hea «noun phase ‘The key word ina noun phene must bea noun. As hal ecu flowing ‘heand ithe ke word this consrction, hoa, st bes nov. Bat ron |mbrpholbgcal pont of view we cannot ll whether heads noun oP when occurs wth fines, Hower, in the caso head, the presence of thes moypn hat here realises the plral in nouns gine vs Us ce MorphoogctHapooes a By contas a [3280 had mst bea vee. I comes after the asia ‘ver fina sit that styl lle by vers, Inthe second example eat has ataced ot ted morph representing the past tense morphene tht is only found in verbs. Furthemere, fom sya point of ew, me now that se he subject an a soho! te objet. The sentence mst sk bave a verb, The ser occurs Between the subj aod the object, (Th ‘order of sentence constituents in gins ubiet Verb Objet) So headed oust be the ver, since it cccus fete the sibjet andthe abet ‘Cocvertion i ao rtrd to at sre drtntea in te Herta (. Marchand, 196% Adams, 1973) and 5 subsumed under affation, by ‘logy to aero aston in intional morphology (.Svtion 231 Snimed tht wero morphs Leones Inking any overt marking) reused Satine in derivatonal morphology aswell or lstance, the we ead ‘led by sulin a ero morph tote noun ad This dane by aloes {othe deiation over ke vette, om the noun vn whee the ‘vee yert-forming sis ie wed, “The use of eo In derivation morphology is controvesl Since neither ‘ne assnal noun brad, or he dare er had, aa overt si we assume tht a sft takes place het, we eod yp With somewhat itd situation where a zero sic onthe moun sd Yo contr ith & ero sation the derived eb, Te moe prudent to eognise conversion tnt wor-frming mechan nd 0 reset hee ee mop ‘much s posuble Soe Sation 623 fr further discsion 6 MORPHOLOGICAL HAPLOLOGY Stomberger (1981) dscuses the phesomenen of morphologies haloes aplogy refers tothe deletion of one of an identeal pat of cement Inthe cae of mortiolgkalhaploloy, the cements const ete moe belonging To diet morphemes. The example in [29 stale Uhcease ta [3294] we have example of the plural tached 6 sous, [298 Stow the sme nouns wit he posesive marr ad [3.2 illastats the ‘as of ba the lua nd he poses co-ocuring. Foal. [529d] pro- ‘is xampe of iru prs combine withthe pl demonstrating The thot constraint apni the o-ocurenc of the morpemes 29) a the cats the dope the ehurbes ive (age) lusts b. thecatseas the dogs paws the church's windows iweu [age! iss ikeu) (aaah ious 4 the onenscar he mis paws the children's toys pine anal ikon)ss Types of Morphes [Asan be en hy thee examples, the attachment of to ential moephs reprterting distinct morphs, plural and possi, voles the meter ‘fhe phological conieat of theo morphs, resulting ta cera degre ‘tumbgty inthe cases in (3.296) Note thatthe mbit doe not ara the ees in [1294 since the iru pr has tren! ponlogcl ‘shops fom the posse 3.7 MORPHOLOGICAL TYPOLOGY We suggested inthe opening chapter that although languages vary enor owns their rture they nonetbless show surprising snriis, THe ‘ay ofthe sgnicant share rata properties whic languages hye i common the domain of language anversa. Many of the univers ae ‘hsrset principles of Unies Grammar that determine the propeies of Tule that grammar of nda languages may have (eg he Set C3ee Condition dicuwed in Sesion 62, “A ineral port of the study’ of wives io language tbe stay of ferences between languages. Tis might lok ofd to esin with. Bait turns on tht dilerences between the trical pttes ound in erent Tangnages apps to coc nin ily restricted ange. Tete re pra tes in which most dferners between langue oo. s (am liner determine where rams eam go inal whe ving ther plenty of jopion) reset paramos detemiae the structural pateras trom which ‘ferent languages may eect, Stucural pti are not randomly dstibuted. There ae @ number of strong preferred patere thu recur ia language ier lnguge, wile cer pullers are rae, oe non-existent (Greenberg, 1963; Come, 198%; and {spill Chom, 1986) The sto of the ange of pater within which Taneuages may sry i the dorsn of language plo. ‘Our conser in this hook with oth the sales and diferences| tbecwac languages inthe wy in which hey form words On the Bass of ‘spi patterns of wor-ermason, agus ecopnise Sie broad rmrpho- lopeal fps (aml called olan) langues: (i) gptinaing (also cas agate angnoges {iy edn (ao ced sytem Tangosses (is) gla (cmetines called incorporating) anepes: (3) tomate angus. We wil nw comidor the morphological types in tara, starting with ample fanatic morphology fom Mandarin Chines Morphoogea Typology ° Bam a sho me fe OM book bused. “He ought the Book" bt eho yige si kin long he cook dip, onelasiier sh very, dalicious ‘He cookal ish hat as very delicious” [Nate A. is short for pefeive apes”. Indias tha an aeton Is Completed OM ie short for “abe marker Le, the morphsmne that indletes the abet ofthe ver (dat rom Li and Thompson, 1978). ‘As you sa 8 fom the morphem-by-morpheme translation, i Chinese ound mores ae infrequent jx a te EnaSsh ranstion demon ‘ates the sme property for Ena. Men the word reba, uted root morphame or cmmpounds af uch 001 ‘Cinzia example ofan analytic langue, that is language where cach opheme may acer asa word insolation: Word etal never have intstonal aes althoagh ey may ear aspect marking Ths, the objet hur maybe viewed an sdependent word. BY const, in ott linasnge types objet markers re aorta inetonal fits that are part aon or proaou. Ia English the subst ronown he contrat with he ‘et pronoun him in sow Lauro Lao son. The cane om he toi he pronoan marks the ange ramet function. Sul, in Engl, mker of ape and tere re sual inftional io of he ‘ecsch ae a cok (eo) Siar i Chinese n [3.30] he tsps opie reise hy x bond morpame, (Note in passing thar there ae many Chinese words conning more tan one morpho Unoly they are compoands hte she (erly “cew-tongue) ‘gsi Ivo shows Croan plow, pion (et-tuninow) “baa Fats and comeing everunderstnd ) ineliget) ‘Vxencise 1c now tur to ater language, Turkish aM ‘ihe hand? eid in my a ‘hm ny handler ny hand fer hehands’ elerinde in my bande 1H vie the words above into morphs and sign cack morph Io a nora {0 He do the morphs matchup with morphs?