208
THE ORIGIN AND EARLY DEVELOPMENT OF THE CHINESE WRITING SYSTEM By WILLIAM G. BOLTZ AMERICAN ORIENTAL SOCIETY NEW H AVEN, CONNE Cf ICUT

The Origin and Early Developments of the Chinese Writing System

Embed Size (px)

Citation preview

THE ORIGIN ANDEARLY DEVELOPMENT OF THE CHINESE WRITING SYSTEM By WILLIAMG. BOLTZ AMERICANORIENTAL SOCIETY NEWH AVEN,CONNECfICUT COPYRIGHT1994 BYTHE AMERICANORIENTAL SOCIETY Al l Rights Reserved ISBN0-940490-78-1 CONTENTS FIGURES.vi PREFACE.vii PROLEGOMENA1 Introduction3 Chapter 1.Writing in General.16 Definition of Writing ....16 Forerunners of Writing..22 PARTONE: THE SHANG FORMATION29 Chapler 2.Writing in Chinese..31 PictographicOrigins.....31 Logographs and Zodiographs.52 Graphic Multivalence.....59 Determinatives ".......,67 Chapter 3. The Multivalence of Graphs.73 Egyptian .75 Sumerian83 00 PART TWO:THE CH"N-HAN REFORMATION127 Chapter 4.Early Legend and Classical Tradition .129 Early Legend..................129 WenXandTzu ..............138 TheLiu shu1\11and the ShuQwenchich t:r;u.143 Chapter 5.TheImpac( of the Chinese WorldView,156 Or-thographic Standardization...... .....156 Graphic Variation..................158 Why the Chinese Script DidNot Evolveinto an Alphabet.168 GLOSSARY OF TECHNICAL TERMS179 ABBREVIATIONS.184 BIBLIOGRAPHY..185 INDEX OF CHINESE CHARACTERS193 INDEX .................199 FigureI Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure7 Figure8 Figure 9 Figure10 Figure11 Figure12 Figure13 Figure14 Figure15 Figure16 Figure17 Figure18 Figure19 FIGURES Inscribe dturtleplastron Inscribedoxscapula. . Examples of Shang oracle-bone inscription characters with ostensibly recognizablepictographicorigins. Neolithic pottery marks fromPanp'o (S' un ... . .. . Neolithicpottery marks fromPan shan and Mach'ang. Neolithicpo ttery marks from tiu wan .. Neolithicpottery marks fromLi ang chu. Examples of Shang oracle-bone inscription charaClers with whichpo ttery marks aresometimes compared . NeolithicpOllery insigniafromLing yang ho. Partial insignefromCh ' ienchai..... . flu vase withe mblem fromPao t'ou(S' un , Neolithic jades withemblems from Liang chu Examples of earlybronze clannamee mblems Examples of clanna me e mblems withthe ya-cartouche Examples of clan nameemblems with a"dagger-axe" Inotif.,........ .. ........... . Sumerian limestonetablct with clcar zodiographic writing....... . . ....... ,.... . Sumeriantranslucent stone tablet with dear zodiographic writing....... . .. . .. .. ........ . Examples of oracle-bone inscription characters with unidentifiable pictographic origins ..... .. The threestages of the developmeOl of the script.. VI 32 33 34 36 36 36 36 37 45 45 45 45 47 49 50 56 57 58 69 PREFACE MyintentioninwritingthisbookhasbeentoJayout inastraightfor-ward and comprehensible way the facts asI see them surrounding the origin andformationof theChinesescript in . the secondhalf of thesecondmil-lenniumB.C.,andof itsreformationandstandardizationintheeh'in-Han era athousand years later.IndoingthisIhopetodispelsome of the wide-spreadmythsandmisconceptionsaboutthenatureof Chinesecharacters andtorestore adegree of commonsenseand clear-headed sobrietytoour understanding of the formand functionof Chinese writing. Iamabletosay"restore"ratherthanthemorepresumptuous"intro-duce" thanks to the past work of twoeminent scholars. Peter S.DuPonceau (1760-1844) and Peter A.Boodberg (1903-1972).Morethan acentury and ahalf ago DuPonceau,thenPresident of the AmericanPhilosophical Sociw ety inPhiladelphia,set fonhaneloquentlyexpressed andclearlyreasoned "dissertation"ontheChinesesystemofwritingwhereinheshowedthat claimsaboutthe exotic,evenbizarre,natureof theChinesescript,andits ostensible"ideographic"basis,arenaiveanduntenable,andthatChinese writing,likewriting everywhere,issimplya graphic deviceforrepresenting speech(DuPonceau1838).AlmostexactlyahundredyearslaterPeterA. Boodbergreiteratedthesamefundamentalthesis,takingashispointof departurethepropositionthattheChinese indevisingtheir writingsystem followed the same general principles that governed the origin and early evow lutionof allother knownformsof writingintheancient world(Soodberg 1937). Muchof thetheoreticalunderpinning of whatIpresent inthismonow graph,especiallyinpanI,isdirectlytraceabletotheworkofthesetwo scholars.Iwasprivilegedtohavespent virtuallythewholeof my"Berkeley inthe'sixties" decade asastudent. bothundergraduate and graduate, with ProfessorBoodberg,andIfreelyandgladlyacknowledgetheextentto whichmywork here isan outgrowth of that association. The actualdrafting and writing of this study waslargely a"Seattleinthe ' eighties" undertaking. and liketheChinese writing systemitself, had afirst formationand,someyearslater,asubsequentreformation.Whenthese ideas were findingtheir first writtenexpression.I wasvery fortunatetohave hadMs.(nowDr.)YumikoF.Blanford(FukushimaYumikomBbElJ""',],') as mygraduatestudent.Ms.Blanfordtookgreatinterestinthework,and spent many hours of many days discussing, scrutinizing,and criticizing each sectionasitcameroughlywrittenfrommydesk.Manyof theideashere VII viiiPreface havetakenshapeasaresultofthoseexchanges,oftenasadirectconse-quence of her suggestions and advice,including numerous cases where she sawthecorrectphoneticexplanationforanoddgraphicstructuremore quickly and more confidently than [ did. Whenthetime came for the refor-mationof thework,lateinthe'eighties,it wasagainmy verygoodfortune tohavehad another talentedand dedicated graduate student, Ms.Laura E. Hess, who took a sustained interest, again with great enthusiasm and under-standing, in the project. and who helped me rethink the material and revise the presentation in everyrespect fromsimplematters of wording and punc-tuationtomajorconsiderationsof factandinterpretation.Wereitnot for thesetwoassociatesthepresentstudywouldbeverymuchmorewanting thanit is.Ihave.of course,exercisedmyoccasionallyhyocephalictenden-ciesinthefaceof goodadvice,andsoneitherMs.Blanfordnor Ms.Hess bearsanyresponsibilityfortheerrors,confusions,andmisinterpretations that may show up here andthere. Manyothers havehelpedandadvisedme inthelong course of writing and rewritingthis work.Asanyonewhohas forgedabook out of anassem-bly(ordisassembly)of papers,notes, jottings,presentations,andother as-sorted writtenbric-a-brac,ratherthan just writingfromstarttofinishina straight line, wellknows,the sources of inspiration, advice, and constructive criticism,crucialand valuable asthey are,become obscured by thetwistings andturningsthat theendeavor takes asitproceeds alongitspathtowarda finishedwork.Butthevalueof thisobscuredhelpisalwayspreservedand reflectedintheshapeof thefinalproduct,evenif explicitrecallof those innumerableinstancesofwelcomeaidis . not 'possible.So,toallofthe unnamedstudents,colleagues,teachers,mentors,critics,andfriends(not mumallyexclusivecategories,nomattertakeninwhatcombination)[ hereby acknowledgeadeeplyfeltandgenuinelyhelddebt of gratitude,in fullrecognitionthatthemeritsof thiswork,whatevertheymaybe,are muchthe greater thanks tothat help. Some names, of course, have not disappeared from memory, and a good measureof adviceandcriticism,oftenof themost detailed,scholarly,and substantialkind,can,Iamhappytosay,becreditedtoindividualnames andfaces.[cannot begintoenumerate or specifytheparticularpointson whicheachof thefollowingpeoplehashelpedme;Ican only saythat the contributions of eachhavebeen substantial,welcome,and sincerely appre-ciated.Those whoreadpart or allof variousdrafts,or whodiscussedparts of it withmevivavoce,respondingwithawealthof thoughtfulcomments andsuggestions,includeLarryDeVries,DavidN.Keightley,LiLing,Roy Andrew Miller, Jerry Norman, Qiu Xigui, Richard Salomon, Barbara Sands, PaulL-M.Serruys,MichaelShapiro,EdwardL.Shaughnessy,KenTaka-shima,andNormanYoffee.InadditionRobertW.Bagleynot onlytaught memuchaboutShangbronzes,inscriptionsandotherwise,buttookthe timetoread,andmarkwithafinestylist'Shand,severalhundredpagesof Prefaceix myinelegant prose,thussparingmeandthereaderbothmanyinfelicities and awkwardnesses.Paul W.Kroll,East Asia editor of the Journal of the Ameri-canOrientalSociety,andeditorof East Asiancontributionstothe American Orientalseries,hasbeenpatientandtirelessintheproductionofthis monograph.Nottheleastof hiseffortshasbeenthecomputer-generated printing of ,many of the Chinese characters that appear herein.Stuart Aque has helped meimmeasurably withthe computer constructing and generat-ingofanumberoftherestoftheChinesecharacters,particularlythe anomalous ones;andDingXiangWarnerhasbeenof greatassistancein preparingthecorrectedpageproofs.Finally,JudithMageeBoltzputthe fullforceofherconsiderablescholarlyabilitiesintohelpingmework throughmanyproblemsof understandingandpresentation,at everystage of thework,neverfailingtoencouragemeonintheendeavor.Toallof theseindividuals-friends,teachers,students,colleagues,andco-conspira-tors alike-I express mydeep gratitude. TheUniversityof WashingtonGraduate Schoolhonoredmein1985as an Arts &Humanities Research Professor, givingmeone term freeof teach-ing,toworkexclusively onthisbook,andthengrantedmeasizeablesub-ventiontoassistinthispublication.TheChinaProgramoftheJackson SchoolofInternationalStudies,undertheDirectorshipofNicholasR. Lardy,alsograntedmeanequallysizeablesubventiontoassistinpublica-tion.Iam very gratefultoboth. PROLEGOMENA INTRODUCTION In 1838 Peter S. Du Ponceal:l. then Presiden tof the American Philosophical Society in Philadelphia. introduced his own study of the Chinese writing system inthis way: Iendeavour toprove,by the followingdissertation,that theChinese char actersrepresentthewordJof theChineselanguage,andideasonlythrough them.Thelettersof ouralphabetseparatelyrepresentsoundstowhichno meaning isattached,andaretherefore onlythe elements of our graphic sys. tern;but, whencombined toge ther in groups, they represent the words of our languages,andthosewordsrepresentorrecallideastothemindofthe readeT.IcontendthattheChinesecharacters,thoughformedof different elements,donomore,andthattheyrepresemideasnootherwisethanas connected withthe words inwhichlanguage has clothed them, and therefore that they areconnected with sounds, not indeed asthelellers of our alphabet separatelytaken,but asthe groups formedbythem when joinedtogetherin the fonnof words.(DuPonceau1838:xi-xii.) DuPonceau foundhimself,inthe1830s,firstahesitant skeptic,later a confirmedopponent,of thethen,asnow,popularlyheldnotionthatthe Chineselanguagewaswrittenwithasocalled"ideographic" script,ascript that waslooked upon as unrelated tothe spoken language, and that instead wasthoughttoregisterandconveymeaningdirectlythroughsomeimag-ined appealtothe eye and mind without any recoursetowords or sounds. I Herecognized that whereusersof Westernalphabets areaccustomedtoas sociating asinglegraph, i.e.,aletter, withanindividual sound,theChinese associatedsingle graphs,i.e.,characters, withwhole words. Chinese charac teis arethus the functional equivalent of those groups of Westernletters we combine into unit sequencesthat stand,by and large, for words. Animportant corollarytothemistakenperceptionof Chinesecharac lers asideographs wastheequallymisleadingbelief that becausetheywere thoughtnottobeboundtospeech,butonlytoideas,i.e.,meaning,the charactersthusconstitutedawritingsystemthatcouldbereadbypeople whohadno knowledgeof theChinese language.Inproof of this somewhat improbableclaim,advocatespointedtothefactthatChinesecharacters were usedreadily by Koreans, Japanese, and Indochinese (called in Du Pon-ceau's time,and inhis book. Cochinchinese), none of whom necessarily had anyknowledgeof theChineselanguage.andbyspeakersof agreatmany mutually unintelligible Chinesedialects. IFor adiscussionof DuPonceau' splaceinAmericanlinguisticsin -general,andof his workinareas otherthantheChinese script,see Andresen1990: 97-104 et passim. 3 4TheOriginand Early DeueWpmmt of theChineseWriting System This confusion stillexiststoday, and stems fromabasicmisunderstan4-jogof thesignificanceof thefactthat Chinesecharactersstandforwords ratherLhanforindividualsounds.2 Bearinmindthatawordisaspoken thing;torefertothewrittenrepresentation of awordasa"word"isacon-venience, but isnot precise. Inasmuch as words,bydefinition,havenot only sound,b:ulalsomeaning.soChinesecharacters,whichstandforwords. thereforealsoalwayscarryameaningaswell.Likeanyother orthography, Chinesecharactersmaybeborrowedtowritethewordsof anotherlan-guage.Butuniikealphabets,whenChinesecharactersareborrowed,the borrowing istypicallyat thelevelof the word,whichincludes meaning,not at thelevelof theindividualsound.Becauseof thisitmayappearthatthe meaning of thecharacter hasbeentransferredalong withthegrnph,espe-ciallywhenthesoundof thewordintheborrowinglanguageisdifferent fromthesoundof thewordinChinese.Infactthecharacterhassimply beenusedLawritethewordinthesecondlanguagethatalreadyhasthe samemeaning lhal the character originally had inChinese, andthere isno 2Seebelow, p.18.The earliest Europeanexpressionof this viewof Chinese characters that Iknow of isfoundinFrancis Bacon 'sTiu Advancementof Uaming.Book II ,sectionXVI, dating from1605: Andweunderstandfurther.that itistheuseof China,andthekingdomsof the HighLevant,towriteincharactersrcal.whichexpressneitherlellersnorwordsin gross,butthingsornotions;insomuchascountriesandprovinces,whichunderstand notoneanother' slanguage.canneverthelessread oneanother'swritings.becausethe charactersareacceptedmoregenerallythanthelanguagesdoextend ... Uohnston 1974, 131). DavidMungellosuggeststhatthesourceof Bacon'sinformationmayhavebeenJuan Conza1es de Mendoza' sHistoritl . .. tkl granReyno tk la.China,published inthelast decades of the sixteenth century, and widely available shortlythereafter in England and on the continent (Mungello1985:184). Baconis confused about twopoints.First, whilethecharacters do not, of course, Mexpress letters," they do express words, and second, while people of differentand may be able toread individual characters, eventhough the ir languages are not mutually com-prehensible,theycannot infactMreadoneanother's writings,"sincereadingoneanother's writings presupposes knowingthe languages,not just the meanings of isolated words written withindividua1characten. By"charactersreal" he seems tomeanthat he thought of Chinese graphs. not asarbitrary signs or marks for sounds likethe letters of European scripts, but having a direct, non-arbitral)' re lationto "things or notions" independe nt of anyLinguisticmediation. It wasthis perceived non-arbitrariness, this "rea1ness, " of the script. that thrust Chinese tothe forefront of consid-e rationinthe seventeenth-century searchfor alinguauniversalis,capturing theattentionof suchfiguresasFr.AthanaSiusKircherandGottfriedWilhelmLeibniz.SeeMungcllo1985. ch. Vl ,"ProtoSinology andthe Seventeenth-Centul)'European for aUniversalLan-guage, " et passim. Thetenaciousholdthat this(mis)perceptionof thenature of theChinesescripthas en-joyed ever since is toa considerable extent, Isuspect,dueto the importancethat wasplaced onit inthis highlyintellectual and philosophical seventeenth-century milieu. Introduction 5 actual transfer of anything other thanthe graphic element itself. The trans-ferisbased on amatching of meaning inthe originallanguage with mean-ing in the recipient language,but the meaning itself is not borrowed.While itistruethat theKorean,Japanese,or Indochinesereaderdoesnotneed to knowChinesetobe abletoreadthe Chinese character,he doesneedto knoW" what. wordinhisownlanguagethecharacterhasbeenborrowedto write.Andhisapprehensionof ameaningwhenheseesthat characteris based on his knowledge of what word it represents inhis ownlanguage. and of themeaningofthatword,notonanythinginherentinthecharacter apart fromthat representation. Thisuseof Chinese characters,asDuPonceaurealized.isnodifferent fromtheuseof Arabicnumerals,e.g.,I, 2,3,.. . in most European writing systems. The graph, or character, we write (3), for example, may be read three if weare reading in English, or tres ifin Spanish, or drei, or tre,or trois,or even san or mi if weare reading in Japanese, or any number of other ways depend-ing on what languagethe character isbeing read in.3 Inthe last instance the graph stands fortwodifferent,but fundamentally synonymous, words in the same language. Each such graph, be it an Arabic numeral or a Chinese char-acter,stands for aword;thefactthat the wordmay be different inpronun-ciationfromlanguagetolanguage,orevenwithinasinglelanguage,is irrelevant tothenature of its writtenforminany particular instance. It certainlydoesnot followfromtheinitself ratherunexceptionalfact that the same graphic signmay stand for the same word in a variety of differ-entlanguages,thatthegraph,beit aChinesecharacteror anArabicnu-meral,standsforanidea.Suchgraphsstandforwords,inanynumberof languages perhaps, but words allthe same. The fact that the Japanese or Ko-reans chose to write their words largely, and in origin exclusively. with graphs borrowed from a different and linguistically unrelated source rather than de-vising a writing system of their own ex nihilo, whatever it may imply of histori-calor culturalinterest.saysnothingaboutthegraphicrenderingof ideas directly, something that tolieoutsidetheprovince of writing. Eventhemostardentadvocatesof theideographicnatureof Chinese characters usethe word "read" when they speak of what it isa speaker of one language or another does vis-a.-visChinese characters. But what does it mean to"read" agraphif not togivethat graph asemanticand aphonetic inter-pretation?Inother wordsChinesecharacters asreadbyaJapaneseorKo-rean speaker bearexactlytherelationtothe wordsof that speaker-reader's languagethatArabicnumeralsbeartothewordsfornumbersinWestern (andother)languages.i.e.,theyrepresent words.DuPonceaucalledsuch 3Sharpanglebrackets,viz. ,(and),willbeusedto set off characters,letters,orother marks whenwe "arereferringtothegraphitself,as opposedtothe sounds or wordsforwhich thegraphinquestionmaystand. Thus, (3)means'the graph3',asopposed tothenumber or word'three', or any other wordthat this graphmightrepresent. 6TheOriginand EarlyDI!IJelopment .o[ theChineseWriting System graphs lexigraphs;wenow more commonly call them logographs,bUllhe terms areequallyprecise.Aboutideographsheobserved:..... anideographicsys-tem of writingisa creature of theimagination, and. .. cannot possibly exist concurrently withalanguageof audible sounds" (DuPonceau1838: xxiv). The important point to recognizeisthat inasmuchas words are aninte-gralpanoflanguage,andhenceofspeech,logographs,whetherofthe ArabicnumeraltypeortheChinesecharactertype,represent elementsof language,andconstituteameansforthedirectrepresentationofspeech, justas sureJyasdolettersof aWesternalphabeticnotation.Thedifference isone of level.Chinese charactersand Arabicnumerals, as wellas ahostof other graphs used in various formsof writing,e.g.,thegraphQinthe con-text"IQBrooklyn," represent speechat thelevel o ftheword;le tters do so att helevelof theindividual sound. Il isalso common, of course, for writing systems to represent speech at a levelintermediatebetweenthat oftheindividualsound,andof theword. Such writing wouldbesyllabic,the syllablebeing that intermediatephonetic e ntity.Intheory the size or level of the linguisti cunit that isrepresented by thee lementsof awritingsystemiswhollyarbitrary.Thatistosay,agiven graphmay standforasingle sound (ormore properly,for asi nglemorpho-phon eme). asgrossomodo most letters do inan alphabeticscript, orfor asyl-labic,asinsyllabariesof themodernJapanesekind,or forwholewords. A graphtha tstandsforasyllableiscalled asyUabograPh,andonethat stands forawordis,aswehavesaid,alogograph,or,lesscommonly,alexigraph. There isno reasonin principle why asingle graph could not represent even an entire phrase, shouldthe speakers and writers of alanguage findit desir-able and usefultodevisesuch graphs.An example of suchagraphmight be thesign(%)standingforthephrasepercent,orthearithmeticsign(+) standing for the phrasedivided by,asin22 ...11=2.Inpractice singlegraphs standing forunits of speechatalevelhigherthanthat of thewordarenot common. DuPonceauexpressesthethree-way distinctionasfollows: . . . Chinesecharactersrepresentthewordsof thelanguage,andarein-tendedto awakentheremembrance ofthem inthemind,they arenot there-foreindependentof sounds,forwordsaresounds.Itmakesnodifference whetherthosesounds aresimpleandelementary,asthosewhichourletters represent, or whethertheyare compoundedfromtwoor threeof thoseele-mentsintoasyllable.Therearesyllabicalphabets,likethatoftheandother languages. and it has never been contended that they do'.lotrepre-sentsounds.AnditmakesnodifferencethattheChinesesyl1ablesarealso words,forthat does not makethemlosetheir character of sounds.BUl,onac-count of this difference,I would not calltheChinese characters asyUahic,but alogographic systemof writing. Thisbeingthecase,itseemsnecessarilytofollow, astheChinese characters areindirectconnexion withtheChinesespokenwords,theycan Introduction7 onlybereadandunderstood.bythosewhoarefamiliarwilhtheoralIan guage. (OuPonceau1838:1lO, emphasis original) DuPonceau' sconclusion,that Chinesecharacters(whenusedtowrile Chinese, and not another language) can only be read bysomeone"familiar withtheor.u language," stands, when seen inthis light,asareasonableand unremarkableobservation.Butinfactwestillfindrespectableexpressions of themistakenconvictionthat Chinese characters are somehowunre lated to language. For example, in aflyerpreparedasexhibition notesto accom panythedisplayof variouskindsofwritingintheBritishMuseumweare toldthat Chinese writingis a"concept script," and that "asaconcept script, Chinesedoesnot dependonthespokenword;itcanbereadwithoutre-gard'0, or evenaknowledge of,'he spoken language" (Gaur1984:2).4 To sechowuntenablethisclaimreallyisweneedonlytoconsiderthe implications ofsuchapossibility.If aknowledgeof thespokenlanguageis not aprerequisitetotheabilitytoreadChinesecharacters,thenallof us, irrespectiveof anytraininginChinese,oughttobeabletoreadthechar-actersof this"conceptscript."Thisisahypothesiseasilytested.Hereisa Chinese character,perfectly common and in everyday use fromtheClassical perioddowntothepresent:!iJ!;hereisanother:fl'ff;andthreemore: 1 l i ~ ; Z. If thehypothesisistrue,andthesecharactersstandformeanings or conceptsdirectly.withouttheinterventionofthemediumof language, thenanyoneshouldbeabletoreadthem.regardlessof hisorherknowl-edge of the Chinese language. That no one whodoes not already know Chi-nesecanreadthemis,of course,trivially obviousand suggeststhatthereis something seriously amiss withthedescriptionof Chi nese writing asa"con-cept script" andtheimpliedcoronaryclaimof a"universalreadabili ty"for its characters. Advocatesof theconcept-scri ptpremise,undaunted,mightinsistthat nothing is really wrong withtheirclaim, rather that wehavemisrepresented thetest,andthat Japanese or Koreans,for example,couldreadthesechar-acterswithoutanyknowledgeof theChineselanguage.Andsothey could. But whenthey do,theyarereadingthem intheir ownJapaneseor Korean language.Andeventhentheyarereadingastring of fiveisolatedwords.If theyhappento knowthemeaning of thesefivecharacterstakentogether as a single sentence, it is either because they have learned something of the Chi-neselanguage,orbecausethewholepatternhasbeenborrowedintotheir languageasanossifiedandsyntacticallyunanalyzableunit withitsoriginal Chinese meaning intact.In any of these cases the Japanese or Korean readers are having recourseto alinguistic entity that correlates Japanese or Korean -4The samepoint ismadeinvirtually thesamewordsinGaur1985:80.Nothinginthis fullerscholarlytreatment isofferedtomaketheclaimanymorepalatablethanitisinthe exhibitionnotes,designedasthey wereforpopular consumption. 8TheOriginand Early Development oj theChineseWriting System wordswiththesecharacters,andmeaningsonlyinassociationwiththose words. This means no morethanthat these characters are used to write cer-tain words in Japanese or Korean, having been borrowed from China at some point in the Chinese Middle Ages for just thispurpose.It has nothing to do with concepts, asreflectedinascript or otherwise. A second possible argument that might be invoked to savethe claim that Chinesecharacterscanbereadwithoutregardtothespokenlanguageis thatthisisnotafairtestbecausewehavewrittenthecharactersherein theirmodernformandthustheoriginalpictographicbasisfromwhich their meanings wouldhavebeen apprehensible directly has been lost.Were wetowritethem.theargument wouldgo,intheir earliest graphicshape. theywouldbereadablewithoutanynecessaryknowledgeof theChinese language.Bythis versionof the claim. Chinese writing isapparently a"con cept script"only initsoriginal form,not inanylater fonn.This isalso em-piricallytestable.at leasttoadegree.Whilewecannot knowhow aperson fromthe late secondmillenniumB.C.might reactto any of thesecharacters whenpresentedwiththemintheirsecondmillenniumB.C.graphicguise, wecan writethemthat wayfor ourselves.and asktowhat extent they seem to convey meaning directly as pictographs, without regardto any knowledge of theChinese language. Intheirearliestknowngraphicformsthefivecharactersthat wecited above appear as~ ,"Ii , 1-,~ ,and~(Kao1980:50,4,494, 230, 89).Mysus-picion isthat these forms,no lessthanthe standard ones first given,are in comprehensibletoanyonewhoknowsnothingoftheChineselanguage. andthatthereisnodirectpictographicconveyanceof meaningherethat couldconceivably justifytheclaimsof theconceptscript advocates.What ever validitythat notionmayhaveinother contexts,it doesnot pertainto Chinesecharacters,modernorancient.Infact,itcannotpertaintoany kindof writingsystem,asweshallshow,becauseit deniestherelationbe tweenwritingandlanguage,i.e.,betweenscriptandspeech.Writingis,in itsturn,aspokenthing.The claimthat itispossibletoread,i.e.,tounder stand,ascriptwhileatthesametimedenyingthatone, mustknow,i.e., understand,the language that the script isused to write. isinherently contra dictory.Thenotionof anykindof ascriptasindependentoflanguage seems onthefaceof it to be asheer impossibility, and yet this isan explicit claim of the "concept-script" advocates.The BritishMuseumflyercontrasts "conceptscript"with"phoneticscript,"whichisdescribedashavingthe "disadvantage"ofbeingdependentonlanguage(Gaur1984:2,emphasis added).This,it issuggested,means that "ideasmust firstbe translatedinto thesoundsof aparticularlanguageandthesesoundsmustthenbemade visibleintheformof conventionalizedsigns"(ibid.).Andthenwhenwe wanttoreadthisphoneticscript wemustreversetheprocess;"thesesigns mustagainberetranslatedbackintothesoundsof the[same]language and fromthereback into the original idea" (ibid.). Introtiw;ti6n 9 Leaving asidetherather formidableassumptionsaboutthe relationbe-tweenthought andlanguagethatthisstatement entails.weneedsimplyto point out that the only part of this descriptionthathasto do withtheingandreading of a"phoneticscript"isthehalf dealingwithmakingthe sounds"visibleintheformof conventionalizedsigns"(=writing).andre-translatingthesesigns"backintosounds"(=reading).Wewouldaddther that twoparts of the statement are as true of the Chinese script as theyareof anyother; eachChinese characterisinfactaconventionalized signthat makes a certain combination of sounds.usually a word(more tech-nically. avisible. and. whenread, it isre-translated into that com-binationofsounds,whichthenisunderstoodashavinganassociated meaning.The point that seemstohaveledtoconfusion, andtotheunten-abledistinctionbetweenaconcept script and aphoneticone,isthesimple factthat Chinesecharactersrender soundsvisibleawholewordatatime, whereasalphabets(thestereotypicalphoneticscript)doit,grossomodo, sound byindividual sound.5 There aretwosenses in whichone can speak of the origin and ment of writing.For want of better labels Ishallcallthesethematerial and thelinguistic.The former referstotheoriginand history of ascript seenas aphysicalobject, where attention is focusedon the script's outward appear-ance. This wouldincludeconsiderationof thepatterns of evolutionof the script's shape, how those shapes were affected bythe kinds of materials avail-ablefor writing.what methods wereusedinthephysicalact of writing,and consideration of the artistic qualities of the various graphic forms.The time and external circumstances of ascript's firstappearance, andthechanging context of itsuse,wouldalsoconstituteanimportant part of thematerial history of writing.Allof these considerations takentogether,combined with numerous other ancillary aspects of the history of writing and of anualscript, I see asthe script'S"outward" or "external" history,and by calling it "materialn Imeantoimply that itsstudy isof ascript asatangibleentity, theoriginanddevelopment of whichcanbetracedfromtheevidenceof 5Notallrecentpublicationsthat mentionChinese writingthismisconcep-tion. Geoffrey Sampson explicitly warns against it:... Chinese writing comes no closer than Englishor anyotherto'signifyingthoughtsdirectly,'ortoexpressing'things'ratherthan ' words.'Chinese script isthoroughly glouographic:it symboli zesunits of aparticular spoken language,namely the Chine5elanguage. withallits quirk! and illogicalities" (Sampson1985: 149).Sampsongoesontogivethreesr.raightfotwardlinguisticindicationstodemonstrate thisclaim.ThefirstisthatsynonymsinChinese,beingdifferentwords,arewritte nwith different characters,in spite of thefact that the "things" or "ideas"that they sland for are the same.The secondhastodo withthewritingof morphemicall y complexwordslikeEnglish buttercup,thethirdwiththewayinwhichthewritingreproducesthegrammarof thelan-guageaswellasthemeaningof thewords.Allof thesesuggestthattheChinesescriptis stri ctly aninstrument towritethe Chineselanguage, and not something independent of it. 10TheOriginand Early Droelapmentof theChineseWriting System physicalobjects and artifacts.This approachtothe study of writing isoften closely allied witharchaeological and art historicalresearch. Incontrasttotheexternalhistoryof ascript wemay speakof ascript's "internal" history;thisiswhatIhavereferredtoasthelinguistichistory of a script.Bythis Imean anaccount of the origin and evolution of a script seen in terms of its relationtolanguage,i.e.,how the script isstructured and op-erates initsprimary function asagraphicrepresentation of speech. Thematerialhistory of writingis-largely an empiricalthing; wecan ob-servethedataasphysicalobjectsanddrawvariousconclusionsfromour observations.Inthelinguistichistoryof ascript.bycontrast,thereisa theoreticaldimensionthat isabsent intheother.The study of the relation betweenscriptandlanguagecallsfortheidentificationof theprinciples that governthisrelation,andthusinvolvesthetheoryof writing.Techni-callysuchatheoreticalstudyshouldbecalledgrammatonomy.Morecom-monly,it iscalledgraphemics. Of thesetwodifferentkindsof historiesof writing,theoneIshallbe concernedwithinthepresent workisthesecond,the"internal,"or"lin-guistic" history, i.e., what wemight callthe grammatonomic history of Chinese writing. There is,of course, no absolute divide betweenthetwo,and consid-erations of ascript'smaterialhistory willoftenhaveadirectbearing onits developmentinthelinguisticorgrammatonomicsphere.Ihavetriedto takesuchaspectsof thematerialhistoryof Chinesewritinginto consider-ation whenever it seems called for.But this study isnot primarlly one of the externalhistoryof Chinesecharacters;forthat wenowhavetheexcellent recent monographbyProfessor QiuXiguiof PekingUniversity(Qiu1988). Rather the present work isconcerned withthe internal structure and evolu-tionof the Chinese writing system,and withtheprinciples that governed its evolution.Asaconsequence of thisapproach,thereishererelativelylittle appealtothe archaicformsof thecharacters-bone or bronze graphs,for example-in contrasttotheirmodern standard(k'ai shu~ . )forms.Nor haveIbeen concerned withthe techniques and procedures for deciphering unknown Shang or Western Chou characters and inscriptions.Important as this is,it isanundertaking distinct from, the grammatonomic history of the writingsystem.Wecanhope,of course,thattheunderstanding, wemight achieveof theprinciples of theChinese script willserve{hecauseof deci-pherment, but decipherment itself isnot apart of thepresent study. Writingarose,asfarasweknow,exnihilo onlythreetimesinold-world antiquity:inEgypt,inMesopotamia,andinChina,andonce Inthenew world,viz .theMayanscriptofMesoamerica.6 Scholarshave,ofcourse, speculated on the possibility that the invention of writing inone or more of theselocales wasinfluenced either directlyor indirectly byitsinventionin 6Ihavedeliberately left the stillundecipheredIndus Valley script out of consideration, and havenot included Mayanhieroglyphic writing fromthepre-Columbian New World in Introduction 11 another.Thereisnopersuasiveevidencetosupportsuchspeculations. WritingseemstohaveariseninEgypt and Mesopotamia at aboutthesame time,inthe mid- or late fourthmillenniums.c.,and inChina not untilthe middle of the second millenniumB. C.at the earliest. The near simultaneity, as weBasgeographical proximity, of theappearance of writing inEgypt and Mesopotamiahas,notsurprisingly,ledtoconsiderablespeculationabout thelikelihood of influence one wayortheother,but there isno indication of actual borrowing ineither direction.Near Eastern scholars allow only for thepossibility,seenbysomeasaprobability,thatthenotion or idea of wril-ingmight havetakenrool inEgypt asaresult of Sumerianinfluence, with-outanyactualborrowingof graphsorsystem(Ray1986:309-10;Fischer 1989:61-62). Ironically,the very factthat China is so remote fromthe Near East,andthat writingdidnotappearthereuntilsomanycenturieslater, has led tothe same kind of speculation,towit, that writing in China wasthe subsequent discussions. If weweretoincludeMayaninthecomparativepart of this study,it wouldfitthegeneralpatternthat seemstoaccount fortheinventionanddevelopment of writing very closely. UntilrecentlytheslandardworkonMayanhieroglyphicwritingwasThompson1971. Thompsondidnot alwaysrecognizethe script asarigorouslyphoneti crepresentationof a reallanguage, and foundhimself increasingly atodds withyounger scholars overthispoint. The firstportions of his chapter ontheprinciples of glyphicwriting,for example,are given overlargelytoconsiderdtions of grAphicstructureandcomposition,physicalarrangeme nt of texts,and the aesthetic qualities of the characters,withonly indirectattentiontotheway inwhichthescriptreflectstheMayanlanguage(Thompson1971 :36-65).Followingthe leadofurijKnorozov,Thompson'smainopponentinregardtothephoneticnatureof Mayanhieroglyphs, scholars now takeit for granted intheirresearchthatthewritingis fun-damenr2Hyaphoneticallybased script. See JUSlesonand Campbell1984,andthereviewby VictoriaR.Bricker1986,and Houston1988,whichhasavel)'fullbibliography of pertinent scholarship.Forabrief,but vel)'inte restingdescriptionof oneparticularlineofresearch seeMorell1986,writingontheworkofDavidStuart.For abriefsummaryinthepopular press ofthe mOstrecent work seeBlakeslee 1989. Of thepre-wargenerationof scholarswhoworkedonthedeciphermentof theMayan hieroglyphs,perhaps themustforcefuladvocateof the stri ctlyphoneticnatureof the script wasthefampusAmericanlinguistBenjaminLeeWharf.thoughhiscontributiontothede-ciphermentof specifichieroglyphsmayhavebeenlesssubstantialthanthatof full-time Mayanists(seeKelley1962:14-15).InapaperreadbeforetheMay.1940,meetingof the EighthAmericanScientificCongress, SectiononAnthropologicalSciences,inWashington, D.C.,Wharf inveighedagainstthestiflingandsterileargumentof whetherMayanhie ro-glyphs shouldbe calledideographic orphonetic.Herecognizedthat this supposeddistinc-tionis,inthecontextofwriting,entirelyvacuous,somethingthatfew,ifany,ofhis contemporariessaw withequalclarity:aconfigurativelinguisticstandpoin tthereis nodifference[betwee n' ideographic'and'phonetic'].' Ideographic'isanexampleof the so-called mentalisticterminology. whichtellsusnothing fromalinguisticpoint of view.No kindof writing,nomatter how crude or primitive, symbolizesideasdivorced fromlinguistic forms of expression .... Allwriting systems,including the Chinese, symbolize simpl y linguis-tic(Whorf 1941 : 483). 12TheOriginand Early Df!IH!lopmentoj theChineseWriting System resultof slow,longdistancestimulus-diffusionfromtheNearEasterncra-dle of Western civilization. Chinesehistorians and archaeologists rightly condemn such conjectures as wholly unfounded,pointing out that there is little indication of such con-tact or influe nce.They alsosometimesmaintainthat findsof neolithicpot-teryfragmt;ntsbearingavarietyof simplescratchesandstrokemarkson their sulfaces,and datingfromasearly as4800B.C.,suggestthat writingin Chinaisactuallymucholderthanhastraditionallybeenassumedonthe evidenceof Shangboneandbronzetexts.Advocates of thisclaimhavelet theirenthusiasmrununchecked,andseemtohavesuspendedtheircau-tiousandcriticaljudgment.[willargueinchapter2thatwhateverthe significanceorfunctionof thoseearlyneolithicmarksmighthavebeen, theywerenot.exceptpossiblyfortheveryparticularcaseof lateTawen k'ou"* t}! 0pictographs,relatedinanydirectorsubstantivewaytothe originof theChinese script weknow fromShang times on. TheapproachIhavetakeninpresentingtheoriginandhistoryof the Chinesewritingsystemis,inadeliberatelylimitedway,comparative.The reasonforthisisthathypothesesabout certainaspects of the development of writinginChinabecomemoreplausiblethantheyothclWisemightap-pear whenwediscoverthat similarprocessesseemtohavebeenat workin the evolutionof writinginbothEgyptandMesopotamia.Thecomparison suggeststhat wecansay withafairdegreeof confidencethat when writing arosein China it foHowedpari passu the samepattern of development inits formative stages asinboth Egypt and Mesopotamia.This was clearly not the result of cross-culturalinfluences,muchlessof chance,but rather thatthe principlesgoverningtheoriginandearlyevolutionof writinginallthree ancientsocieties-Egypt,Mesopotamia,andChina-werefundamentally thesame.Inother words,inthedevelopment of their writingtheChinese did not follow"some mysterious esoteric principles that set them apart from therest of thehumanrace,"asP.A.Boodberg already counseledushalf a century ago(Soodberg1937:331),but invented writingaccordingtowhat look like general, I amtemptedto sayuniversal,principles and patterns. Thebrief notesinMorell(] 986:55)andevenmorethediscussionin Campbell (1984:11-16) suggest thattheoriginand development of Mayan hieroglyphic writing followedthe same principles wecanidentify as govern-ingtheevolutionof writinginEgypt,Mesopotamia,andChina.Campbell (1984:12)summarizesthestagesasfollows:(a)truewritingemerges with logographicsigns;(b)the ,firststeptoward"phonetidsm,"that is,phonetic flexibilityintheuseof graphs,is"rebus"writing,orwhatwemaycall "punning";(c)phoneticcomplements,i.e. ,determinatives,arise;and(d) logographs come tobe used for their sound value alone, i.e.,they are "dese-manticized." This,inanutshell ,isthe early history of allknown WTitingsys-tems,and it is particularly satisfying to see now that Mayan writing confonns to this general pattern so closely.]fwe wishto claim this pattern as universal, IntroductiQn 13 the evidence from Mayan will not stand in the way.(Note that the fourth stage, that of desemanticizauon, demands special attention inthe case of Chinese, andinfact5infradealswiththequestionof whythisnevercame about fullyin China.)Recognitionof the possibleuniversalityof theseprin-ciplesgivesusafirmbasisfortheeventualdevelopment of asoundgram-matonomic theory,that is, atheory of writing. In. the present work,apart fromthepreliminary discussion in chapterI, thepossibility of ageneraltheoryof writingisonlytouched oninpassing, andmorebyimplicationthanbyexplicit statement.Mypurposeisinstead to describethenature and internal structure of the Chinese script fromthe timeof itsinventioninthemiddleof the Shang agedowntotheperiodof itssl3.ndardizationintheHan,andtodispelsomeof themisconceptions thathavelong surroundedit.Ihavedividedthediscussionof thisspanof roughly amillennium and ahalf intotwoparts:(I) theoriginand early de-velopment of thescript inthe Shang, which Ihave called the Shang Forma-tion,and(II)theregularizationandstandardizationofthescriptinthe Ch'in-Hanperiod.whichIcall&.heCh'in-HanReformation.Withthedis-covery and availability inthe last twenty years of aconsiderablebody of pre-HanandearlyHanmanuscriptswecanseemoreclearlythanheretofore theexact nature of thepre-Han,non-standardized.script,and assessmore accurately the effects of the Han standardization. This inturn enables usto identify previouslyunknownfeatures of the "reformation." Thetwopartsof thisstudydifferfromeachother inonefundamental respect.PartIisanefforttopresentanobjective,scientificallyfactualac-countoftheoriginanddevelopmentofChinesewritingin. theShang period,basedondirect scrutinyandanalysisof thecharactersthemselves. Part n. by contrast, forms itself around a consideration of how, in the Ch'in-Hanera amillenniumlater,theChineseperception of writing anditsrela-tiontolanguage, quite apart fromthe actual structure andhistory of either, shapedthe subsequent historyof theChinesescript.Part Iis,then,essen-tiallyculturallyneutral;detached, wemight say,fromaconcernwithother of Chinese civilization.Part IIin contrast deals withacentral part of theearlyculturalandintellectualhistoryof imperialChina,takingasits startingpoint thetraditionalChineseworld-viewandtheplaceof Chinese writing init.It ends withasuggestionthat the subjectiveperception of lan-guageandscript,andof therelationbetweenthetwothatcharacterized Chinesethinking intheCh'in-Hanperiod wasasmuchresponsibleforthe fact that the script remainedlogographic and did not movein the direction of ana1phabet asany purely linguisticfactorsmight havebeen. Thesetwoparts,takenasaunit, account forthe wholeof thelinguistic history,inthesensedefinedabove,of theoriginand early development of theChinesewritingsystem.Exceptfortheverycurious,butalsoveryob-scure,emergenceof whatappeartobelocalizednon-standardvarietiesof pre-HanChinesewriting-asseen,forexample,inthecharactersof the 14TheOriginand Early Dl!fJelopmmtof theChineseWriting System Ch'u silkmanuscript-the period betweenabout1000B.C.and 200B.C.did notwitnessanyfundamentalchangeintheprinciplesLhatgovernedthe structure of the script or its operation.7 There was, of course,considerable changeintheoutwardappearanceofthecharacters,intheinventoryof frequentlyusedgraphs,andinother aspectsof what wereferredtoasthe external. or material,history of the script.But these did not affect the inter-nal,theoreticalhistory of Chinese writing inany significant way. Whenwecometotheeh'in-Hanperiodtheprinciples governingthe structureandoperationof thescript begintoshowat leastapotentialfor significant change. In the third and second centuries B. C.the Chinese seem tohavebeguntoperceivetheir writing system ina waythat,had it actually fulfi lled itspotential, wouldhave likely entailed the widespread useof afew common graphs to stand not for syllables inherently associated with specific meanings.butforfundamentallyasemanticsyllablesthatcouldrepre-sentwhatevermeaning,i.e.,word.wascaJledforinaparticularcontext. The eventual result of such a development might well havebeen a syllabary, perhaps evenultimatelyanalphabet.This didnot, of course,happen;nor wasit arealpossibilityatanylatertime,eveninthefaceof thepowerful influenceof Westernalphabetictraditions.Thismeansthatinsofaraswe areconcernedwiththeinternal,linguistic,andtheoreticalhistoryof the Chinesewritingsystem,thereareonlytwocrucialperiods,whatIhave calledthe Shang Formation ontheone hand andtheCh'in-Han Reforma-tionontheother.Althoughseparatedbyagapof nearlyamillennium, thesetwoperiods areequallyimportant toafullaccount of thehistoryof the script.This iswhythepresent workisdividedintotwoparts,one deal-ing withthe first formation of the script,the second withits reformation, or rather the reaffirmationof that original formation, athousand yearslater. Because writingof anykindisno more andnolessthanagraphicrep-resentationof speech(thisdefinitionwillbediscussedformallyinchapter I),tostudyitsnatureandhistorywemustoftenhaverecoursetothe speech,that is,tothelanguage,thatthewriting writes.Inthe caseof Chi-nesecharactersatthetimeof theirinvention,thatlanguagewasthelan-guageof theShang people,i.e.,what wemaycallShang Chinese,orEarly ArchaicChinese.Forthewritingsystemof theeh'in-Hanperiodthelan-guagewas,obviously,aformof Chineseaboutathousandyearsremoved fromtheShang.andthiswemightcaBLateArchaicChinese.Ideallywe should have aknowledge of the phonetic structure of both of theseperiods of Chinese,and couch our remarks about the script,and how it represents the language, accordingly.But in fact the study of Chinese historicallinguis-ticshadnot yet reached thepoint were wecan say withany specificity what thephonetic structure of Shang Chinese was.Evenforthelanguage of the 7OntheCh'usilkmanuscript see Jao1958,Hayashi1972,Ts' ai1972,Barnard1972. 1972-73. Ch' en1984.and Li1985. Introduction 15 Hanwestillfacemanyunresolvedproblems andunansweredquestionsof considerablemoment.Linguistshavesofargenerallyhadtosatisfythem-selves withreconstructing asingle form of pre-Han Chinese, called typically OldChinese(abbreviated OC),andhavethenhad toacceptthe shaky cor-ollary that this willdo forallperiods of pre-Han linguistichistory. Currently wemight saythat there are availablefour distinct and to some extent. competing reconstructions of thephonetic structure of Old Chinese. The earliest of these, and the onemost accessibletonon-linguists.isthat of thelateBernhardKarlgren.codifiedinhisdictionary-formatworktitled Gmmmata&ricaRecen.sa(1957;hereafterabbreviatedGSR).Theonly other reconstructionthathasbeendescribedcompletelyenoughandsystemati-cally enough inpublished formto allow relatively easy use isthat of Li Fang-kuei(1971.1976).Theothertwo,bothof whichd e s e r v ~seriousattention, arethat of E.G.Pulleyblank(l973a,1977-78,1982,1984a)andthatpro-posed jointlybyNicholasC.BodmanandWilliamH.BaxterIII(Bodman 1980,Baxter1980).Neitheroftheselasttworeconstructionsisyetfully enough developedinavailablepublicationstobeuseablefOTour purposes here.IhavethereforechosentouseLiFang-kuei ' sreconstructiondespite thefactthat insomerespectshisproposalsareconservativeandartificial. TherearetworespectsinwhichIhavemodifiedLi'sreconstruction throughout:(i) wheninmy opinion theevidence callsfor a consonant chis-terinacertainworddifferent fromthatwhichLireconstructs,Ihavenot hesitatedtodivergefromhim;and(ii)Ihaveuniformlyreconstructedthe Old Chinese source of the Middle Chinese departing tone (ch 'u sheng~ .) asfinal-s ratherthanfinal_h.B 8Raxter' sreconstructi onof OldChineseisnow.atpageproof lime(summe r.1993), aV'diJablein aformidable,and richly informativevolume.SeeWilliamH.Baxter, AJlandboolt of Old Chimse Phonology. Berlin &NewYork:Moutonde Cruyter,1992. 1.WRITINGINGENERAL DEFINITIONOFWRJTING Allworksonwritingandwrltmgsystems,whethergeneralorscriptp specific.contemporaryorhistoricallyslanted,mustfacethesameissueat theoutset:howto define writing.Inconfronting thisproblem not afew au-thorsfindthemselvesinitiallyproposing descriptions of writing thatinvolve twothings,visualsignsandtheactofcommunication.Theythentryto forgeaformaldefinitionof writing by specifying apreciserelationbetween thesetwoelements. Giventhat weperceive writing as in some sense the visualcounterpart to speech,andwerecognizethefunctionof speechtobechieflythecom-munication of ideas, wequitenaturaHY'associatethevisibleformsthat con-stitutewritingwiththecommunicationof ideas.Thusweendupwitha relationalanalogyof thefollowingsort:writingistovisualcommunication aslanguage isto oral communication. If weacceptthisanalogy-writing:visualcommunication::language: oralcommunication-we areforcedtoadmit aswritinganyandallvisual signs or marks that convey meaning, e.g.,the skulland crossbones on abot-tleofmedicine,thecigarettewithacirdearounditandaheavybar throughit onthe wallof apublic room,thered cross onthe side of anam-bulance, and so forth,Yetif weadmit all such visual signs aswriting weend upwithadefinitionof writingthat goes wellbeyondour originalintuitive sensethat writing issomehowthe visualcounterpart tospeech. If,ontheotherhand,werecognizethatwhenwesay"thepurposeof writingistocommunicateideas" what wereal1ymeanisthat "thefunction of writing isprecisely to communicate what is communicated bythe speech iliatthe writing represents,"werestrictthescope of writingtothosevisual signsthemeaningof whichismediatedbylanguage.Inotherwords,the communication aspect of writing is .onlyan adjunct tothe fact that the writ-ing stands forlanguage, and it isthe language that isthe mechanism for the communicationof ideas.Theskullandcrossbonesis,tobesure,avisual signthat communicates averyspecificmeaning.But inthat actof cqmmu-nicationthere isno and automatic linguistic value necessarily associated withthe visualsign, The samepicture of the skull and crossbones couldbe"read"variouslyas'poison,''poisonous:'hazardous,''pirate,'or even'sku))andcrossbones.'1Becauseof thislinguistictheskull 1Ihaveappropriatedthe Mskulland crossbones" example fromY.RChao (1968:101). 16 Writing inGeneral 17 ,' andcrossbonesgraphisanexampleof thecommunicationof anideadi-rectlyrather thanone governed or mediatedbylanguage.Onthis basis we would deny it the status of writing. To do otherwise leads tochaos: we would havetoadmit as writing every image.painting. graphic symbolor iconthat evokedameaningfulassociationinthemind of thebeholder. Wemaythus define writing very simply asthegraphic representation of speech;and. awritingsystem.then. asany graphic means forthe systematic representationof speech.Likealldefinitions.thisdefinitionexpressesa judgment.It wouldbepossible.of course,todefinewritingdifferently.as, for example, any visualsignor mark that conveys or communicates meaning irrespectiveof itsrelationtolanguage.Inmy judgment suchadefinition doesnot clarifythenature or history of what weintuitivelythink of aswrit-ing anybetter thanadefinitionthatrestricts writingtothosegraphic signs thathaveadirectrepresentationalrelationtolanguage.Thebroader,and lessprecise, definitioninfactcomplicates the issueconsiderably,because it introducesnumerousconsiderationsthatarenotpertinenttothekindof writingthatrepresentsspeech,i.e.,towritinginthenarrower sense-and this.aswesaid above, leads tochaos.Certainly for our purposes here, if not in general,nothing isgained,andmuchislost,byLaking what Iwouldcall anon-linguistic viewof writing. 2 Thecommunicationaspectof writingis,bytheabovedefinition,sec-ondary,existingonlyasanautomaticconsequenceofthefactthatthe speechthatthewritingrepresentsservestocommunicatemeaning.M o r e ~ over,whetherornotanindividualsigninawritingsystemcommunicates meaning depends onthelevelat whichthat signrepresentslanguage.Let-tersofanalphabet,forexample,donottypicallycarrymeaning,only sound,becauseinmostlanguageswrittenwithalphabetsmostindividual sounds do not have anyassociatedmeanings.In Englishthelettersn,e,g,I. s,h,forexample,normally stand only forsounds. and do not communicate ameaning inisolation(exceptasthenames forthoserespectivelelters,of course).The lettersianda,incontrast, stand forsounds and insome cases 2This definitionmatchesthesensethatSaussureseemstoexpresswhenhesays"[a] languageanditswrittenformconstitutetwoseparatesystemsof signs.Thesalereasonfor the existence or thelatteristorepresent that former"(Saussure1983: 24).Eventhough Sau-ssuresays"language,"i.e.,langue,not"speech"(parole)inthispas!lage(seeEngler1989: 66b),itseems likelymat he wasrererringto "spokenlanguage"or "spokenutterances,ftnot tolangue inthe moreabstract sensc (see Vachek1973;10). This in tumallowsror an under-standing or "speech" and ""writing" astwocomparable but independent realizations(orrep-resentations, or manirestations) or language,the firstauditory,the second visual. Suchanunderstanding wouldgivetoWTitingastatus different rromtheoneIhaveal lowedinthedefinitionadoptedinthischapter,andrromtheoneSaussurewouldlikely havecountenanced.Whilethetheoreti calimplicationsorthisdifferentunderstandingor writing arenot without interest. Iamnot convinced that they are essentialto an understand-ing or writingproper. andinanycasetheyseemtomenottoimpinge significantlyonthe developmental and evolutionary mattersthat I shallbe dealing withhere. 18TheOriginand Early Developmentof theChineseWriting System (for i , only when \'ITittenI) for words,i.e.,they may carry meaning.But to be considered writingtheyneednot communicatemeaning.only sound.The communication of ideas inwriting is, as wehavesaid,entirely afunctionof the fact that the writing represents speech. It is just thismisunderstanding aboutthewayinwhichwritingconveys meaningthatisresponsibleformuchof thewidespreadconfusionabout thenatureof Chinese characters.BecauseChinesecharacters forthemost part represent speech at the levelof the word,or at least of the morpheme, not at thelevelof theindividualsound asour letters generally do;andbe-cause words, or morphemes,havemeanings. Chinese characters are inevita-bly associated withmeanings as wellaswithsounds in a waythat thegraphs of Westernalphabeticalsystemsarenormallynot.But theassociationwith meaning.i.e .thecommunicationaspect of thecharacters,existsonlyasa consequenceof theassociationwithsound,thatis,withwordsofaguage,or whatwemaycallsimply"speech ... 3 Whathasoftenhappenedis that academic analystsand casualcriticsalikehaveemphasizedthe link be-tweencharacter andmeaning attheexpenseof theprimaryandessential link between character and sound. Alllanguages haveboth words and morphemes, the former consisting of oneormoreof thelatter,andthelattertypicallydefinedasthesmallest meaningful element of alanguage.Becauseof thecharacteristicallymono-syllabicandisolatingstructureof Chinese,especiallyatthestage,it isnot misleading to speak of Chinesemorphemes astantamount to words,andtothink of the worditself asthe smallest entity of thelanguage that has both asound and ameaning.4 For any word wecanidentify twoaspects:sound and meaning.Whether thesetwoaspectsexistseparatelyandindependentlyof awordisnota linguisticquestionbutaphilosophicalone,onapar withthequestionof whether "whiteness" and "horseness" exist as separate entities apart fromthe white horsethat wecan see,smell , touch,and ride,and weneed not. fortu-nately,answerthat questionhere.For our purposes it issufficient torecog-nizethesetwoaspects of any word. Ishall adopt aslightlymodified version of Boodberg'sterminologyandconventions,andcallthe"sound"compo-nentor aspect of aworditsphoneticaspect,and whennecessaryabbreviate this with the upper-case letter P.Similarly, I shall call the meaning of a word itssemantic aspect,andabbreviatethiswiththeletter S(seeBoodberg1937: 331-33).Everywordhasthesetwoaspectsbydefinition,irrespectiveof whether or not it has a written form. Writing, as wehave said, consists of visualsigns or marks,thoughnot all visualsigns or marks qualify as writing.A single visual sign or mark wehave See Sampson1985:149(citedinnole 4 of the Introductioninfra) . . 4On the much debated question of the monosyllabicnature of Chinese and itsimplica-tions forthe writing system,seeBoltz 1989: A-3ff,and note 6 there. Writing in19 , called agraph.This we can abbreviate as G.Wecantreat pronunciation and meaning.i.e.,phoneticvalueand semanticvalue.asdistinctivefeaturesof graphs,regardless ,of whether aparticular graphiswriting or not.Therela-tionbetweena' graphandthesetwofeatures of "sound"(P) and"meaning" (S),canbe anyone of thefollowing,the"plus" sign(+) indicatingthatthe featureinquestionisassociatedwiththe graph,the"minus" sign(-),indi-cating that it isnot: (I)G (2)G (3)G (4)G [-P,-S] [-P,+S] [+P,+S] [+P,-S] Type(I) graphs, withneither anassociated pronunciation nor meaning. are merely idiosyncratic or randommarks or drawings,andhaveno bearing on writing. Type(2),on the other hand. withno established pronunciation, but carrying arecognized meaning,are ,exactly likethe skuHand crossbones signor the red cross on the side of anambulance. They are visual signs that communicate meaning but becausethey haveno automatic and unambigu-ous relationtolanguage,i.e.,becausethey are[-P1,they are, by definition, not writing. The essential and indispensable featurethat must be present for a graph or systemof graphs to qualify as writing isphonetic representation.That is, writing must represent speech.This means that it must be[+P].In Trager's , terms,writingisdefinedas"anyconventional systemof marksordrawings . .. that represents the utterances ofa language as such" (Trager 1974:377). Asearlyas1933Bloomfieldhadalreadyexplicitlystatedthat writingmust bear a"fixedrelation"to linguisticform(Bloomfield1933:283).Andmore recently Serruyshasdefinedthe graphs of truewriting asnecessarilygratedinasystem,"and"resultinginavisualrepresentationof alanguage" (Serruys1982:455). Whenawritingsystemarisesthatutilizesasinglegraphtorepresenta singleword.asisthecasewithChinesecharacters,thatgraphistype(3), [ + P,+S] . But the graph stands forthe meaning of the wordonly byvirtue of standing forthesound of thewordinquestion.Consider,forexample,the English word' eight' .Atthelinguisticlevel ,thatis,attheprimary andfun-damentalleveloflanguage proper, this wordhaslwO aspects,thephonetic, [eyt] , and the semantic,themeaning'eight'asthenumber betweenseven andnine.Atthegraphiclevel,thatof writing,whichisentirely secondary andderivative,thatistosay.whichcannot existexceptinrelationtothe primarylinguisticlevel,wecanrepresentthiswordwiththecharacter (8). The relationof thegraph(8)tolhe word'eight'canbe diagrammed sche-matically thus: 20TheOriginand Early Developmentoj theChineseWriting System graphic level: (secondary,derivative) linguisticlevel: (primary,fundamental) bullae > numerical tablets > full w r i t i n g ~cannot 28TheOriginand EarlyDevelopmentoj theChineseWriting System and(ii)caseswherewecanactuallyobservethedirectdevelopmentfrom precursor or forerunnertoactualwritingarerare.This doesnot mean,of course,that theearliest stageof writing wasnot ~ a r g e l ypi ctographicinori-gin;tobesure,allof the evidence suggests that it was.But forthemost part wehavenoactual,tangibleexamplesof pictographicprecursorstowriting thatwecanconfidentlysayare,ratherthan"mightbe, "theprototypesof specificgraphsin early script. Usingthedistinctivefeaturedesignationswesetoutabove,apicto graphic precursor is[- P,+5] ,that is,ithas ameaning (depicted directly by thegraph),butnoconventionallyassociatedpronunciation.Whenitac-quires suchapronunciation,becoming[+P,+8),it isthenby definitionan exampleof writing.Itispreciselythetransitionfrom[-P)to[+ P]that markstheshiftfromnon-writingprecursortowriting,andthatwewould liketobeabletodocumentwithspecificexamples.Butpictographicex-ampl es ofthis shift are elusive. 8chmandt-Besserat's evidenceseemstopro-videsuchexamples,but only fornon-pictographic graphs.What thismeans isthatthepictographic forerunners wenormally associate withtheoriginof writingarenot theonly source,orthesolemechanism,foritsemergence, andinfactarenotthesource withthemost clearlyidentifiablelinktothe advent of writing overall. beobserved. andprobably does not obtain.He concludes that asattractiveasthissequence might appear."itremains,atpresent,unsubstantiated" ( 1990: 56).Hethengoesonto raise otherequallyseriousobjectionstotheproposal,pointingout.forexample,thatbullae are knownthat containtokens not consistent withtheimpressions onthe outside surface. ForOurpurposesherewehaveacceptedthepossibili tythatSchmandt-8essaratiscor-recttosomedegree.because that allowsUStodiscussapossibleconcrete caseinwhichthe transitionfroma[-P,+S)graphtoa{+P,+S]one inconnectionwi ththe originof writing might havetakenplace.It may tum out that this phenomenonwas muchless monolithic and muchmorc haphazardthanitsproponents wouldhaveusbelieve.but it stillmay havebeen oncof several(many?)contextsinwhichatransitionfromnon-writingscmasiographs(0 writtenlogographs.Le .aKphoneticbreakthrough," occurred. PARTONE THESHANGFORMATION 2.WRITING INCHINESE OR1GINS The earliest formsof recognizable writing inChina aretheoracle-bone inscriptions(oftenabbreviatedOBI)ofthelaleShangperiod.ca.]200-1045B.C.,so-calledbecausetheyareprincipallyoracularordivinatoryin cootenl.) Theseinscriptionsarefoundincisedon(hescapulaeof oxenor sheep and onturtle shells,typicallythe ventralshell,calledaplastron,but occasionally alsoon the carapace.Inscriptions on oxscapulaeare common, thoseon sheep scapulae arenot. The script that these inscriptions reveal, although a fullydeveloped writ-ingsystemalready.stillpreservesunmistakabletracesof itspictographic origins. To illustratethiswemustbeginbyasking what wemeanby"picto-graphic."Figures 1 and 2 show atypicalincised plastron and scapularespec-tively.Figure3gives examplesof Shang oracle-boneinscriptioncharacters thataregenerallyconsideredtohaveidentifiablepictographicorigins. What willbe immediately apparent fromthese inscriptions and fromthis list isthathardlyasinglecharacter canactually beregarded aspictographic.if by"pictographic" wemean agraphthat depicts athingrealistically enough forusto identify it without knowing what wordthegraph stands for.Inthe simplestterms wemight saythat atruepictograph ought toidentifyathing totheviewer.andthat intumcallstotheviewer'smindthewordforthat thing.Schematicallythissuggeststhatthe"reading"of apictographpro-ceeds bythe mental sequence: PIcroGRAPJ.1>THI NG>WORD,akind of two-step processinwhichthelinguisticentity.theword,isintroducedonlyatthe second step and is associatedwiththe pictograph only throughthe interme-diary of theactualthing. This two-step process from THINGto WORDis exactly the reverse of the pro-cesswecustomarilycallreading.Reading,asitiscommonlyunderstood.is theprocessof assigning soundsto graphs.and of comprehending meaning IThe century and ahalf between1200 and1045 B. C.istheperiod to whichKeightley as-signs the Shang oracle-bone inscri ptions,representing the time of the reigns of the last eight (ornine)Shangkings.fromWuTingItTtoTiHsin1i. SeeKeightley1978:xi ii.For 1045 asthe yearof theChou conquest of the Shang, seeNivison1983.Nivisonhasinsubse quent publications.and invariousprivately circulatedworkingpapers. aswelluinviva voce discussions,acknowledgedmat thereissome questionabout thecertainty of the1045date. Itmayhavebeen1040instead'.SecNivison1990:156-57,andfootnote4.Shaughnessy (1991: accepts1045 ascorrectand gives anexcellent summary ofthekinds of data pertinent tothe question, and the kinds of argumentsthathavebeen offered . 31 32TheOrigin and EarlyDroelopmmtof theChineseWriting System FIG.I.Inscribedlurtleplastron . (FromChang1965) fromthe sounds, i.e., fromthe words that the graphs represent. The scheme forthis wouldbeGRAPH >SOUND >MEANING,whichistantamout:t1toGRAPH> WORD>THINGwhere "thing" isthemeaning of the word. This isprecisely op-positetothereading,or rather interpreting. of apictograph.A pictograph stands for a thing in the real or imagined world, and only secondarily and in-directly for aword.It sets off anassociativeprocess of THING>WORD.Inthis literal sense apictograph isnot akindof writing.Only whentheassociative processisfromWORDtoTHINC,i.e.,SOUND>MEANING,canagraphbesaidto constitutewriting.becauseitsdirectassociationiswithaword,not witha Writing inChinese 33 Fl c. 2.Inscribedoxscapula.(FromCh' u1961) thing. And if its association is witha word, its pictorial aspect isno longer es-sential ,thoughitmaystillbe vestigiallypresent.What areoftensaidtobe "pictographs" inthe Shang script are actually graphs of this latter type, asso-ciated withwords first,and withmeanings only through words.Eventhough insomeinstancesthecharactermaybepictographicallyidentifiableasa thing, e.g.,the Shang graphIfJ'formu ' eye'(no. 4 of figure3) or~for kuei ' turue' (no.12of figure3), functionally it stands for a word first , and a thing only by virtue of the fact that that wordmeans that thing. We shall call these graphs zodiographstodistinguishthemfrompictographs.(Seebelow,p.54) 34 The Originand EarlyDeuelopment of theChinese Writing System OBImodernmodern graphcharacterreadingmeaning I. t "siang'elephant' 2. 1\

niao'bird' 3. Ii 0 . 'ou orifice,mouth' 4. . /liT 13 ...'eye' 5. Fl.,ueh' moon' 6. IB EE , ' im 'cultivated fi eld ' 7. f;I:nu' (kneeling) woman' 8. Jt chi' (osier)9. 1 ;;R, 'ien ' overhead '> 'sky,heaven' 10.'f yang' sheep, ram' II. 1 1.li,...'horse' 12. Iilikun'turtl e' U. f! ,u' li5h' 14. l!U 'ing'tripod. cauldron ' nwi'grasp. ho ldin thehands' 15. FIG.3.Examplesof Shangoracle-boneinKriptioncharacterswithostensiblyrecognizable pictographic origins. Theappearanceof writinginChinaaround1200B.C.isconsiderably later than itsappearance inMesopotamia or Egyptinthe secondhalf of the fourthmillenniumB.C.Thisoccasionally givesrisetothe suspicionthat haps writing in China owes its origin to some remote and indiscernible influ-ence fromthe Near East. There isno tangibleevidence known at present to suggest that this wasthe case, or that Chinese writing isthe result of any kind of stimulus-diffusion,however indirect, frompoints outside of China. Writing inChinest 35 The apparent latenessof theinventionof writinginChina incompari-sonwithitsappearanceelsewherehaspromptedsomearchaeologistsand historianstoseek c;videncethat wouldshow that writinginChina wasactu-allyinvented much earlier thantheoracle-bone inscriptionevidencewould suggest.TothisendsomeChinesearchaeologistsandhistorianshave claimedthatthemarksandsignsfoundinlargenumberonpotteryfrag-ments. and similar artifacts fromwidelyscatteredneolithic sites across north andwestChinaconstitutetherealoriginof Chinesewriting,whichthen wouldbeinevidenceasearly asthe fifthmillenniumB.C.,significantlypre-dating anything known fromEgypt or Mesopotamia. Theneolithic marks for which these claims are madefallinto twoclearly distinctgroups.Thefirstconsistsof primitivemarksthatresemblelittle morethanscratches,eachconsistingof anywherefromonetofiveor six strokes,arrangedinvarioussimpleangularconfigurations,paintedorin-cisedonthesurfaceof ceramicpotsherds.Themarks of the second group. bycontrast, are carefully executed depictivedesigns. Themarksof thefirstgroupsometimes seemtosuggestarudimentary kindof rally-keeping; at other timestheyappeartobeno morethandeco-rativedesigns,orperhapsidentifyingemblemsorinsigniaof somekind. Theredoesnot seemtobe anymeaningfulorder of repetitionor concate-nationthat would leadusto suspect anythingmorethanthatthese areran-domand largelyunorganized,unsystematic markings. Inmost cases asingle potsherdhasonlyoneor twomarksonit;pieceswithmoreareinthemi nority.Moreover, whilethe archaeological record of pre Shang China shows a widescattering of suchfinds fromabout 5000B.C.downtothe Shang con tinuously,most of these consist of a very fewmarks persite, sometimes only one or twoindividualscratches at a given location.2 This suggests that there isno underlying systemtothe marks, andthat there wasno pattern of usage of themarks by any significant number of people eveninone site,much less over an area extending beyond a single locale. Those few marks that arc found inmorethanone siteare sosimpleand general , e.g.,singlestrokeslikeI , -, II.or crossesand angles likeX , +,1\ , that inalllikeHhood they arose independently ineachdifferentlocale,andareonlyfortuitouslysimilar or identi cal to signs used elsewhere.Ina fewplacesmore than just ascattering of potsherds havebeen foundwithmarks onthem, and these areillustrated in figures 4 through 7.Even in these cases there is no evidence that the marks constituted a systematic devicefor tally-keeping or writing of any kind.3 Some archaeologists, palaeographers, andhistorians wouldliketo see in this group of neolithicmarks,illustratedinfigures4through7anddating 2 For at:oncise summary, with a cumulative chart, oCthese neolilhic marb, see Ch'en 1985. ,For detailed summaries of thesefindsand background information, see Cheung 1983. Cheung gives a fullbibliography of the Chinese and Western studies of theseneolithic mark-ings upto1979.Figures 4through 7 arefTOmQiu1978. 36TheOriginand EarlyDevelopmentoj theChinese Writing System , " x+II1 1'T FIG.4.Neolithicpotte rymarks foundonfragme nts fromPan p'o ts' un*ltttl ; near modem Hsi-an,Shcnsiprovince.Yangshaof!lJiiJculture.sitedatedbyC-14techniquesto4800-4200B. C.(FromQiu1978) I x + o f FIG.5.NeolithicponerymarltsfoundonfragmentsfromPanshan* tlJandMach' ang Jil&:inRansuprovince.Machia)'30. a ; ~ t t iculture ,sitesdated1.02700- 2000B.C.(From Qiu1978) II -1/11X+ /\L o -tln FIG.6.NeolithicpotterymarksfoundonfragmentsfromLiuwanN , Letuhsicn .111 . Chinghaiprovince .Machia)'200glj{1j;culture.sitedatedto2400- 2000B. C.(FromQiu 1978) Ix FIG.7NeolithicpotterymarksfoundonfragmentsfromLiangchu.a.nf Inearmodem Hang-chou, Chekiang province . Liang chu culture; C-14 dating of sitenot available.but said byQiuXiguiLobecontemporarywithLungshanftLlJculture.ca.3(K)O-2000B. C.(From Qiu1978) toaseadyas4800B.C. ,theoriginof Chinesecharacters,thusmakingthe advent of writinginChina asmuchasamillenniumearlierthanineither the Near East or Egypt (Ch' en 1978, Ho 1975, YU1973). [n their enthusiasm forthishypothesistheypointtothesimilaritiesingraphicshapebetween variousShang oracleboneinscriptioncharacters of ca.1200-1050 B.C.and these marks fromthe fifthmillenniumB. C.as evidence that the latter are the prototype, or source, of the former. The most frequently cited examples are showninfigure8. Writing in Chinese 37 OBImodern graphcharactermeaning I.'one 2. -' two' 3. = ' three' 4.a I1!l ' four' S.X E. ' five' 6. + {;'seven' 7. J( J\ 'eight' s. + ' ten' 9.II it ' twenty' 10.T 7J' 'reveal' 11 . t-.3S: 'jade' 12. ~.!l!. ' tumulus' 13. =It-# 'water-well' 14. A A ' enter' IS. t:1 0 ' mouth' 16. t!/J 'grot.ss' FIG.8.Examplesof Shangoracle-boneinscriptioncharacterswithwhichneolithicpottery marksare sometimes compared. K.C.Changhassaid about thesemarks that, inhis opinion, ... the ovenvhelming majoriry of the ceramicmarks,both Shang andpre-historic,weremarkersandemblemsof famil ies,lineages.clans or some divi-sionsof these.andassuch,attemptstomatchthemwithknowncharacters and to translatetheminto meaningful sentences areunlikely tobe productive inmost cases.(Chang1980:245) 38TheOriginand Early Develapmentof theChineseWriting System There aretwoprima faciereasons toagree withChang and to doubt that theneolithicmarks haveany direct relationtoShang characters.First,iden-tificationbasedsolelyongraphicsimilarityisnotoriouslyriskyandincon-clusive,allthemorcsowhenthegraphsareverysimpleangular configurations of strokes.If thearchaeologists or palaeographers wereable toshowanyindicationthattherewasasimilarityinfunctionormeaning be-tweentheneolithicmarksandtheShanggraphs.theywouldhaveafar moreconvincingcasethantheydoatpresent.Butthereisnosuchevi-dence,becausethereissofarnowaytoknowwhateitherthefunctionor meaning of the marks ontheseearly artifacts was.Second,the sheer extent of timeinvolved,from4800B. C.to1200B. C.,aperiodof overthreeanda halfthousandyears,weighsheavilyagainstanyconnectionatall.Onthe faceof it, it would appear virtually impossiblethat thenascent seeds of writ-ing couldhavegerminatedinthemid-fifthmillenniumB. C.but not grown intoanything approaching a real writing systemuntil more thanthree thou-sandyearslater.Suchadevelopmentwouldbeincomprehensible;writing systems simply do not evolvethat way.If apotential for writing arisesinthe formof graphs or marksstanding fornames or words,no matter what kind or how limited,that potentialmust either fulfillitself apace,culminating in aviable,full-fledgedsystem,or witheranddie.Ahalf-waywriting systemis nosystematall,andthereislittlelikelihoodthat aninchoateattemptat writing wouldor couldremaininakindof limboor suspendedanimation forseveralmillenniabeforeachievingtheformof atruewritingsystem.If thatpotentialsystemdidnot developintorealwritingreasonablyexpedi-tiously,there would be no reasonfor apeople to preserveits embryonic bits and pieces.As apracticalmatter a writing systemissomething that iseither achievedquickly, or not at all. ChengTe-k'unhasclaimedthat"theinventionanddevelopmentof writing in China wasacontinuous process covering aperiod of no lessthan 6000 years" (Cheng1982:22).But writing isaninvention. not the end prod-uctof anevolutionarydevelopment,andlikeallinventionsitmusthave beentheresultof amomentaryoccurrence.Inthiscase.themoment was whensomeonerealizedthatagraphorsigncouldstandforawordor name.orsomeothermeaningfulunitof language.Priortowritingthere may havebeen, aswehave discussed above, any number of marks, drawings, and pictographs of variouskinds that prefigured inaphysicalsensetbe ad-ventof writing.Andsubsequenttoitsinventiontherecertainlyhavebeen continualchanges,refinements,simplifications,extensions.andmodifica-tions,of allsorts,that affectthewritingsystem.Allof thesefeatures.both beforeand aftertheinvention of writing, maybe loosely considered evolu-tionaryaspects.But theinventionitself wasnot anevolutionbut akindof realization, and itmust havebeen apunctual event. If that realization came as early as4800 B. C.inChina, then what, wemust ask.happened toit forthe next three thousand yearsbefore it isseen againintheShang? The fact that Writing in Ch.inese 39 thereisno signof atruewritingsystemduringthisentireperiodindicates rather compellingly that theantecedents of the Shang script,and therefore of whatweknow~ sChinesewritingingeneral,cannotbefoundinthese simpleneolithic marks fromthe third,fourth. or fifthmillennia B. C. Anumberof scholarshaverecentlylookedupontheappearanceof Egyptianhieroglyphic and Sumerian cuneiform writing similarly asmomen-taryinventions ratherthanastheproducts of aprolonged evolution.Apro-pos of Egyptian. JohnD.Rayhas allowedforthepossibilitythat "one mind mayhaveformulatedthebasicprinciples"(Ray1986:311).AndH.G. Fischerhassuggestedthesamething withhis comment that "sincethehi-eroglyphicsystemdoesnot seemtohaveundergone alongperiodof incu-bation,itmaywellhavebeenconceivedbyasingleindividual"(Fischer 1989: 66).In discussing the when and where" of the origin of the Sumerian cuneiform script,MarvinA.Powell hassaidthat it cannot bethought of as havingarisenthrougha"slowaccretionof signaftersignfromgeneration togeneration:thereisnot asingleinstanceinthehistoryof writingfora communal-evolutionary inventionof ascript.Individualsinvent.Thecom-munity of usen;modify,adapt,elaborate,refine,addto,andtakeaway,but theydonot invent"(Powell1981:422).The samethingcanbe saidof the inventionof theChinesescript.Theonly surmisethathasany substantive foundationat a11isthat theChinesescript wasinventedsomewhereinthe regionof theearly Shang state,inaboutthemiddleof thesecondmillen-niumB.C.Givenour current knowledge,alleffortstopushitsoriginearlier bycenturies,or inthemost enthusiastic cases,bymillennia,are unsubstan-tiatedspeculationandwishfulthinking.Thepossibilitythat asignificantly earlierdatefortheinventionoftheChinesescriptmaysomedaybees-tablishedonthebasisofconvincingpalaeographicandarchaeological evidencecannot,of course,beperemptorily ruledout.But whensuchevi-dence comes to hand it willhave to consist of morethan impressionistic and sporadicmatchingsbetweensimplemarksof undetermined soundor sense andlater,knowngraphs.It willhavetocomeforthinaformunambigu-ouslyrelatabletotheShangscript anddemonstrablyrepresentativeof the Chinese language. It isoftenobservedthat writinginMesopotamiaaroseinaneconomic contextof accountingpractices:reckoning,tallying,auditing,andother generalproceduresof record-keeping.PierreAmiet,Conservateur enchef of theDepartement des Antiquites orientales at theLouvreremarksthat: ... lesmelropolesdesdeuxregionsadjacentes jouerentunroledccisif: UrukenSumeretSuseaupieddesmontsZagros,00.furentorganisesles premiersEtatsdignesdecenom,d'abordparuneruptureaveclatradition prehistoriquequesymbolisaitlapoteriepeinte,puisparI'elaborationd' une comptabilite devenue indispensable a la gestiond'une richesse cnorme.Cette comptabiliteamenacommenaturellementlacreationdusystemed'ecriture 40TheOriginand EarlyDevelopmentof the ChineseWriting System encoreeleme ntaire,trespartiellementpiclographique,largementabstrail, quietaitappeJeadevenircuneiformeetallaitetreadopteetadapteparla plupart des peuplesdeI'ancienOrient.CelteecrilureestatleSleea Uruk a la findeI'epoquedememenom,vcrs3300avoJ -C.,alorsquelesvoisinsde ffiemecultureDepraliquaienlquelacomptabilitenumerale.Ecritureet comptabilitefurentmisesenoeuvreparuneadministrationsacerdotalequi palronna un art rcsolumem fealiSlepar opposition a lastylisationpropre aux prehistoriques.(Amict 1982:19) Writinginthe same exhibitioncatalogue from which Amiet's comments abovearecited, JeanBOllero.Direcleur d'Etudes at theEcolePratique des HautesEtudes,IVe sec.,observes: FondeesurI'agricultureceri:aliereintensiveet)' eleyageengranddu menubetail,Ielout e ntreles mainsd ' unpouvoir centralise, elles'estrapide-me ntempetreedansuneeconomietentaculaire,quirendaitinevitableIe controlemeticuleuxdesmouvementsinfinis,etinfinime nlcompliques,des biensproduits ct misencirculation.C'estpour subvenir a celtetache,en la fadliltantetlagarantissantparlamemorisation,queI' onamisaupoint l' ecriture:defait,pendantplusieurssicdesapresson.. invention .. ,ellen'a sema presqueriend ' autre.(Souero1982:28) Jack Goody has devoted an e ntirechapterof one of his recent books on writingandsocietytoexaminingtheimplicationsof thisapparentcause-and-effectrelationintheancie ntNearEastbetweeneconomicandcom-mercialactivityontheo nehandandtheadve ntof writingontheother (Goody 1986:45-86)4 Theargument for acause-and-effect relationinMesopotamia,between society' sneedtokeepaccurateandcomplexcommercialaccountsandthe inventionof writing to accommodate that need, seems generally defensible. Becauseof this there might be atendency to regardthe identificationof any societalneed that may besupposedtohavestimulatedthe invention of writ-ing as consti tuting indirect evidence forthat invention, even in theabsence of directindepe nde nt primary evidence for theexistence of writing itself. In other words, wemight want to saythat,if wecould first determinethat wr it-ing could arise in response to cer tainprecisely specified social conditions or needs,andthe nshowthatthoseconditionsorneedsexistedatacertain time,wewouldbe in apositionto argue fortheinventionof writing at the timeinquestionevenwhenthereisnoknowndirecttangibletestimonyto the actualexistence of writinguntillater.Ido not think suchanargument wouldbetenable.If thedirect evidence for writing intheancient Near East inthemid-fourthmillenniumB. C.werenot asabundant asit aClual1yis,I doubt that the cause-and-effect relationbetween accountancy needs and the 4ThetwOpassagesquotedabove,fromAmietandSottero,arebothcitedbyCoody ] 986:48-49, in English . Writing inChinese 41 . invention of writing would appear ascompelling as it does. Thereis no good reason, after all,to think that commerce inthe mid-Shang wasany less com-plexthan it wasin .Mesopotamia,and yet writing didnot ariseto accommo-dateaccountancy requirements inChina asit didintheancient Near East. It wouldbeniceif wecouldidentify some socialconditions or needsin Chinathatmighthavestimulatedtheinventionof writingthere,compa-rable thewayit isthought that theaccountancyrequirements of increas-ingly complexcommercialactivitydidinMesOpotamia.Becausetheextant Shangtextsarevirtually alldivinatoryinscriptions,it issometimesclaimed that writinginChina aroseinareligiousand sacrificialcontext.5 This sur-miseleavesunansweredoneof themajorquestionsinconnectionwiththe inventionof writingin China,viz.,whyit arose just whenitdid,apparently inthe middle or latesecond millenniumB.C. David N. (1987,1989) has givenserious attention to this prob-leminconnectionwithhisstudiesof mensurationandcalculationinlate neolithiccraftsmanshipandtechnology.HeinvokesColinRenfrew' ssug- , gestionthat writingmayhavedevelopedintheAegeanasameanstodeal withtechnologicalproblems of mensurationandreckoning inbronzecast-ing,and he thenallows for thepossibility of suchaconnection in China, es-pecially inthe east-coast cultures, where herecognizes apropensity for what he calls a'componential cast of mind" (Keightley1987:112). It was in the east, as opposed tothe northwest, where ceramic and bronze artifactsareseentohavebeencomponcntiallyconstructed-afeatureof theirmanufacturethat demandedprecisemeasurementandfit.Theskill withwhichritualjadeobjectswerecarved,or correctly,abraded,to exact dimensional specifications also called,Keightley suggests, for an extra-ordinary degreeof precisioninmeasuring.Keightleywouldliketoseethe inventionof theChinese script,componential asitunarguably isalready in 1200B. C. ,asafurthermanifestationof thissame"componential"mentality that he hasassociatedwiththeneolithiccultureof theeast coast of China, and at the same rime asa responseto the needfor precision intechnological mensuration, as Renfrew suggests it to havebeen inthe Aegean(seeKeight-ley1987:110,112,Jl6). Inhislater and fullerstatement of thishypothesis (Keightley 1989: 192-98)Keightleylinks the need forrecordingmeasurements withthe religious andlineageconcernsforveneratingthedeadandvalidatingthestatusof thelivingdescendants.Thishedoesbyrecognizingtheimportanceof ce-ramics and jades, especially thepreciselycrafted jadepi'discs'andtS 'ung ' tubes', justthekindsof itemsthat wouldrequireprecisemeasuringin 5Weshouldalsoallowforthepossibilitythat writingW"dSusedintheShangperiodin mundaneaswellasreligiouscontexts,butthatwehavenoarchaeologicalvestiges,and hence no direC[ or specificknowledge,of those everyday contexts, becausethematerialsthat wereused wereperishable,unlikethebones and shell s of the divinatory inscriptions. 42TheOriginand Early Developmentof theChineseWriting System theirproduction, asgravegoods.Giventhislink,KeighlleysaysMwecansee that religiousand lineageconcernsmighthaveprovidedimportant encour-agementfortheinventionof awrittenscript"(Keightley1989:197),and again.. . .. lineage-relatedactivities-suchasthemanufactureof mortuary jadesintheNeolithic,andof ritualbronzesintheShang,aswelJasthe creationof somesystemof lineageidentifi cation-mighthavestimulated the development of writing"(ibid.,emphasis added inbothinstances). If Iunderstandtheargument.threepoints having to do withtheinven-tionof writing arebeing proposed here:(i)that because jade, ceramic, and bronze items weremanufactured ineast-coast neolithiccultures byahighly developed technologythat involvedacapacity for precise measuring,and in thecasesofceramicsandbronzesl concomitantlyinvolvedthecarcful fittingtogether of picces(i.c.,components),writingmayhaveariseninre sponsetotheneedforameanstoregisterandmanipulatethesemeasure-ments;(ii)that thereligious and mortuary import of many of theseitems as they wereassociatedwithlineageconcerns gavcafurtherurgencytotheir propcr manufacture, and thus afunher encouragement totheinventionof writing;and(iii)thatthecomponentialmentalitythat wasresponsiblefor thiskindof ceramicandbronzeproduction(aswellasforanumberof otherfeaturesthatKcightleyidentifiesasinsomesensecomponentialin naturea ndrequiringexactmethodsinmensuration;seeKeightley1989: 195)wasalsoresponsibleforthecomponentialstructureof theChinese writing system,that is,forthefactthat the script evenin its earliest attested fonnconsistsof charactersthatareoftenconstitutedof twoormoreele ments,typicallya"phonetic" and a"semantic" element.6 Beyondthis,it iscrucialtoKeightley'sproposalthatwerecognizethe cleardifferencebetweenthe"holistic"(Keightley'sword,1989:193)and "non-componentiallyinclined"culturesof theneolithicnorthwest,onthe .one hand, and the"componential predisposition" of the neolithic eastcoast cultures,ontheother.Noteinthisconnectionthatthekindof simple strokesandangularmarkingsthat wedescribedaboveassometimesiden-tifiedasprecursorsof Chinesewriting(figures4through7)aretypicalof northwestneolithicculture sites.Of theexamplesgiven,onlythoseof the Liangchu site(figure7)arefromaneast-coast culture.Thesecond group of neolithicmarks.discussed below,areby contrast uniformly fromsites of theeast-coastregion,andarecharacteristicallycomponentialfigures,ei-thercomplex geometricdesignsorrealisticrepresentations of identifiable things.The distinction, in other words,betweenthe twogroups of neolithic 6The terms "phonetic" and "semantic" arenot precise or well-defi ned, but are often in vokedinconnectionwiththestructureof Chinesecharacters.Inhis1989 articleKeightJey doesnotusetheseterms,but inthe1987articleherefersto"phoneticandsemanticsym-b o J s ~(p.I J2).TheprecisetechnicaltermsarephonopMric andsemantic thtenninative respec tively;definitions are giveninthe Glossary. and afulldiscussion of this aspect of theChinese scri pt willbe foundlater inthepresent chapter. Writing in Chinese 43 markscorrespondstypologicallytoKeightley'sdistinctionbetweenthe northwest andtheeastcoastneolithiccultures. ThekeywordinKeightley'sstatementofhishypothesisaboutmen surationandtheinventionof writingismight.Needsassociatedwiththe demands of precisioninthe measuring and manufacturing of secular or re ligious objectsmight havehad something to do with theinventionof writing inChina,but thereisasfarasIknownotangibleordirectevidencethat theydid.AsIhavealreadyclaimed,andasIhopetoillustrateconcretely, the invention of writing followedpari passu the same pattern of development everywhere, it arosethroughouttheancient world,includingMayanwriting inthe New World.The componential structure of graphs at acertainpoint initsearlydevelopmentisapartof thisuniversalpattern.Ifthecompo nentialstructure of Chinesecharactersisasignificant, andnot merely for tuitous,refl ectionof the"componentialcastof mind"thatKeightleyhas identifiedascharacteristicof theeastcoastneolithiccultures,thenthose other culturesand societies intheancient world where writing appearedde novo,viz.,Egyptian,Mesopotamian,andMayan,must alsohavebeenof the same"componentialmentality."Keightleyseemstoall owforthisproposi tionimplicitly inhisconcludingremarkstothis sectionof the1989article: The argument isnot that only thosewhomadecomponentialpots werelikely to invent a componential writing system. The argument is rather that, giventhe increasing social and craft complexity evident inthe LateNeolithic, writing was more likely todevelop firstinthe region where suchhabits of organization,in various aspects of life, were more pronounced and valued. {Kcightley 1989:19B} The implication,ifone wereto extrapolate this point to cultures beyond theChinesedomain,seemstobethatintheancientNearEast,inEgypt, andintheMayancivilizationof CentralAmericathelevelof "socialand craft complexity"musthavebeensuchthat writing waspronetoariseina waythat it wasnot elsewhere.Thisis,onthefaceof it,not animplausible proposition,but neither isittestableor demonstrable inany waythat I can see,andthereforeit isof somewhat limitedmoment. As wehave said, writing isknown from independent evidenceto exist no earlier thanthe late Shang, that is,fromabout1200 B.C.Wemight look with increasedconcentrationforindependent evidencethat writingactually ex isted in the late third millennium 8.c.-the time when these mensuration re quirementsmighthave,byKeightley'sconjecture,stimulateditsinvention (Keightley1989:198).To myknowledgeno suchevidence yethas beenun-covered,but that certainly does not meanthat it willnot someday be found. At the same time, rather thantry topush the advent of writing in China back toadatesignificantlyearlierthanthattowhichdirectarchaeologicalevi dence attests, wemight also look for changes in those mensuration and man ufacturing aspects of the technological world of the second hal f of the second millennium 8.C.intheeast-coast culturesthat might beassociatedwiththe appearance of writing just at the time when weknow it to be first in existence. 44TheOriginand EarlyDroelopment of theChineseWriting System Quiteapartfromthesereasonstobeskepticalof claimsforafifth, fourth, or third-miUennium origin forthe Chinese script,there isafurther consideration that has to do withhow writing originates and evolves in prac-tice.NotwithstandingthethesisofDeniseSchmandt-Besseratdiscussed above, everything weknow about the originof writing in antiquity, whether inEgypt.Mesopotamia, or China. suggests that writing arose inthe main via theinventionof pictographsof concrete,depictableobjectsoracts,and evolved fromthat to a of fullgraphic representation of speech.How thishappened in China isdiscussedinconsiderabledetailinthefollowing chapters. Suffice it to say for now that inview of the central role that we rec-ognizepictographs to haveplayedintheformativestage of writing,totake theneolithicscratches,whichareinnowaypictographic,asthesourceof Chinesecharacters,manyof whichhavedearlypictographicformsinthe Shangperiod,isto violatethefamiliarpatternaccordingtowhichwriting seemstohaveariseneverywhere.There is,of course,no reasonto deema familiarpatterntobedecisiveinthefaceof evidencetothecontrary.But inthis case there is no real evidence to the contrary, and the recognized pat-tern,moreover. is,as weshall see, apan of an overalltheory that has an ex-planatory as wellas descriptive capacity to account for the early evolution of writingingeneral. The second group of neolithicmarksthat wereferredtoaboveconsists of quitedifferentkindsof graphs,asshowninfigures9,10,II,and12. ThesegraphsareallassociatedwiththeLiangchuit and Ta wenk'ou * til 0cultures,consideredbyChinesearchaeologiststospantherather long period fromasearly asthe middl eof the fifthmillennium s.c. downto about 2000 B. C.The graphs in figure 9are found on pieces of pottery vessels fromtheTa wenk'ousiteat LingyanghoinShantung province. Figure10showsapartialgraphseenonapotteryfragmentfromCh'ien chaifro.inthesameregion,whichseemsdearlytobeasecondoccur-rence of the fourthgraph of figure9. Thegraphillustratedinfigure11appearsonwhatisdescribedas"a flat-backedhuvase"foundataburialsheatPaDt'outs'unl: ml tf in Shantungprovince,another Tawenk'ousite(Cheung1983:328).There-markablething about this graphisthatit appears