origin and development of the paleohispanic scripts: the ... · revista portuguesa de arqueologia....

107

A B S T R A C T ThefirstattestedindigenouswritingsystemintheIberianPeninsulaisthe

so-called “Southwestern script”. I argue that this system is of Phoenician inspiration and

simultaneouslyofferanexplicativemodelforitsdevelopmentasanalphabet,lateronalso

the source of the Iberian semi-syllabary. This account, made primarily from a linguistic

standpoint,ultimatelyintendstodiscloseorthographicrulesofthescriptandphonological

featuresofitsunderlyinglanguage.

R E S U M O OprimeirosistemadeescritaindígenadaPenínsulaIbéricadocumentadoéadesignada

“EscritadoSudoeste”.Aqui,defende-sequeestesistemaédeinspiraçãofeníciae,simultanea-

mente,oferece-seummodeloexplicativoparaoseudesenvolvimentocomoalfabeto,consti-

tuindoposteriormentea raizdosemi-silabário ibérico.Esteensaio, realizadosobretudode

umaperspectivalinguística,temcomoderradeiropropósitoaveriguarregrasortográficasda

escritaecaracterísticasfonológicasdalínguasubjacente.

1. Introduction

ThreemajorwritingsystemswereusedintheIberianPeninsulainprotohistoricaltimes—thatis,fromtheendofBronzeAgedowntotheearlymomentsofRomanoccupationoftheterri-tory—towritelocalPre-Romanlanguages.TheyarereferredtogenericallyasPaleohispanicscripts.Theseweretheso-called“Southwesternscript”andSoutheasternandNortheasternIberian.

Withtwovarieties,theIberianscripthasbeenthusnamedinreferencetoanIronAgearchae-ologicalculturethatisassociatedtoabulkofPre-Romanpopulibroadlyknownas“Iberians”.ItsNortheasternvariety(henceforthNEIberian)wasusedinanareathatroughlycorrespondstotheSpanishLevant,whichisthereasonwhyitisalsoreferredtoas“Levantine”.AlthoughNEIberianhasbeendeciphered,itsunderlyinglanguage—alsoknownas“Iberian”—stilllargelydefiestrans-

Origin and development of the Paleohispanic scripts: the orthography and phonology of the Southwestern alphabet1

MIgUELVALéRIO*

REVISTA PORTUGUESA DE Arqueologia. volume 11. número 2. 2008, pp. 107–138

Miguel Valério Origin and development of the Paleohispanic scripts: the orthography and phonology of the Southwestern alphabet


lation.Ontheotherhand,theSoutheasternvariety(henceforthSEIberian)remainsforthemostpartincomprehensible.Whilesometextsappeartobeinthesamelanguage,itseemsitwasalsousedforsomethingotherthanIberian.EventhoughitisdifficulttoestablishchronologicallimitstotheuseofSEandNEIberian,mostscholarswouldagreetoplacethembetweenthefifthandthefirstcenturiesBC,i.e.intheIronAgeII.TheNEIberiansemi-syllabarywaslateradoptedbytheCeltiberiansofthehinterland,whoadaptedittotheirIndo-Europeanlanguage.Apartfromthese,itisalsoworthmentioningtheexistenceofgreco-Iberian,anIonian-basedalphabetusedtowritetheIberianlanguage.Fortheirgeographicaldistribution,seeMap1.

ThescriptknownasSouthwestern(henceforthSW)hasalsobeendubbed“South-Lusitanian”,“Tartessian”or“Bastulo-Tartessian”amongotherdesignationsthatappealtoethnicandethno--geographicboundaries.SincetheethnolinguisticsituationoftheprotohistoricalIberianPenin-sulaisstillpoorlyknown,hereImakeuseoftheconventionalgeographic-baseddesignation.ThecorpusofSWcomprisestodaynearlyonehundredinscriptions,thevastmajorityofwhichwerefoundinsouthwestPortugal(thusitsdesignation)intheregionsofBaixoAlentejoandAlgarve.Some,however,hadtheirfind-spotsintheneighboringSpain,namelyinthearealocatedbetweenExtremaduraandWesternAndalusia(Map2).

Thechronologyofthisscriptistroublesome.Despitetheexistenceofsomegraffitionpot-tery,mostofthetextswereinscribedinstone-madestelaewhosefunerarynatureisrecognizedbuttheseareusuallyfoundtakenfromtheiroriginalcontextandreusedasconstituentsoflatertombs,oftenbeingbrokenforthatpurpose.Buttheiroriginal functioncanbesurmised,andnot justbasedonhowfrequentlytheyturnupinnecropoles.ThestelefromAbóbadaI(Fig.1),forinstance,wasdividedintotwomainfunctionalsections.Thelowerpartoftheblockwasshapedtothepur-poseofbeinginsertedontheground,allowingthesteletostandvertically.Theremainingsectionfunctionedasthesurfacewherethewrittenmessageandtheiconographicmotifwerecarved.ThesamelogicappliestothesteleofMestras(Fig.2).Raisingwrittenanddecoratedfunerarystelaewas,ofcourse,across-culturalpracticeintheIronAgeMediterranean—inadditiontoSyriaandAnatolia, itwasalsopresentintheAegeanandItaly,amongotherregions—butintheIberianPeninsulaitneednottobetakenasacaseofdiffusion2.

Map 1 DistributionofthePaleohispanicscripts. Map 2 ApproximatefindspotsofSWinscriptionsupto1996(Correia,1996,p.162,Fig.15).

Origin and development of the Paleohispanic scripts: the orthography and phonology of the Southwestern alphabet

Miguel Valério

REVISTA PORTUGUESA DE Arqueologia. volume 11. número 2. 2008, pp. 107–138 109

Thedirectionofwritingisnotstable:itmaybeleftwardorrightwardandoftenfollowstheboustrophedontechnique.Textwasfrequentlyinscribedwithinstraightlines,calledcartelas(roughlytranslatableas“bands”).InthesteleofAbóbadaI,forinstance,theinscriptionconsistsoftwosec-tions,writtenleftwardsandrightwards,andonlytheleft-orientedportionofthetextiscontainedinaband(Fig.1).Word-dividersareveryrare;textno.35fromMestras(Alcoutim,Portugal)isanexception:wordsareseparatedbyverticalstrokes.

Fig. 1 SteleofAbóbadaI/textno.48–35x40cm(Correia1996,p.118).



Henceanyapproachseekingtoseparate lexemescanonlyrelyondistributionalanalysis,thatis,onecanonlyidentifythelimitstopossiblewordsbylookingforrepeatedsign-groupsindifferentinscriptions.Thescriptyieldsonegoodexampleofhowthismethodcanbeappliedtoundividedtexts.Thereisasequenceoftwolexemeswhichisrepeatedinaconsiderablenumberof inscribedstelae,accordinglychristenedas“funerary formula”. Ipresentbelow instancesofsign groups from some (not all) texts where those two lexemes can be identified and distin-guishedfromtherestofthecontent.Inoneinstance(text48,AbóbadaI),oneofthesigngroupsoccursalone.Theunderliningandverticaldividersareanartificialcreationofmine,meanttohighlightthelexemesinquestion.ForthesakeofsimplicityhereIstrictlypresentthesignsright-wards:

Fig. 2 SteleofMestras/textno.35–85x62cm(Correia,1996,p.105).


Miguel Valério


Text9: ...‡"9|Mòa9|nòwk9ni

Text13: ‡ÇZÇ|ia9ò|Mòa9|nòwk9nòi

Text15: ...i|Mòa9|b9

Text17: ...9ò|Mòa9|nòwk9nii

Text38: ..."9|Mòa9Mò|nòwk9n±i

Text47: ...òn"9|Mòa[9]Mò|nòwk9n±i

Text48: iaÇòL"Çsi9L|nòwk9n±i|sÇ...

Text54: s|Mòa9Mò|nòwk9n±i

This example is quite informative, as we are able to isolate two possible stems which fre-quentlyappearinconnectionand,inaddition,showtracesofsuffixationandevensomeagree-ment i.e. whenever stem 1 appears with suffix X, stem 2 usually (not always) comes next withendingY:

Stem1: Mòa9- > Mòa9-Mò

Stem2: nòwk9n- > nòwk9n-±i

Someauthors,readingk9niiandk9n±i askoniiandkonti, respectively,haveconnectedthesepresumedwordswiththeethnonymConii~Cynetes,thenameofaPre-RomanpeopleknownfromClassicalsourcesthatabidedinthesouthernareasofmodern-dayPortugal.Distributionalanalysis,however,showsthattheyareinrealitypartoftwolargerlexemesandverylikelyunrelatedtothatethnonym.

Somestelaewereretrievedfromfunerarycontextsinarchaeologicalexcavations.Archaeolo-gistswhostudied thematerialsdated thosecontexts inwhich theywere foundtobetweentheseventh and fifth centuries BC (Mederos & Cabrero, 2001, pp.101–103, with references) butbecausethestelaeinquestionhad,asarule,beentakenoutoftheiroriginalpositionsandreusedtobuildnewfunerarystructures,weknowthattheyarenecessarilyearlier.Forthisreason,itisdif-ficulttoestablishthetimelineofthewritingsystemcontainedtherein.

Thereispresentlynoholisticeditionofthecorpus.With81inscriptions,Correia(1996)wasthelastmostcompletecorpuspublished.InthisarticleIfollowthetextnumbersgiveninhiswork(Correia,1996,pp.167–168),whilepresentingsubsequentinscriptionsaccordingtotheirprove-nance,i.e.bymodernplace-name.

Amongtheinscriptions,onequitesingularobjectstandsout:theso-calledSignaryorTableofEspanca, found in the municipality of Castro Verde, in Baixo Alentejo, Portugal (Correa, 1990,p.132).Ithasnoarchaeologicalcontextandisundated.Thetable(48x28x2cm)isatwo-linewrit-ingexercisethatincludestwenty-sevensignsengravedintheupperrowandlessregularlyrepeatedinthelowerline.Forthisreason,itisoftenadmittedthatthefirstandupperrowductuscorre-spondstothestandardizedsystemaspresentedbyamasterscribetohisstudent,whoselessimpres-



siveletteringliesbelow.TheoutstandingcharacterofthisobjectmakesitofparamountimportancefortheunderstandingoftheSWscript—eventhoughtheabsenceofsomesignspresentinthestelaefromitssetofletterscloudsthenatureoftheirrelationship.Thenumberofsignsinthetableandthepresenceofvowelsshowthatthesystemrepresentedthereincanonlybealphabetic,whereastheorderofthefirstthirteenletterscloselyreflectsthatoftheWest-Semiticalephats.

The direct Phoenician derivation of the script is widely accepted (De Hoz, 1996, p.201;Rodríguez, 2000, inter alia), even though some proposals that reject it and favor a greek role(Schulten,1940;Untermann,1985,1997)—orbothgreekandAnatolianrole(Beirão,1990,p.118;gomes, 1997, pp.11–12; Silva & gomes, 1998, p.163) — in its design have been offered3. Thesourceofthewritingsystemwillbethefocusofsection2.

EvenmoreconsensualisthenotionthatallPaleohispanicscriptsareultimatelyinterrelatedandithasalwaysbeenassumedthattheyhaveacommonsouthernorigin(Correia,1996,p.8).Inotherwords,itisadmittedthattheSWscriptmighthavedevelopedintoSEIberian,whichthengavewaytoaNEvariant.

Regardingthedeciphermentstudiesonthescript,manyscholarsseemtoendorsetheviewthattheSWsignsmatchthoseofIberian,notjustinshapebutalsophonetically.Thedecipher-mentofIberian,namelyitsNEvariant,wasaccomplishedfromthe19thcenturyonwardsmostlywithbasisoncoinageandinscriptionsassumedlybearingonomasticelementsknownfromlatergreekandLatinsources,aswellasotherepigraphicmaterialbearingbilingualorquasi-bilingualLatin-Iberiantexts.Theresultwasacuriousmixed-systemofsemi-syllabicnaturewithsyllabo-graphicsignsforstopsandalphabetic(i.e.phonemic)onesforotherconsonantsandvowels(Med-eros&Cabrero,2001,p.97).Thissystemwasadvanced intheworksofgómezMoreno(1922,1943)andeventuallybecameconsensual.Thehistoryofdeciphermentandthelistofworksdeal-

Fig. 3 TheTableofEspanca(Correia,1996,p.22).


Miguel Valério


ingwiththenatureofPaleohispanicwritingsystemsarelong:IsubmitthereadertothedetailedaccountofMederos&Cabrero(2001,p.97),withrespectivebibliography.

WithparticularrespecttotheworkontheSWscript,thecurrentstateofaffairsderivesfromthepublicationofgómezMoreno’slastmajorworkinwhichheconsideredeveryprotohistoricalscriptofthePeninsulaasthesamesemi-syllabicsystem(gómezMoreno,1961,p.187).Aroundthesametimeanotherscholar,Schmoll,arrivedatsimilarresults:hepostulateda25-signsystemin which the 5 vocalic signs matched those of SE Iberian (Schmoll, 1961, 1962). Furthermore,SchmolladdedanotherpeculiardetailbynotingthataftereachsignthatmatchedanIberiansyl-labogram,aletterwiththecorrespondingvocalicvaluefollowed—e.g.tu‑u:

Southwestern SE Iberian

dÇ tu‑u d tu

Indeed,Correa(1990,p.136)pointsouthowsurprisingitwouldbeforsignsofboththeSWscriptandtheIberiansemi-syllabariestoyieldexactlythesamephoneticvaluesanddefendsresortingtointernalanalysisasamediumtoestablishingdifferences.OneoftheadjustmentsheadvocatesisSchmoll’sobservationthatsyllabogramsarealwaysfollowedbyalphabeticsignswiththecorrespond-ingvocalicvalue:e.g.ta,ti andtu arealwaysfollowedbya,iandu,respectively.Evenifoneassumes—assomeseemedtodo—thatthisissomesortofscriptio plenamechanism(similartothatofAncientNearEasterncuneiformsystems)usedtorepresentlongvowels(i.e.ta–, tı−, tu–),theresultisabizarresituationinwhichtheunderlyinglanguageonlyallowedstopsinfrontoflongvowels(!).This“vocalicredun-dancy”,asperCorrea(1990,p.136),is“asurprisingandexclusivetrait”ofthescript.

VowelsSyllabic (stop) signs

B K T

A a M c x4E 9 Bb k ‚I i U R ±

O ᚬ 0 g+ <>U Ç 3 Y d

Regardingmethodology,theabovementionedideathatthereisacompleteornearcompletecorrespondencebetweentheSWscriptandtheIberiansystemsisdebatable.Mostreadingspub-lishedinthe literature,namelythatofstopssigns(bV,kV, tV),aredirectcalquesoftheIberianconventionaltranscription(Fig.4)andtacitlyinsinuatethatthelanguagehadafullphonologicalcorrespondencewithIberian.

Thepreconceptionthatitmustbesoduetotheirmereformalresemblanceandrecognizedrelationshipissimplyuntenable.WeknowofnumerouscasesthatcounterthisassumptionbutIwillcontentmyselfwithone(whichwasincidentallyputforwardbyJ.deHozinthe10th Interna‑tional Colloquium on Paleohispanic Languages and Cultures,Lisbon2009).ThisisthecaseofMycenaeanLinearB,decipheredin1952byMichaelVentris.givenitsgeneticrelationshipwiththeCypriot

Fig. 4 Mostconsensualtableofvaluesforthesyllabicstopseriesandvowels(Rodríguez,2000).



syllabary(decodedsincethe19thcentury),whichweknowtodayultimatelyderivesfromMinoanLinearA,scholarsinitiallyattemptedtoreadLinearBwithbasisontheCypriotvalues.Thatturnedoutunfruitful. In fact,althoughrelated, theCypriot syllabaryandLinearBsharedonlyelevensignsthatmatchinformandonlyeightorninethatareequivalentinformandphoneticvalue(Valério,2008,p.62).ApartfromthefactthattheCypriotsyllabarymostprobablynotatedadiffer-ent language, this was certainly caused by the geographical and chronological gap between thescripts.Thepointtoretainhereis:thelessclosetwogenetically-relatedscriptsareinspaceandtime,thelessfruitfultheircomparisonwillbe.Thisadvisescautionintheapproachtothedevel-opmentofdifferentwritingsystemsintheIberianPeninsula.

Bearinginmindtheproblemshereintroduced,thisarticlesetsouttoadvancesolutionsforissuesrelatingtodecipheringapproachestotheSWscript,theirmethodsandtheirreadings.ThefirstpointneedingclarificationistheprecisesourceoftheSWscript:Phoenicianorgreek?Then,theforeseeablealphabeticoriginofthesystemengendersonefurtherproblem:theoutcomeofIberianasasemi-syllabarylacksa logical justification.Suchadevelopmentmightseem“unnatural”asasyllabaryrepresentsalessflexiblesystemwhencomparedtothealphabet.Iam,ofcourse,notclai-mingthatwritingisabsolutelyprogressive.Imerelyassertthat,incomparisontosyllabaries,alpha-betsallowtoexpress individualphonemesandahighernumberofphonemesequences (namelyconsonantclusters,fairlycommon,forinstance,inIndo-Europeanlanguages)withfewmorethantwentyletters,muchlessthanthoseofasyllabary.Wecannotforget,however,thatthechoiceofaparticulartypeofscriptislargelydependentontheinherentfeaturesofthelanguagebeneath4,andthustherearesomeexamplesofworldlanguageswhosesoundsystempromptedtheshiftfromanalephat/alphabettoasyllabary.Thiswillbeimportanttotheanalysisherepresented.

Indealingwiththegradualtransitionfromthesourcealephatoralphabettothesemi-syllabicIberian,Iwilldiscussasolutiontotheproblemsraisedbythecurrentinterpretativemodel(s)oftheSWscript;and,indoingso,IwillalsoadvancenewreadingstosomeSWsigns.Iwillrelyonbothinternalandexternalanalysis(i.e.comparisonofsignvaluesinboththesourceanddescend-antscripts),asalsooutlinedinRodríguez(2000).Tounderstandwhatstructuralmodificationsoccurinagivenscriptwhenawritingsystemistransferredfromonelanguagetoanother,onemust take in consideration the eventual differences in their phonetic inventories. In our case,knowledgeofthestructuresofthedonorandrecipientscriptandofthephonologyoftherecipi-entlanguagewillpromptthereconstructionoftherecipientlanguage’sphonology5.Therefore,themostimportantmethodologicalpointinmyapproachistheappealtoalinguisticframework,whichIbelievehasnotbeenfullyexploredyet.

2. The source of the Southwestern script

ThissectionaddressesthequestionoftheoriginoftheSWscript.IhaveacknowledgedabovethatPhoenicianisthemostconsensualcandidateforitssourceamongscholars,whereasgreekisthealternativeofaminority.Thestartingpointofthepresentanalysis,however,intendstobetabula rasa.

2.1. The paleographic evidence

Thefirstanalyticalparameteristhealphabeticalorderofthescript’ssigns.ThearrangementoftheTableofEspancacloselyreflectsthatofthePhoenicianalephat,albeitta–wcomesbeforewa–w


Miguel Valério


(Fig.3)—atraitoftengivenintheliteratureasagreekfeature.Theorderoftheletters,however,isaveryconservativeandenduringaspectofalphabeticwritingsystems:thearrangementofWest--Semitic alephats was already established in the Ugaritic cuneiform alephat of the second--millenniumBCanddidnotchangemuchuptoourdays.BecauseinUgaritic/w/cameafter/t/,Rodríguez(2002,p.193)fairlysuggeststhatthepositionofupsiloningreekmightbeinheritedfromaheterodoxSemiticsignarythatfollowedtheUgariticorder.Andinanycase,thechangeofpositionsbetweentwoneighboringlettersintheTableofEspancaisnonprobative—especiallywhenwemustconcludethattheabecedaryinthistableandtheSWscriptingeneralarepossiblynotrepresentativeofthesamescript.

Moresignificantconclusionscanbedrawnfromtheinspectionofthemorphologyoftheletters.Ingeneral,their linearityandverticality(cf.e.g.aandz)wouldsuggestagreekratherthanWest-Semiticorigin.However, it issymptomaticthatthreeletters,thosecognatetogreekbeta,iotaandtau,aremuchclosertotheirPhoeniciancounterpartsthananycorrespondingearlygreekform(Sass,2005):

SWTable of Espanca

(undated)

PhoenicianKulamuwa orthostat(late 9th century BC)

Early GreekDipylon jug

(mid 8th century BC)

notattested( elsewhere)

ThefinalbutessentialaspectisthevocalicsystemofSW.Wehaveseenthatthemostconsen-sualschemepresentedintheliterature,andbasedonthecomparisonwiththelaterIberianscripts(seebelowfordiscussion),isthatofapentavocalicsystemwiththefollowingreadings:

a 9 i ᚬ Ç

a e i o u

Proponentsofagreekoriginmustassumeadifferentconfiguration,necessarilyinterpreting9as/o/orasimilarvowel.Fromanihilisticviewpoint,onecanmakeanimportantobservation:º,theobviouscognateofgreekepsilon,behavesintheSWscriptnotasavowelbutasaconsonantalsign(thiswillbedemonstratedspecificallyinsection3).Anotherpointtobeaddressedbelow(see3.2)regardsinternal(besidesexternal)evidencesuggestingthat9representsafront/highvowel(e)andᚬ writesaback/roundedone(o),aspredictedinthetableabove.

Itisacceptedthatthefive-vowelschemeisaninnovationofgreek(eventhoughvowelsignshaveemergedindependentlyinotherwritingsystemsofSemiticoriginacrosstheworld),whereastheWest-Semiticalephatswerebynatureconsonantal.However,thesourceofvowelletterscanbetraced back to them. Ugaritic, Phoenician and Aramaic scripts all made use of a special devicecalledmatres lectionis (whichincidentallyledtothecreationofvowelsignsinotherSemitic-derived



scriptsoftheworld).Thisconsistedintheuseofsomeconsonantandsemivowel(moreaccurately,approximant)letterstorepresenttruevowels,namelyintherenderingofforeignonomastics.Par-ticularly,thePhoenicianalephatmadelimiteduseofthismechanismtowriteLuviannamesinCilicia(Sass,2005,p.139),wherethelocalIndo-EuropeanAnatolianlanguage(writteninAnato-lianhieroglyphs)wasspokenalongsidePhoenician.ItshouldbenotedthatEarlyIronAgeLuvianhadonlythreevowels,/a/,/i/and/u/,notatedinthefollowingmanner:

Letter Phoenician matres lectionis (Luvian onomastics)

’a–leph /’/(glottalstop) /a/

yo–dh /y/(semivowel) /i/

wa–w /w/(semivowel) /u/

greeks,whopossiblyborrowedthePhoenicianalphabetinCiliciaitselforintheneighboringPamphylia6andhadforehandknowledgeofthematres lectionis,pickedthesethreeWest-Semiticlettersandtheirsecondaryvocalicvalues,addingeando(laterontwomoreletters,forthecorre-spondinglongvowels):

Letter Phoenician Greek

’a–leph /’/(glottalstop) /a/

yo–dh /y/(semivowel) /i/

wa–w /w/(semivowel) /u/

he– /h/(voicelessglottalfricative) /e/

cayin /c/(voicedpharyngealfricative) /o/

Thispointisfundamental.TheSWscriptcontainsthesamebasicvowellettersusedinthePhoenicianmatres lectionis,butthesignsassumedforeandodonotcoincidewiththegreekones(Rodríguez,2000,p.26)7:

SW Phoenician Greek

9=/e/< cayin >o=/o/

ᚬ=/o/< ’a–leph(?) nocognate

Inshort,thesamecayinthatbecameomicroningreekmighthavedevelopedintoeintheIbe-rianPeninsula,whereas it isnotclearwhatthesourceofowas.Rodríguez(2002,p.192,n.10)ingeniouslysuggeststhatoisderivedfromarotated’a–leph,citingaparallelchoiceinthedesignoftheYiddishscript.Moreover,Phoenicianhe–,whichyieldedgreekepsilon,hadadifferentfateasa


Miguel Valério


consonantalsigninIberia(seesection3).Togetherwiththeshapesofletters,thissignificantvoweldivergencecorroboratesthecommonly-sharedviewthatthescriptisofPhoenicianinspiration—allweneedtoassumeisthatitunderwentindependentverticalization,aprocessnotunseeninwritingsystemsoftheworld:

Phoenician’aleph cayin yo–dh ’a–leph (?) wa–w

SWa

ae

9

i

io

ᚬu

Ç

2.2. The archaeological evidence

Archaeologysupportstheconclusionsdrawnfromthepaleographicalanalysisofthescript.TheadoptionofwritingintheIberianPeninsulaispartofaprocessof“Mediterraneanization”increasinglyintensifiedsincetheBronzeAge.BecausetheSWscriptisattestedwithsecurityfromthe7thcenturyBCon,itmustnecessarilydatefromthattimeorearlier.ThiscoincideswiththearrivalofPhoenician(namelyTyrian)tradersandsettlerstothePeninsula,placedbyscholarscon-sensuallyinthe9thcenturyBCorshortlyafter.

Arruda(1999–2000,p.259)conceivesthat“populationsofeasternorigin”weresettledintheareaofthegibraltarStraitfromthebeginningofthe9thcentury.TheappearanceintheSpanishterritoryofpotteryofSyro-Palestiniantypologies,namelyamphorae,datedequallytotheeighthorseventh-centuriesBCisprobative;theindigenoussettlementofElCarambolo(Camas,Seville)yielded(levelsD-IVandC-III)platesofthinrimandeasternfeatures,withCypriotparallels,alsofoundatthesiteofCastilloDoñaBlancaandtraditionallydatedtoatleastthemid-8thcenturyBC(Mederos&Cabrero,2001,p.106,withreferences).

IntheIronAgeI,theexistenceofsettlementsinthesouthwesternareasoftheIberianPenin-sulawhichareinterpretedbyarchaeologistasPhoenician,basedonthepredominanceofmaterialcultureofeasterncharacterandonarchitectural remainswithparallels in theSyro-PalestiniancoastinsiteslikeSantaOlaia,Abul,AlmarazandCerrodaRochaBranca,contrastswithasmallnumberfindingsofAegeanpotterysuchasthefragmentofakrate–rorpyxisfromCallesdePalos(Huelva),datedtoeithertheLategeometricI(760–730BC)ortheMiddlegeometricII(800–760BC),orafragmentofanEuboean(?)skyphosdatedtothethirdquarteroftheeighth-centuryBCfromCalledelPuerto9,alsoinHuelva(Arruda,1997,pp.39–40,44,58–59).

ItisnonethelessconsensualthateffectivegreekcolonizationoftheIberianPeninsulashoreswasundertakenlater.Weknowthatgreeks,Ioniansinparticular,preferablyestablishedthemselvesintheLevantinecoastofthePeninsula8,wheretheyleftbehindwell-documentedandarchaeologi-cally attested settlements such as Emporion (modern-day Ampurias), Rhode–(Rosas) and Mainake–(Málaga),amongothers.ThiswasthemotivationforastrongHellenicinfluenceinthecultureofIberians from the sixth-century on, manifest in such phenomena as the emergence of a graeco--IberianalphabetofIonianbase(IronAgeII).ItwasfurthermorethetriggertosomemodificationsintheNEIberiansemi-syllabary(seesection3).Evidently,theSpanishLevantwasanaturaltacticalchoice,sincePhoenicians(laterfollowedbyCarthaginiansinthesixthcenturyBC)werealreadywell-



-established in the south and southwesternshoresaheadoftheAegeancolonizingimpe-tus. In short, Phoenician, not greek, was athandtobeborrowedbylocalpopulationsofthesouthwesternPeninsulainIronAgeI.

IntheaforementionedCalledelPuerto(no.6),inHuelva(interpretedbyarchaeolo-gists as an “indigenous” settlement withroots in theLateBronzeAgeandnotasanexogenousemporium),oneamphorawithagraffito was found that deserves considera-tioninthissection.Theamphorawasrecov-ered from level IIa (= c. 650–600 BC) andbelongstothetypeVuillemotR-1(=Trayamar1).TheseareconsideredthetypicalwesternPhoenician amphorae of the first momentsofthecolonization,andwereusedfortrans-portingwine.Regardingthegraffito,fromasequence of at least three, possibly moresigns,onecanunfortunatelyreadwithsafetyonlythefinaloneduetotheconditionoftheobject.ThislastsignhasbeeninterpretedasaPhoenicianyo–dhintheliterature(Mederos&Cabrero2001,p.105,withreferences),but,inreality,itisunmistakablyakintotheiofanyofthePaleohispanicscripts.

WhiletheshapeoftheletteralonewouldnotallowdiscerningwhetherthiswasSEIberianortheSWalphabet,theformermaybeexcludedwithbasisonthechronologyoftheartifact.Onecanimaginea7thcenturytradingvesselarrivingataportofsouthwesternSpainwithacargoofdozensofamphoraecontainingwineandotherprovisions.Possibly,therewassomeintermediaryentityoperatingattheportofHuelvathatwasresponsiblefortheredistributionofproductsamongtheindigenousworld.Inscribingapersonalorplace-nameonthecontainerswouldbeasuitableman-ner of distinguishing which one was meant for whom or where. This is a procedure attestedthroughouthistoryandcross-culturally.Ultimately,Ithinkthisitemisofutterimportance:whileitisnotdirectevidenceoftheroleplayedbyPhoeniciansinthedesignofthelocalscript,itisaproofofearlycontactsbetweenWest-SemiticmerchantsandusersoftheSWalphabet.

Therefore, independent evidence (the sum of the paleographical and archaeological data)supportsthealreadyconsensualviewthatPhoenicianwasthesourceoftheSWscript.

2.3. Additional signs

ThealphabetoftheTableofEspancaandtheSWscriptaspresentedinthestelaeincludenotonlyabasiccoreofsignstakenfromthePhoenicianalephatbutalsoasetofadditionallettersthatarenotofWest-Semiticfashion.TheyarepresentedintheTableasfollows:

T U o Q Ü g

Fig. 5 graffitoonamphoraofthetypeR-1fromCalledelPuerto6,Huelva(FernándezJurado&Correa,1988–1989,p.131,Fig.2/1)anddetailwiththereadablesign.


Miguel Valério


Thesesigns,usedtocoverphonologicalgapsleftbythePhoenicianletters,couldbe1)theproductof indigenousdesign;2)back-formationsfromPhoenicianletters;3) inspiredbyotherwritingsystems.Wehaveseenthatinthecaseofo,alsopresentinthisextraset,thesecondpossi-bilityisprobablythecorrectone.

Itwillbeexaminedwhichofthesepossibilitiesapplytowhichextraletters,butwhilethistopicwouldbelonginhere,itmustbepostponedtosections3and4,whereimportantpointsthatrelatetoitwillemergefromthediscussionformulatedthere.

3. From an alephat to an alphabet and from an alphabet to a semi‑syllabary: internal reconstruction of SW

3.1. The Iberian language and the development of its semi‑syllabary

AfterperceivinghowthePhoenicianalephatwasadaptedtoafullalphabetwithvowelsintheIberianPeninsula,weneedtoclarifythemotivationbehinditslateradaptationtoasemi-syllabarywas. The key-point is Iberian itself. Understanding how this script related to the phonologicalfeaturesofitsunderlyinglanguageprovidesimportantinformation.

The Iberian semi-syllabary is considered as such because it possesses thirteen individual(phonemic)signsforvowelsandmostconsonants,butthreeseriesofsyllabogramsforstops—bilabials,velarsanddentals(Correa,1994,p.268).TheNEvarietyevendevelopedgraphicvariants(markedwithoneextrastroke)ofthesyllabogramsofthekandt-seriestorepresenttheirvoicedcounterparts(notethatIberianhadvoicedandvoicelessvelaranddentalstopsbutonlyavoicedbilabial):

% ba 12 ka (ga) C ta (da)

( be :9 ke (ge) HK te (de)

* bi ;∙ ki (gi) O∏ ti (di)

, bo @ ko (go) to (do)

/ bu B ku (gu) TU tu (du)

Ontheotherhand,thescriptwasnotequippedtoexpresssequencesofthetypestop+con-sonant+vowel(SCV).Ofcourse,thiscouldwellbenotthereflexofanactualfeatureofthelan-guagebutratheralimitationofthewritingsystem(Correa,1994,pp.278–279).But,again,whywouldthescriptacquirealesspracticalformandmisrepresenttraitsofthelanguagethataplainalphabet could accommodate? In reality, SC clusters are absent even from texts written in thegreco-Iberianalphabet(i.e.theuseoftheIonianalphabettowritetheIberianlanguage),fittoembodysuchconsonantalclusters.

Indeedstop+consonantsequencesweremissinginIberian,aspatentinthetranscriptionofforeignnameswithsuchclusterswithCVsyllabogramswhosevocalicvaluereflectseitheradeadoranepentheticvowel:



Latin > Iberian

Flaccus bi‑l‑a‑ke (§4.7.3)

Fabricius ba‑bi‑r‑ki (E.7)

Gaulish

Blandus(*blandos) ba‑l‑a‑n‑de (B.1.125)

Celtiberian

*Segobriga se‑ko‑bi‑r‑i‑ke‑s (A.89)

Fromaphenomenologicalpointofaview,itisworthanalyzingtypologicalparallels.Thereare some languages in the world today that likewise have no such consonantal sequences, likeJapaneseandsomePolynesiantongues.TheJapanesephoneticinventory,forinstance,isbasedonsyllabicsets.Apartfromfivevowels(a, i, u, e, o)andthenasalsound(n),allothersyllablesinthelanguageareconsonant+vowel.Consonantclustersinloanwordsarealwaysbrokenupwithvow-elsandwordscontainingafinalconsonantotherthannareaddedanecho-vowel,oftenooru (Kay,1995,p.69).Notetheexamplesbelow,withspecialemphasisonthetreatmentofclustersofthetypestop+consonant:

English > Japanese

fax fakkusu

Christmas Kurisumasu

club kurubu

present purezento

stress sutoresu

Conveniently, Japanese came to use two syllabic writing systems (apart from the Chinesecharacters—kanji)calledHiraganaandKatakana,thelatterbeingusedmostlyforwordsofforeignorigin(Akiyama&Akiyama,2002).ThebasiccoreofHiraganaismadeof55syllabograms.

ThecasewithIberianislessdramaticbecausethelanguageonlydisallowsclustersoftheSCV(stop+consonant+vowel)type;SVC(stop+vowel+consonant)sequencesareallowed.Inotherwords,whileJapaneseconsistsalmostentirelyofopensyllables(cf.Na-ga-sa-ki),Iberianhasalsoclosedoneswithnonstopconsonantsinsyllable-finalposition.Forinstance,awordlikea‑bar‑kis(textC.4.1)containstwoconsecutivesequencesofthatkind(V-SVC-SVC),whichmeansthattwoconsonantsmaycometogetheracrosssyllableboundaries(abaRKis)andformheterosyllabicclus-ters9.UnlikeJapanese,then,Iberianwaswrittenwithasystemcontainingsyllabogramsforstopsalone(i.e.SVsigns)andindividual(i.e.phonemic)lettersforconsonants(C),sincethesecouldbefollowedbybothvowelsandotherconsonants.Thissystem,asemi-syllabary,wasthebest-fittingfortheIberianlanguage.


Miguel Valério


Thisframeworkunveilsthemotivationforasemi-syllabaryinitslastformativestage,butwestillneedtoclarifyhowthechangeoperatedbetweentheborrowingofthePhoenician-inspiredalphabetandtheshapingupofSEandNEIberian.That“intermediarystage”needingelucidationistheSWscriptitself.

3.2. The formation of Southwestern script

AsalreadynotedbyRodríguez(2000,p.29),whatSWhasisaseriesofsignsderivedfromPhoenicianstoplettersbut,withoutexception,alwaysfollowedbyvowelsigns:

Phoenician SW script NE Iberian

letter value sign ante (vowel) syllabogram value

d d Ç u T tu

t. ± i i H te

t x a a C ta

g c a a f ka

k k 9 e : ke

k. R i i no cognate

WeobservethattheSWscripthasawholesetofsignsdescendantfromthePhoeniciandentalstopletters,eachusedalwaysbeforeaparticularvowel:thusT+A,T+I,T+U(whereT=dental).Ontheotherhand,thesamesignsbecamesyllabogramsofthet-seriesinIberian,afactthat,aswehaveseen, ledscholars (whobelieve that those signshad identicalvalues inbothscripts) toassume“doublespellings”ofthetypeta‑a,ti‑iandto‑o.Thesearephonologicallydifficulttoaccountforand,infact,unnecessary.

IsuggestthatthemotivationforthisphenomenonisthesameasinIberian.Apartfromthesecombinationsofstopsignsandvowels(S+V),theSWscriptusesindividualphoneticsignsforallotherconsonants(n, m, l, r, s,etc.)whichmayoccurbeforeanysign.Itisplausible,then,toassumethat,likeIberian,thelanguagewrittenwithSWlackedSCVsequences.

Becauseonlythreestopseriesexist(conventionallytranscribedasB,KandTasseenabove;Fig.4)wemayalsohypothesizethatthelanguagehadnovoiceoraspirationcontrast,thuscon-tainingonlyplainvoicelessstops(/k/,/p/,/t/)initsinventory.ThisissupportedbyIberian:despitewritingalanguagewithbothvoicedandvoicelessdentalandvelarstops(/k/and/g/,/t/and/d/),thesemi-syllabaryhadoriginallynovoicingdistinction10,theabovementionedvoicedvariantsofsyllabogramsbeingalaterinnovation11:



SWLanguage Writing >

IberianWriting Language

/p/ *P > B /b/

/k/ *K > K /k/~/g/

/t/ *T > T /t/~/d/

IfweassumethatthesoundsystemofthelanguagespokeninthesouthwestoftheIberianPeninsulaincludedonly/k/,/p/,/t/,weconcludethat,astheyimportedthePhoenicianalephat,itsabundanceofstopletters(d, t, t. , k, g, k. , b, p)becameredundant.Insteadofdroppingtheadditionalsigns,therecipientsofthescriptwouldthenhavedecidedtouseallofthemforthesamepurpose,placingeachofthembeforeonevowel.

Thereisstrongstructuralevidencetosupportthisifweattemptasortofinternalreconstruc-tionofthePaleohispanicscripts.LetustaketheexampleofthePhoeniciandentalletters(da–let,te–thand ta–w): inaccordancetothepresenthypothesis, theywouldhavebecomeredundantbecauseonlyonetypeofdentalstop(/t/)neededtoberepresented.Itfollowedthateachofthemwascon-nectedtoasingledistinctvowel,inaclearattempttoavoidanarchyinthescript—animportantprincipleofwriting.Thus:

d Ç t + u

± i t + i

x a t + a

Toprovethatthisdevelopmentisnotfarfetched,wemaycomparethegreekalphabetwhichimportedthethreesibilantsoftheWest-Semiticalephat:s.ade(/s./),samekh(/s/)andšin(/∫/).TheserepresenteddifferentphonemesintheWest-Semiticlanguagesbutbecausegreekpossessedonly/s/(Brixhe,2007b,p.26)thenewHellenicalphabetwindedupwiththreeredundantsignsforasingle sibilant. Thus s.ade became the letter san, used in some variants of the Hellenic alphabetbeforedisappearing,andsamekhwasusedwithkhiorkappainthecompoundsXΞorKΞthatrep-resentedthecluster/ks/,intimemerelyabbreviatedtotheletterksi(Ξ);onlyaftercenturiesdidsigma(<šin)becamethesolegreeksibilantletter.

So farwehaveaccountedfor threestop-vowelcombinations.Ofcourse, twoothervowelsexistedthatneededtoberepresented.Andsowefindthat<(evidentlyderivedfromd)occursbeforeᚬ,i.e./o/(texts11,19and62).Eventhishasalogicalexplanation:agraphicvariantofthestop-signconnectedtouwasusedforobecausethesetwovowelsaresimilarinnature,bothbeingbackandroundedvowels.Similarly,agraphicvariantofthestopletterusedbeforeiwasdevelopedfore,andthesetwoarefrontandhighvowels12.Thesequencej9appearsintexts11,20,35,48andpossiblyon thenewly found inscriptionofSabóia13.Thuswe find that the scriptwasadaptedbasedonalogicphonologicalscheme:


Miguel Valério


Front/high vowels ± t (+ i) > j t (+ e)

Back/rounded vowels d t (+ u) > < t (+ o)

Throughthismechanics,afulldentalstopsyllabicseriesemerged:

xa taj9 te±i ti<ᚬ todÇ tu

Withthisinmind,insteadofassumingdoublespellings,asotherscholarsuphold,Isuggestthesimplertransliterationta(i.e.t‑a),representing/ta/intermsofsoundperformance.

Rodríguez(2000,p.25,2002,pp.189–190)mentionsinpassingtheinterestingparallelofthege’ezscript(incidentallyalsodescendentfromaSemiticalephat:SouthArabian),thewritingsys-temofSemitic languagesofEthiopia,mostnotablyAmharicandTigrinya.ge’ezconsistsof26basicconsonantalsignsandasetof7diacriticalvowels,whicharecombinedtoformthecompositesigns. Each of those main signs represents a consonant + vowel (CV) combination and eachunmarkedconsonantalsymbolcanbecombinedwithanyofthesevenvowels.Word-finalconso-nantsandconsonantclustersaredenotedwiththeconsonantsignplusthediacriticofthevowel ,functioningasananaptycticvowel(Comrie,2009,pp.614–616).Rodríguez(2000,p.25)mentionedthissystemonlytonotethat“thevocalicsignsare[progressively]absorbedasappendicesbytheconsonantalonetothepointwherecombinedsignsareconfigured,makingthisscriptlooksyl-labic”.Unfortunately,whatthisscholardidnotfurtherunveil(tomyknowledge,atleastinthisarticleofhis)isthatoncemorethissystemisintimatelyrelatedwiththephonologicalnatureofthelanguage.Notsurprisingly,Tigrinyahasarichphoneticinventorywhosesyllablesmay,however,onlyformCVorCVCsequences.Whenthreeconsonantsoronedoubleconsonantandasimpleonecometogetherinwords,clustersarebrokenupwiththeinsertionofananaptycticvowel ;likewise,whentwoconsonantsoradoubleonewouldturnupinword-finalposition,iemergesafterthem;finally, if this is caused by a suffix, the same anaptyctic is inserted before it (Rehman, 2007).Amharic,ontheotherhand,allowsC+r/lclustersinword-initialpositionasingra ‘left’andblen‘pupilofeye’,buteventhesemaybeunderstoodasgi–‑raandbi–‑len (Comrie,2009,p.596).

OnemayalsocompareKharos.t.hı−,agandharianscriptofAramaicoriginthatconsistedofacoreofconsonantalcharacterstowhichdiacriticalvowelsignswereappended.Thevocalicdiacrit-icswereappendedtoconsonantsinspecificpositions,whichrevealsthattheinventor(s)ofthesystemhadsomeknowledgeofphonology:thuse and i (frontvowels)areplacedonthetopoftheconsonantalsignandoandu(backvowels)areaffixedbelowit(glass,2000,p.13).ThislogicfindsadirectmatchintheSWdesignofthedentalseries.

Itisimportanttostress,atthispoint,thattheEthiopianandIndianscriptsarenotsyllabarieslikethetwoJapanesekana,noreven semi‑syllabaries:vowels,albeitconnectedtoconsonantalsigns,are still identifiableas independent fromthemas inany alphabetic system.This sameprincipleappliestotheSWscript.

Thereorganizationofthestopsignshereproposedisconfirmedbytheothertwoseries.WehaveseenthatWest-Semiticg,kandk. gavewaytothecombinationsk+aandk+eandk+iinSW.



Then,withthesameschemeseenaboveforthet-series,thepredecessorofIberianko(whichappearsintheTableofEspancabutisofobscureorigin)isalwaysusedbefore/o/,whileagraphicvariantofitformsthecombinationk+u:

Phoenician SW script NE Iberian

letter value sign ante (vowel) syllabogram value

g c a a f ka

k k 9 e : ke

k. R i i — —

— — g+ ᚬ o @ ko

— — Y Ç u — —

Aswemayobserve,RwasnotpresentinNEIberiananymore(itdidexistinSEIberian,butthereitsvalueisuncertain).Thereisprobablyagoodreasonforthis.WeknowthatlateronaevolvedtorinIberian,becomingidenticalwithther.Thelatterthenhadtobedisambiguatedandwaschangedintow.SincethisnewshapewasnowsimilartoR,itispossiblethatthesignhadtobereplacedtoavoidconfusion.Forsomeobscurereason,Yhadnocontinuationaswell.

ThemostpoorlyunderstoodSWseriesisthatofthebilabialstops—whichIwilltranscribeasp+Vinaccordancewiththeworkinghypothesishereendorsed.OnlytwosignshaveacandidlyidentifiableevolutionfromPhoenician,whichisonlynaturalbecausetwoisthenumberofbila-bialstopsrepresentedinthatalephat(West-Semiticlanguagesdidnothaveanemphaticbilabial*/p./).ThusthesignsderivedfromPhoenicianbethandpeyieldedp+eandp+o(Correa,1990,Fig.3;Rodríguez,2000,p.31)respectively:

Phoenician SW alphabet

letter value sign ante (vowel)

b Bb 9 e

p 0 ᚬ o


Miguel Valério


Moreover,NEIberiansyllabogram0(bu)isthedescendantofp+o.Regardingshape,itappearsthattherectangularSWpowasformedwhenitbecamemuddledwiththelowerrowofthebandsofthestelae(Rodríguez,2000,p.27).

OnetrickycaseinthediachronyoftheIberianwritingsystemsisº.Ontheonehand,thisaprioriisthedescendantofWest-Semitiche.Ontheotherhand,thehomomorphicsigninSEIbe-rianhasbeeninterpretedbyFaria(1992,p.41)asbeonthebasisofthepresumedidentificationoftwoIberianpersonalnames,bersirandbekor,ontwoinscriptions(Untermann,1987,pp.293,299–300,306)aswellasthesupposedsimilaritywithsignbi(DeHoz,1976,p.300).Moreover,coinsfromAlcácerdoSal(SWPortugal)datingtothesecondandfirstcenturiesBCcontainthefollow-inginscriptionintheSEvarietyofIberian(Faria,1992,p.39):

™/ß´û

Fariareads,fromtherighttotheleft,be‑u‑i‑bu‑m(eventhoughheseemstohesitatebetweenmandninthelastsign).Thiswouldbeanindigenousplace-namethat,likesomanyinthesouth-westernareasoftheIberianPeninsulapreservedinRomansources,endsin-ipo(cf.Olisipo, Sisipo, Collipo,etc.).Inreality,thereading-i‑pu‑n,oreveni‑po‑n, ismorelikelyasitconformstotheformofthesuffixknownthroughLatintransmission—notethatSEIberianprobablywrotenotIberianbutanotherlanguageinWesternAndalusia.Itakethelastsignoftheinscriptiontobeavariantofn,notm,whichisfeasibleonmorphologicalgrounds.

ThustheevidenceofSEIberianconfirmsthereadingofSW0aspoand,mostimportantlytothepoint,substantiatestheideathatitsownûisbe.Basedonformalresemblance,someauthorswould not hesitate to attribute the same phonetic value to its SW predecessor. The sign doesbehavelikeastopletter,butitalwaysoccursbeforea(texts19,48,67,75),note,andsoinalllike-lihooditfillsthegapofpa—asolutionwhichisinharmonywithitslaterSEIberianvalue14.Inthiscase,therecipientsofthescriptsimplygaveanewusetoasignthathadnone(i.e.thelanguagehadnosoundequalorclosetoavoicelessglottalfricative),usingittofill inagapinthemostdefectiveseries.NotethatintheTableofEspancaºdoesnotappearinitsoriginalPhoenicianposi-tion:ithasbeenmovedforward,closertothesetofadditionalsigns.

As for pi, itmightbe representedby thehapaxÖ,whichprecedes i in text35 (Fig.2). InRodríguez(2000,p.44,S-306),thesignisgivenas:

ThefactthatitappearshorizontallyandthescratchesintheinscribedstonepreventedthisscholarfromidentifyingthissignwiththeoneinthepenultimatepositionintheTableofEspanca (Ü),whichappearsvertical. ItshouldbenotedthatthisisoneoftheextrasignsintheTableofEspanca.Iwillnotinsistonthisproposition,however,sincetheonlytwopointsinfavorofitaretheremoteaffinitytosomeinstancesofNEIberianbe(seetablebelow)andtheneedtoeliminateanemptyslotinthebilabialseries.

Presently,Ialsocannotofferasolidproposalforpu.Thecreativelogicoftheothertwoseriesofstopswouldsuggestagraphicvariantof0.Indeedwehavesimilarcharacterslike1and2,butthesedonotseemtobehavelikestopsigns(seesection4).SoIleavetheholeintheseriesunfilledforthemomentbeing:



SW SE Iberian NE Iberian

º a pa ã ba % & ba

Bb 9 pe û be (?) ( ¤ be

i pi (?) uncertain uncertain * + bi

0 ᚬ po uncertain uncertain , . bo

? Ç pu / bu / bu

ThetableabovesuggestsSWpa and pe switchedplacesinSEIberian,eventhoughthemotiva-tiontothischangeisnotclear.Inanyevent,thedataseemtosupportthereadingsherearguedfortheSWsigns.

Stillwithregardtothebilabialseries,thismodelcontrastswiththeworkofRodríguez(2000,p.36),whoproposesthatM=baandf=bi(?).Hisanalysisentailstwoproblems.UnlikethoseSWsignswhosePhoenicianpredecessorsandIberiansuccessorshaveassuredstopvalues,thesetwolettersarenotusedexclusivelybeforeonevowel:Moccursnotonlybeforea(severaltimes,inthesocalled“funeraryformula”)asRodríguezclaims,butalsobeforee(text64);andfbeforea,e(text35)andi(texts11,25,42,64,75).ThisprinciplehasbeenlargelyignoredinpreviousapproachespossiblybecausescholarsareattachedtotheideathatstopsignshavethemselvesaCVvalue.Inanyevent,whenappliedtotheSWsigns,themodelherepresentedisproductiveandprovestosolvesomereadingdifficulties.

WithrespecttoM,alternativereadingsputforwardintheliteraturedefineitasasibilant(e.g.,gomes,1997,p.12).OneofthepointsinfavorofitisthefactthatNEIberianhasahomomorphicsignwiththevalues (Siles,1979,p.81).Asibilantwouldbeagoodsolutionforasignthatisfol-lowedbydifferentvowelsinthecorpus.Ontheotherhand,Misalsothebestcandidateform;Cor-rea(1990,Fig.3)madeasimilarproposal.Andsince/m/isarecurringphonemeinlanguagesoftheworld,itisnotlikelythatthescriptwouldlackit.ThedecisiveargumentisprovidedbytheTableofEspanca,wherethesignplacedinthepositionofWest-Semiticmemissimilartothatsameletter(Fig.3),whichseemstoconfirmitsvalueofalabialnasal.

Ofcourse,thispresentsoneobstacle:thediscontinuityrevealedbythehomomorphicsignthatisasibilantinNEIberianandthedifferent-lookingsignforminthesamescript.

NE Iberian

s m

{Å|} i

WemustnotexpectNEIberiantoshowfullcontinuity(andweknowitdoesnot)becausecertainlythephoneticinventoryoftheIberianlanguage(IronAgeII)wasdifferentfromthatofthe


Miguel Valério


SWscriptlanguage(IronAgeI).Onegoodexampleofthisisthereplacementof9foreasthesignfore,whichcertainlyhappenedduetotheinfluenceoftheepsilonofthegreekalphabet.NEIbe-riandevelopedinthemodern-dayCataloniaaroundthe5thcenturyBCwhen(asdiscussedinsec-tion2)severalgreeksettlementshadbeeninstalledintheregion.Onthewhole,theNEIberiansemi-syllabarywasunderstronginfluenceofthegreekalphabet,sobothitseandsmayhavebeeninspiredbyepsilonandsigma.

Inshort,everythingsupportsthereadingm,ultimatelyconfirmedbytheWest-Semiticpre-cursorofthesign:

Phoenician mem SW (stelae) SW (Espanca)

AsperU,itisclearthatitisalsoaconsonantotherthanastop.Iwillforthatreasondiscussitinsection4.

3.3. Two possible exceptions?

Iwishtoclosethissectionbyaddressingtwoinscriptionsofthecorpusthatappear,atfirstsight,todefytheruleofstop+vowelcombinationwhichIhaveadvocated.ThesearethestelaeofAlcaládelRío(fromSeville,Spain;text75)andBenafim(Loulé,Portugal),thelatterpublished(gomes,1997)aftertheholisticeditionofCorreia(1996).Accordingtotheeditio princeps,thestelafromBenafimincludestwicethecombination0ᚬinaccordancewiththeorthographicrule.Butthenastrangesequence(aR0Ç,fourthrow)andahapaxlegomenon( ,firstrow)aredocumentedinthedrawingsandphotograph(gomes,1997,figs.2–4).Thecasewiththehapaxlegomenonmaybesolveduponacloserinspectionofthephotographprovided:

Fig. 7 ThefirstrowsofthestelefromBenafim(gomes,1997,Fig.3)andadetailofthedoubtfulsign.



Thehapax appearstobeactuallyabadlyerodedᚬ—thisisalsothereadingproposedbyFaria&Soares(1998,p.156).Thedifficultsequenceinthefourthrowismoreproblematicduetotheconditionofthewrittensurfaceofthestone.Inmyopinion,regardlessofthebadconditionoftheinscriptionatthispoint,onemayreaditas*aR1Ç,whichwouldnotbeinconsistentwithanyruleofthescript:

Inanycase,howeverthereadingofthistextmaybedubious,itisalso,inmyopinion,insuf-ficienttodiscardtheextremeregularityoftherestofthecorpus.

Theonlygraphictestimonyoftheotherproblematicaltext(75)whichIcouldfindis1961drawingofgómezMoreno(withnoscale)intheworkofCorreia(1996,p.145):

Fig. 8 Thefourthrowofthesteleandtheproblematicsequence(followinggomes,1997,Fig.2).

Fig. 9 Text75fromAlcaládelRío(Correia1996,p.145,accordingtogómezMoreno,1961).


Miguel Valério


Aswemaysee,perfectlyregularsequencesarepresentinthetext(dÇ, ºa,±i,etc.)butsomeothersareintriguing:e.g.gappearsbeforeastopsigninoneinstanceandonceitisevenisolated(!).DespitesomeproblemsinthedrawingofgómezMoreno—forinstance,thesequencedncanonlybeamistakefordÇ—therearesomefeatureswhicharenoteasilyexplained.Nonetheless,thisinscrip-tion,astheoneabove,stillincludesaconsiderablenumberofregularsequencesthatconformtotherulesoutlinedinthissectionandinultimateanalysisnoneofthetwoendangerstheirvalidity.

3.4. From the SW script to the Iberian semi‑syllabaries

Ihopetohavedemonstratedinthissectionthatthreesets(dental,velar,bilabial)ofS+VsigncombinationswereusedasaruleintheSWscript.ItwastheexistenceofstopsignswithidenticalconsonantalvaluesbutbeingusedincombinationwithdifferentvowelsthatpavedthewaytothedevisingoftheIberiansemi-syllabary.Verylikely,theadaptorsoftheSEIberiansyllabaryspokealsoalanguagelackingstop+consonantsequences—thisscriptwasassuredlyusedforIberianinalaterstagebutwedonotknowyetwhatlanguageitwroteintheearlieststages.Bythen,theymusthavefoundthatithadbecomepurposelesstokeepaddingvowelsignstothestopones.Tobesure,thesyllabicvaluehadalreadybecomeintrinsictothem.Thatis,itwasalreadyevidenttothereaderthatthecross-likeletterwastawithouttheactualvowelnexttoit.Itispossiblethattheinscriptionnr.81conventionallyattributedtotheSWcorpusbutfoundinCañamero(Cáceres,Spain),isatokenofthistransitoryphase,ifnotalreadyoneofthefirstexamplesoftheSEIberiansyllabary:

4. Non‑stop consonant letters

TheSWsignthatderivesfromPhoenicianh.e–t(avoicelesspharyngealorvelarfricative—thetwoCanaanitephonemesmergedinPhoenician;Woodard,2008b,pp.86–87)hasagreatnumberofgraphicvariants(givenasS-201–205,S-308andS-309inRodríguez,2000)withdifferentkindsofextrastrokes.Allinall,thesignprecedesanumberofdifferentvowelsandisattestedbeforeaconsonantonlyintheproblematicaltextfromAlcaládelRío.Evenso,itdoesnotbehavelikeatypicalstopsignandthereforeitmuststandforanothertypeofconsonant.Avoicelessvelarfrica-tive(/x/,whichIwilltranscribeashypotheticalh)orasimilarsoundispossible,takingintoaccountthevalueoftheoriginalPhoenicianletter.Sinceoneofthefifteensignsinthethreeseriesofstops,pu,isstillunidentified,itneedstobeverifiedwhetheragraphicvariantof0—whichwouldhave

Fig. 10 Inscriptionno.81(Correia,1996,p.151,withreferences).



becomeambiguouslysimilartotheletterinquestion—wasusedassuch.Inanyevent,theplenti-fulvariationsofthesignareworthexamining:

"+ e(text34,47,54,60,61),o(text75),u(texts9,27)3 + e(text15;Sabóia),u(41,48,71)4 + a(text15,25,35),e (text15),o(text51),u(51)1 + e(texts10,26) 2 + e(text9),o(text17)∙ + a(text23)

Mostofthevariantsoccurwithmorethanonevowelandallofthemseemtobeformsofasinglesign.Inalllikelihood,isanonstopconsonantalsignthatunderwentexceptionalstyliza-tionwheninscribedinstelae.Thisiscorroboratedbythenew(yetunpublished)stelefromMesasdoCastelinho(Almodôvar,Portugal).Thisnewitemcontainsthelongesttextfounduptothedate(nearly90signs),butitcontainsratherunusualandhighlystylizedvariantsofwell-knownsigns.Onemuststillawaitfortheeditio princepsbut,sinceIhadalreadytheprivilegetoinspectthestoneintheAlmodôvarSouthwesternScriptMuseum,Imayreportthatitcontainsthefollowingembel-lishedvariantof‚ beforeu:

Apartfromit,thetextalsoincludes1beforeu,eanda,whichreassuresthehetero-vocaliccharacteroftheletter.

Anotherproblematicsigniswwhich,followingotherscholarswhocompareittoasimilarsigninIberian,Itaketorepresent(possibly)somesortofliquid,transcribedasr —onlytodistin-guishitfromr,justassandz,whomusthaverepresentedtwodifferentsibilants,aretransliter-atedsands,respectively.

AletterthatalsodeservescommentisU,whichwehaveseenisanon-stopconsonantalsign.Thisarrow-likeletterhasaparallelinPhrygian,whereanhomomorphicletterrepresentsthepala-talizedoraffricate(t∫ or ts?)thatresultedfrom*/ke/~*/ki/andaccordinglyappearsinfrontofeandiinthewholeofPhrygia(Brixhe,2004,pp.26–27;Adiego,2004,p.302).ThisPhrygianlettercontinuedinlaterAnatolianalphabets:thesamesigninLydianistransliteratedascandprobablystandsforanon-palatalizeddentalaffricate/ts/(Yakubovich,2009,p.45); inCarianthesoundwrittenwiththissign(transliteratedasτ)is“somekindofcoronalobstruent,probablyanaffric-ate”butitsprecisevalueisstillunknown(Woodard,2008a,pp.57–58,66).ItisacceptabletodaytothinkthatgreekwasthesourceofthePhrygianalphabet—andthusallAnatolianalphabets—becausetheysharethesamevowelscheme(i.e.epsilonandomicronfromPhoenicianheandcayin),eventhoughthelatterisattestedatleastatthesametime,ifnotearlier(seenote3).EvidencetodaysuggeststhatthegreekalphabetprobablyfirstemergedinsoutheasternAnatolia,fromwhereitwouldhavebeentransmittedtoPhrygianareas(seealsoabove).Thearrow-likeletterwasaPhry-gianinnovationandInowthink,asCraigMelchertsuggeststomeinapersonalcommunication,thatthisinnovativeAnatoliansignwasinspiredbytheLuvianhieroglyphzi/abecauselocalAna-toliansneededtorepresentasoundabsentfromgreek,whosealphabetwasthemodelfortheirs.Indeed,findingsofPaleo-Phrygianalphabeticinscriptions(Brixhe&Lejeune,1984)andLuvianhieroglyphicones (Hawkins,2003,pp.142–143,map4)overlap in someareasof south-centralAnatolia.


Miguel Valério


Since Phoenician presence is also attested in Cilicia,where monumental bilingual Luvian-Phoenician inscrip-tionsliketheonefromKARATEPEwereerected,itispossi-ble, thoughverydifficult toprove, thattheywereawareofthe existence of such a sign. If there is any relationshipbetweenthissignandtheAnatolianone,asortofvoicelesspalatalized(/ty/)oraffricate(/ts/)dentalispossible.Itis,inanycase,suggestedbyitsappeareanceinanelementoftheso-calledfuneraryformula, nawk9nUi= narkenUi(text42).Sincethisisarareoccurrence,itmaybeaspellingvariantofnawk9n±i=narkenti (texts13,31,38,47,48,54), i.e.thesignmayrepresenttheoutcomeofearlier*/tyV/.Buteventhisisofdoubtfulvalidity,sincewehaveseenintheIntroductionthatnawk9n=narken-isastemthatmaytakedifferentsuffixes.Iwilltranscribethisashypotheticalzbut,duetothefragilityofthisproposition,Iwillnotinsistonit.

WhilethereisanevidentmorphologicalsimilaritybetweensomeofthelastsignsoftheTableofEspanca(T,U,Qandg)andsomelettersoftheAnatolianalphabets(apparentlyalllaterthanSWexceptforPhrygian),theircompatibilityintermsofsoundisundeterminable—correspond-enceismuchmoreoverwhelmingatbothlevelsbetweenthebasiccoreofthescriptandPhoeni-cian.Andoneshouldbearinmindthatwhensuchbasicandlinearshapesareinquestion,theyoftencanemergeindependentlyindifferentwritingsystems.

5. Rare letters, hapax legomena and graphic variants

Thereisalsoanumberofhapaxlegomenaandraresignsthatmustbegraphicvariantsofothersigns,lapidaryerrors,misreadingsorhaveyetanotherexplanation.Otherwisetheamountofsignswouldsurpassthatofaregularalphabeticsystem.Sothesedifficultsignsneedtobeexplained:

S-105 ñandS-303(invertedñ)S-105isattestedontext26.Thesemustbegraphicvariantsofz.

S-121îEitheramistakeoravariantof w.

S-301ëThisoccursintext38,whereitappearstofunctionassomesortofword-divider(seeFig.12).Theimportantsectionoftheinscriptionmayberead:poti * anakerto...andsoon.

Fig. 11 Anatolianhieroglyph*376zi(left)anditslikelyalphabeticsuccessor(right).

Fig. 12 Detailoftheinscriptionno.38,MealhaNovaI(Correia,1996,p.108).



S-3028Adoublingofq?

S-304ÇItappearstobeanerrorintext59,whereperhapsitwasmeanttober:

S-311Thissignoccursintext67,betweenaande.Mightitbetheresultofanill-orientednthatbecamedoublewhenthesculptorattemptedtoamendit?

S-312}It appears only on an inscription foundnearthePaleochristianbasilicaofMértola,Portugal(Faria,1994),beforee—itispre-ceded by one vowel and followed by two.Asequenceoffourvowelswouldberatherunlikelysoitmustbeaconsonant.Wefindthattheshapeofthesignandthefollow-ingeguaranteeitasavariantofk,k(e).

S-305òandS-313ÏS-305occursintext64andS-313isusedtwice in the inscription from Mértola. ItappearsalwaysbeforeiandispossiblythatitisadoubletofU(whichismostlyattestedbeforethisvowel).

S-314;Theoneinstanceofthissignisintext28.Acloseinspectionoftheinscriptionrevealsthatthisisactuallylandtheextra“leg”inthedrawingisactuallypartofascratchinthestone.Thelineinquestionthusreads:]uarh(?)oli[15,whichisasign-grouprepeatedintext17.Wecannotignore,however,that;isindeedattestedintheTableofEspanca.ThisisconnectedtothequestionofwhethertheTableofEspancarepresentsadifferentwritingsystem(sinceitlackssomeSWsigns),aproblemwhichremainsunsolved.

Fig. 13 Detailoftext59,fromgavião(Correia,1996,p.129)

Fig. 14 DetailoftheinscriptionofMértola(photographoftheauthor).


Miguel Valério


6. Final considerations

Along these pages answers to questions raised in the introductory section were sought.Althoughwedepartedfromatabularasastartingpoint,ourfirstinquiryinrealityconcernedaconsensual idea requiring confirmation: we conclude that all evidence discussed above, paleo-graphicandarchaeological,supportsaPhoenicianoriginoftheSouthwesternscript.

Phoenicianh. e–t

šin? sa–mekh nun me–m la–medh reš

SWh(?)

“"1

s

z

s

s

n

n

m

M

l

l

r

r

Phoeniciangı−mel kaph k.o–ph he– be–th pe– ta–w t.e–th da–leth

SWk(a)

c

k(e)

k

k(i)

R

p(a)

º

p(e)

b

p(o)

0

t(a)

x

t(i)

±

t(u)

d

IthadbeenassumedpreviouslythatbecauseSWwaspentavocalic,itwasnecessarilyderivedfromgreekratherthanPhoenician.Inreality,however,weknowthatSemiticalephatswerethesourcetowritingsystemsthatdevelopedvocaliccomponentsindependentlyindifferentregionsoftheglobesuchasEthiopiaandIndia.AsalreadynotedbyRodríguez(2002),thishasbeengreatlyignored,perhapsbecauseofasomewhatprevailingEurocentricviewthatfocusesmainlyonthehistoryofwritingfromtheCanaanitealephatdowntotheLatinalphabet.

Fig. 15 Inscriptionno.28,AmeixialIII(Correia,1996,p.98).



Consequently,werealizethatSWaddedtwovowelstothethreebasicanduniversalvocalicsignsofthePhoenicianmatres lectionis,andtheselectioncriteriaforthosetwolettersiscompletelydivergentfromthegreekoptions.Besidesfivevocalicletters,thesystemofSWcontainedtwenty--four consonantal signs, fifteen of these representing the special category of stops (one is stillunidentified).These fifteen signs belonged to three different series that corresponded to velar,bilabialanddentalstops.Eachserieshadfivesigns,oneusedalwaysandexclusivelyincombina-tionwithoneofthevowelsinthesystem:thislaidthefoundationsoflaterIberianstopsyllabic(CV)signs.Theremainingnineconsonantsignsthatdidnotdenotestopswereusedfreelybeforeanysign.Theresultisthefollowing29-signsystem:

Vowel signs Non stop consonant signs

a a s s n n

9 e z s ś M m

i i U z(?) l l

ᚬ o " h (?) r r

Ç u w r (?) — —

Stop signs

ca ka ºa pa xa ta

k9 ke b9 pe j9 te

Ri ki i pi(?) ±i ti

+ᚬ ko 0ᚬ po <ᚬ to

YÇ ku (?) pu dÇ tu

Asarule—andwehaveseengoodexamplesofthat—writingsystemsaredevisedtoconformtoasmuchaspossibletothephonologicalprofileofthelanguagestheyexpressandthusthereis,toagreatextent,a relationshipbetweentypologyof languageandtypologyofscript.Butevenwhenonesystemisadaptedtowriteanewlanguagethatisgeneticallyandtypologicallyunrelatedtotheoriginalone,whichnecessarilyleadstoadaptations,thenewadaptedformofthescriptwillinevitably preserve certain traits that denounce the previous state of affairs. This has been theprincipleappliedhereand,inthecaseofSWandsubsequentIberiansemi-syllabaries,itledtotheinference of some phonological features that accounted for exceptional orthographic rules inthesesystems.Thisstrategyyieldsasortof“scriptinternalreconstruction”.


Miguel Valério


Inourcase,thisexerciseleadsustotheconclusionthatvoiceandaspirationwerenotdistinc-tiveinthelanguageoftheSWscript,whichwouldhavepossessedonlythreeplainunvoicedstops:/k/,/p/,/t/.Thismeans,onetheonehand,thatdifferentstopsignsfromthePhoenicianalephatwerereorganizedinonlythreebasicstopsets(velar,bilabialanddental);ontheotherhand,itjus-tifiestheeffortofIberiantomarklaryngealfeatures(i.e.voicing)inthevelaranddentalsyllabicsigns,whichculminatedinthe(re)creationofgraphicpairsforvoicedandvoicelesssyllabogramswhichhadbeenlostinSW.

Moreover,thealignmentofstopsignswithindividualvowelsreflectstheinexistenceofclus-tersofthetypeSCinthelanguageofSW.Thelatterhadprominentconstraintsonsyllablestruc-tureandtautosyllabicconsonantclustersmusthavealwaysbeenbrokenupwithananaptycticvowel; nevertheless, as in Iberian, heterosyllabic clusters are possible: cf. the aforementionednarken-oruarman(hereintransliteration),anothermuchrepeatedsign-group(texts11,25,51,61,63).

Suchphonological features(lackofvoicecontrastandheterosyllabicclusters)are impor-tant to our knowledge of the language, especially because they suggest a non-Indo-EuropeanlanguageandresembleourpictureofIberian16.Butforthemomentbeingwemustbecautiouswithsuchconsiderations,becausethesetraitsarenotstraightforwardindicationofgeneticaffil-iation.Suchfeaturescouldbedevelopedindependently,sometimesresultingfromarealcontactbetweenunrelatedlanguages.ThussomeIndo-Europeanlanguages,likePersian,lacktautosyl-labicclusters;andithasbeenproposedthatHittite,anotherIndo-Europeanlanguage,hadnovoicedistinctionbutmerelyallophonicvoicing (Kloekhorst,2008,pp.21–25),possiblyduetosubstratuminfluence.

Finally,IwouldliketounderlinethatthesystemIproposeisnotmuchdifferentthanthatalreadyadvocatedbyRodríguez(2000,2002)—whoseworkIcameacrosswhenthewritingofthistext was already ongoing. Our proposals diverge, however, with respect to the strictness of theorthographicrulesandthedistinctionbetweenstopandnon-stopconsonantletterswhich,asaconsequence,producedifferentreadingsforsomesigns.Themostsolidcaseisthe“serpentine”sign,whichIreadasm basedonsolidindependentevidence:1)theshapeofPhoenicianmem;2)thepositionofthesignintheTableofEspanca;3)itsusebeforemorethanonevowel;4)itsabilitytosolvetheproblemofabsenceofalabialnasalinthescript.Noneoftheotherproposalsmeetsthesecriteria.Moreover,Iofferaworkinghypothesisthatjustifiesandsubstantiates,fromalin-guisticstandpoint,therare(butnotunseen)developmentofwritingintheIberianPeninsula,withitsgradual,partial“syllabification”—anaccountthathasbeenmissingsofar.Nevertheless,IamsatisfiedbythemanypointsofconvergencebetweenwhatIpresenthereandtheworkofpreviousscholars because, evidently, the validity of one’s readings is strengthened when independentapproacheshaveleadtosimilarresults.

Thesereadingsneed,ofcourse,additionalconfirmation,whichwouldcorrespondtosubse-quentstagesindecipheringwork:i.e.exposingthelanguagehiddeninthescriptintermsofpho-nology,morphologyandsyntax,achieving,atthesametime,thehighestpossiblecompatibilitybetweenthe textsandelements (namelypersonalandplace-names)knownby indirect sources,namelyingreekandLatintransmission.Thepurposeofthisessaywassolelytoprovideasetofaccuratelydefinedandregularorthographic rulesandphonetic readingsaccordingtoasoundmethodology,hopefullypavingthewayforthesecondstep.



NOTES

* New University of Lisbon – Faculty of Social and Human Sciences1 Thislong-thoughtarticlecouldonlytakeshapethankstothe

incentiveofmanypeople,friendsandcolleagues.IamgreatlyindebtedtoRodrigoBanhadaSilva(CityMuseumofLisbon),MárioVarelagomes(NewUniversityofLisbon),IlyaYakubovich(UniversityofChicago)andMáriogouveia(NewUniversityofLisbon)who,apartfromdiscussingandreviewinglatedraftsofthispaper,providedimportantreferencesandmaterial.JoséMalveiro,PauloAlexandreMonteiroandAlexandreFernandes(NewUniversityofLisbon)alsosuppliedmewithimportantliterature.IamfurtherthankfulforthepatienceofCarlosSimões,EdgarFernandes,FilipeOliveiraandJoanaBruno(NewUniversityofLisbon),whoreaddifferentdraftsofthisarticle,discussedthemwithmeandsuppliedimportantopinionsorideas.Finally,IamgratefulforthepreciousconceptualhelpofBrentDavis(UniversityofMelbourne).Iam,nevertheless,thesoleresponsibleforthefinalviewsupheldhere.Furthermore,Ihaveattemptedtogiveduecredittoeveryideacontainedhereinwhichisnotoriginallymine.However,duetothespectacularlylargenumberofworksproducedonthesubjectinthelastcenturyandmyimpossibilitytoaccessmanyofthem,Iapologizeinadvanceforanypossibleomissions.

2 Actually,suchpracticecannotbedisconnectedfromthelocalreality.TherewasawidespreadLateBronzeAge(c.1200–800BC)traditionoferectingwhatappeartobetombstoneswithdepictionsofgoods,mostlyweaponsbutalsoluxuryobjects(fibulae,mirrors,combs,etc.)and,occasionally,representationsofpugnaciousorhuntingscenes.Thoseiconographicmotifshavesustainedtheideathatthemonumentsinquestioncelebratedeadwarriorsormembersofabelligerentaristocracy.AconsiderablenumberoftheseLBAstelaehavebeenretrievedfromtheregionsofBeiraBaixa(Portugal)andExtremadura(Spain),butalsoinotherareastothesouth(seeCardoso,2002,pp.392–396forasummaryofthetopic).Itissuggestivethat,atleastinonecase,aLBAstelewasreusedinEarlyIronAgeandinscribedwiththeSWscript(text80,fromCapoteinHigueralaReal,Badajoz,Spain).ThepossibilityofsomesortofcontinuuminthetraditionoferectingdecoratedgravestoneswithastrongcomponentofsymbolicpowerinthesouthwesternregionsofthePeninsulaisnottobeoverlooked,especiallyifwetakeintoaccountthedepictionofanarmedwarrioratthecenterofthesteleofAbóbadaI(Fig.1).

3 Anyproposalconnecting,evenpartially,thewritingsystemsoftheIberianPeninsulawiththeAegean-Cypriotpre-alphabeticsyllabaries(LinearA,LinearB,Cypro-MinoanorCypriotsyllabary)astheonesuggestede.g.inTovar(1951)andPérezRojas(1986)istotallyunfoundedandmustbediscardedonchronologicalandpaleographicgrounds.

4 Thesuccessofthealphabetisinpart“accidental”asitisassociatedwiththeextensivenessofPhoenicianandHellenicmaritimetradeinIronAgeMediterranean,whereIndo-Europeanlanguagesproliferated.Likewise,theexpansionofIndo--EuropeanlanguagesfromWesternEuropeinthemodernworlddictatedbyHistoryjustifiesthepredominanceofLatin-derivedalphabetstoday.

5 ForamorecompleteexpositionofthistheoreticalframeworkonemayseeStephens&Justeson,1978,pp.275–276.

6 HereIdonotfollowNaveh’s(1973,1982)viewthatthegreekalphabetderiveddirectlyfromtheCanaanitealephatataveryearlydate(mostlybasedonthefactthatlikegreek,earlyCanaaniteshowedsomelinearityintheshapesofitsletters).

Instead,IadheretothemoreconsensualthesisofaPhoenicianorigin(startingwithCarpenter,1933;Sass,2005,withreferences).Today,evidencethattheHellenicalphabetwasinventedsomewhereinsouth-easternAnatolia(PamphyliaorCilicia)isgainingadepts(Yakubovich,2007,p.218).Onetheonehand,itcanbearguedthatthepresenceofgreek-speakingsettlersinPamphyliagoesbacktotheLateBronzeAge(Yakubovich,2008,pp.190–195,withreferences).Ontheotherhand,Phrygian,whichlikeothersubsequentAnatolianalphabetsmustbeadescendentofgreek(becauseofthechoiceofvowels),isfirstattestedongraffitionpotteryfromgordiontodaydatedto“beginningoftheeighthcentury,orafullfiftyyearsbeforethefirstassuredlygreekdocuments”(Brixhe,2007a,p.278).TheconclusionisthattheearliestgreekinscriptionsmaystillbewaitingtobefoundinsoutheasternAnatolia,wheregreek-speakingpopulationsmusthavebeenincontactwithPhrygians.

7 giventhatgreekhadnoglottalfricative,thenameoftheletterhewouldhavebeen“heard”(i.e.perceived)as/e/bygreekears,thusmotivatingitsborrowingastheletterepsilon,asBrixhe(2007a,pp.284–285)pointsout.Thesamescholararguesthatthecayinmayhavebeenusedfor/o/withbasisontheacrophonicprinciple:thesignisagraphicdepictionofan“eye”(=cayin)andallthreegreekwordsfor‘eye,eyesight’beginwithanooro–(!oφθαλμoVς,#oμμα,w! ψ).

8 ProponentsoftheHellenicoriginofthescriptalsoputemphasisonClassicalauthorswhoreportgreekexpeditionsbeyondtheStrait,suchasHerodotus,whonarrates(I.163)howIoniansailorsfromPhocaeareachedasemi-mythologicalkingdomnamedTartessos(ΤαρτησσoVς)beyondtheColumnsofHeracles(i.e.thegibraltarStrait),thusintheIberianPeninsula.TheTartessianking,Argantonius,invitedthePhocaeanstosettleinhisterritoryand,whentheydenied,hestillofferedthemgoldtobuildwallsaroundtheirpolisinAsiaMinor.Tartessosistraditionallyidentified(StraboIII,1,6and2,11)withtheterritoryaroundthebasinoftheriverguadalquivir,whereinhistoricaltimesLatinsourcesplaceanindigenouspeoplenamedTurdetaniorTurduli.Onasidenote,itseemsthattoastem*Trte–-theRomansaddedtwonativeethnonymicsuffixes,-taniand-uli(whichareequivalent;cf.anotherethnonym,BastetaniandBastuli),whichHerodotusonhisownaddornedwiththeAegeantoponymicending-σσoς,quitefrequentintheAegeancoastofAnatolia(cf.Halikarnassos,thegreekauthor’sownhometowninCaria).Theassociationbysomeofthissemi-mythologicalindigenouspeoplewiththeSWscriptisthebasistosomeofthelatter’salternativenames,despitethefactthatthevastmajorityofthecorpuscomesfromthePortugueseregionofBaixoAlentejo,notfromtheareaoftheguadalquivirinSpain.

9 Wedofindinstancesofword-finalstopsinIberianthroughgreektransmission:e.g.gaibigaitandsalirg(Michelena,1979,p.25).Onceagain,Japaneseiselucidativefromatypologicalviewpoint.DespitethesyllabicnatureoftheNipponeselanguage,sometimesianduarenotpronouncedbetweenvoicelessconsonants.Theyalsodisappearinword-finalpositionwhenstressisonthepenultimatesyllable:e.g.ikimasu‘go’(presenttense)ispronounced/ikimas/;likewiseikimashita‘went’ispronounced/ikima∫ta/becausestressisontheantepenultimatesyllable.Itispossiblethatword-finalstopsinIberianarealsotheresultofaccent-drivenlossoffinalvowels:salirg<* SA ‑lir‑gV.Nevertheless,syllabicscriptswerestill“apropriate”forbothlanguages.


Miguel Valério


10 WemaycompareLinearBwhich,regardlessofthefactthatMycenaeangreekhadvoiceddistinction,inheritedanduseduniquekandp-seriesforsixdifferentgreekphonemes:/k/,/g/,/kh//p/,/b/and/ph/.Onemaypresumethat,LinearBbeinganadaptationofLinearA,theMinoanlanguagedidnothavevoicedoraspiratedvelar/bilabialstops.

11 ThismeansIberianhadtoreinventsignsforvoicedconsonantsthathadoncebeenavailableintheoriginalborrowedscript(Phoenician).OnemaycomparetheevolutionofvelarsignsinalphabetsfromgreektoLatinalphabet.Thegreekalphabetpossessedgammaandkappa(for/g/and/k/)butwhenEtruscansborrowedandadaptedthatsystemtheykeptonlyonesign(derivedfromgamma)for/k/,C,becausetheirlanguagehadnovoicedstops.Lateron,RomansdevelopedtheirownalphabetwithbasisontheEtruscanoneandwereforcedtousethatletterforboth/g/and/k/,sinceLatinhadvoicecontrast.Cwasused

ambiguouslyforcenturiesbeforeagraphicvariant,g,wasdevisedspecificallyforthevoicedstop.

12 ContraCorrea(1990,Fig.3)andRodríguez(2000,p.31),whopropose?tobete.Onthissignseebelow.

13 IthankJoséMalveiro,whokindlyprovidedmewithapictureoftheinscription.

14 Likewise,Correa(1990,p.Fig.3)hasproposedthehypotheticreadingp(a).

15 Thereading]uarpoli[isalsopossible.16 Thisresemblanceregards,a priori,onlytheexclusivenessof

tautosyllabicclustersinbothlanguages,butithasbeensuggestedthatIberianalsohadnovoicecontrastandthatthegraphicvariantsofstopsignswereusedforallophonicvoicedstops—eventhoughthisdoesnotseemlikelygiventheexclusiveuseofbeta(andnotpi)inthegreco-Iberianinscriptions.

BIBlIOGRAPhy

ADIEgO,Ignasi-Xavier(2004)-LosalfabetosepicóricosanhelénicosdeAsiaMenor.InBÁDENASDELAPEÑA,Pedro;TORALLASTOVAR,Sofía;LUJÁN,EugenioRamón;gALLEgO,MaríaÁngeles,eds.-Lenguas en contacto: el testimonio escrito.Madrid:ConsejoSuperiordeInvestigacionesCientíficas,pp.261–270.

AKIYAMA,Nobuo;AKIYAMA,Carol(2002)-Japanese Grammar. 2nd Edition.NewYork,NY:Barron’sEducationalSeries.

ARRUDA,AnaMargarida(1997)-As cerâmicas áticas do Castelo de Castro Marim.Lisboa:Colibri.

ARRUDA,AnaMargarida(1999–2000)-Los Fenicios en Portugal.Barcelona:UniversidadPompeuFabra.

BEIRãO,CaetanodeMello(1986)-Une civilisation protohistorique du sud du Portugal(1er Age du Fer).Paris:DeBoccard.

BEIRãO,CaetanodeMello(1990)-Epigrafiada1.ªIdadedoFerrodoSudoestedaPenínsulaIbérica.Novosdadosarqueológicos.Estudos Orientais.Lisboa.1,pp.107–118.

BRIXHE,Claude(2004)-Corpusdesinscriptionspaléo-phrygiennes.SupplémentII.Kadmos.Berlin.43,pp.1–130.

BRIXHE,Claude(2007a)-Historyofthealphabet:someguidelinesforavoidingoversimplification.InCHRISTIDIS,Anastasios-Phoinos,ed.-A History of Ancient Greek. From the beginnings to Late Antiquity.Cambridge:CambridgeUniversityPress,pp.277–289.

BRIXHE,Claude(2007b)-LesalphabetsduFayoum.Kadmos.Berlin.46,pp.15–38.

BRIXHE,Claude;LEJEUNE,Michel(1984)-Corpus des inscriptions paléo‑phrygiennes.Paris:éditionsRecherchesurlesCivilizations.

CARDOSO,JoãoLuís(2002)-Pré‑História de Portugal.Lisboa:EditorialVerbo.

CARPENTER,Rhys(1933)-Theantiquityofthegreekalphabet.American Journal of Archaeology.Boston,MA.37,pp.8–29.

COMRIE,Bernard,ed.(2009)-The world’s major languages.2ndEdition.London:Routledge.

CORREA,JoséAntonio(1990)-LaepigrafíadelSuroeste.IngAMITO,Teresa,ed.-Arqueologia Hoje, I. Etno‑Arqueologia.Faro:UniversidadedoAlgarve,pp.132–143.

CORREA,JoséAntonio(1992)-Laepigrafiatartesia.InHERTEL,Dieter;UNTERMANN,Jürgen,eds.-Andalusien zwischen Vorgeschichte und Mittelalter.Köln;Weimar;Wien:Böhlau,pp.75-114.

CORREA,JoséAntonio(1994)-Lalenguaibérica.Revista Española de Lingüística.Madrid.24:2,pp.263–287.

CORREIA,VirgílioHipólito(1996)-A epigrafia da Idade do Ferro do Sudoeste da Península Ibérica.Porto:Etnos.

FABIãO,Carlos(1992)-Opassadoproto-históricoeromano.InMATTOSO,José,ed.-História de Portugal, vol. I: antes de Portugal. Lisboa:CírculodeLeitores,pp.77–299.

FARIA,AntónioMarquesde(1992)-Aindaonomepré-romanodeAlcácerdoSal.Vipasca.Aljustrel.1,pp.39–48.

FARIA,AntónioMarquesde(1994)-UmainscriçãoemcaracteresdoSudoesteachadaemMértola.Vipasca.Aljustrel.3,pp.61–63.

FARIA,AntónioMarquesde;SOARES,AntónioM.Monge(1998)-UmainscriçãoemcaracteresdoSudoesteprovenientedaFolhadoRanjão(Baleizão,Beja).Revista Portuguesa de Arqueologia.Lisboa.1:1,pp.153–160.

FERNÁNDEzJURADO,Jesús;CORREA,JoséAntonio(1988–1989)-NuevosgrafitoshalladosenHuelva.Huelva Arqueológica.Huelva.10–11:3,pp.121–142.

gLASS,Andrew(2000)-A preliminary study of Kharos. t.hı− manuscript paleography.MADissertation:UniversityofWashington.

gOMES,MárioVarela(1997)-Estelaepigrafadaenecrópole,daIIdadedoFerro,deBarradas,Benafim(Loulé).Al-ulyã.Loulé.5,p.9-22.

góMEzMORENO,Manuel(1922/1949)-Deepigrafíaibérica:elplomodeAlcoy.Revista de Filología Española.Madrid.9,pp.341–366.

góMEzMORENO,Manuel(1943)-Laescrituraibérica.Boletín del Seminario de Arte y Arqueología.Valladolid.8:1,pp.13–32.

góMEzMORENO,Manuel(1961)-Laescriturabástulo-turdetana(primitivahispánica).Revista de Archivos, Bibliotecas y Museos.Madrid.69:2,pp.879–951.



DEHOz,Javier(1976)-LaepigrafíaprelatinameridionalenHispania.InJORDÁ,Francisco;DEHOz,Javier;MICHELENA,Luis,eds.-Actas del I Coloquio sobre Lenguas y Culturas Prerromanas de la Península Ibérica (Salamanca, 27–31 mayo 1974).Salamanca:Universidad,pp.227–317.

DEHOz,Javier(1990)-ElorigenorientaldelasantiguasescriturashispanasyeldesarrollodelaescrituradelAlgarve.InPresenças Orientalizantes em Portugal. Da Pré‑História ao Período Romano.Lisboa:UniversidadeNova,1,pp.219–246.

DEHOz,Javier(1996)-Elorigendelasescrituraspaleohispánicasquinceañosdespués.InVILLAR,Francisco;ENCARNAÇãO,Joséd’,eds.-Actas del VI Coloquio sobre Lenguas y Culturas Prerromanas de la Península Ibérica(Coimbra, 1994).Salamanca:Universidad;Coimbra:Universidade,pp.171–206.

HAWKINS,JohnD.(2003)-Scriptsandtexts.InMELCHERT,H.Craig,ed.-The Luwians.Leiden;Boston:Brill,pp.128-169.

KAY,gillian(1995)-EnglishloanwordsinJapanese.World Englishes.Oxford.14:1,pp.67–76.

KLOEKHORST,Alwin(2008)-Etymological dictionary of the Hittite inherited lexiconLeiden;Boston:Brill.

LEJEUNE,Michel(1970)-Lesinscriptionsdegordionetl’alphabetPhrygien.Kadmos.Berlin.9,pp.51–74.

MEDEROSMARTíN,Alfredo;RUIzCABRERO,Luis(2001)-LosiniciosdelaescrituraenlaPenínsulaIbérica.grafitosencerámicasdelBronceFinalIIIyFenicias.Complutum.Madrid.12,pp.97–112.

MICHELENA,Luis(1979)-Lalangueibère.InTOVAR,Antonio;FAUST,Manfred;FISCHER,Franz;KOCH,Michael,eds.-Actas del II Coloquio sobre Lenguas y Culturas Prerromanas de la Península Ibérica (Tübingen, 17-19 de junio de 1976).Salamanca:Universidad,pp.23-39.

NAVEH,Joseph(1973)-SomeSemiticepigraphicalconsiderationsontheantiquityofthegreekalphabet.American Journal of Archaeology.Boston,MA.77,1–8.

NAVEH,Joseph(1982)-The early history of the alphabet.Jerusalem:MagnesPress;HebrewUniversity.

PANAYOTOU,Anna(2007)-Pamphylian.InCHRISTIDIS,Anastasios-Phoinos,ed.-A history of Ancient Greek. From the beginnings to Late Antiquity.Cambridge:CambridgeUniversityPress,pp.427–432.

PéREzROJAS,Manuel(1986)-Epigrafíatartésica.InTartessos(Revista de Arqueología,Extra;1).Madrid:zugarto,pp.74–81.

REHMAN,Abdel(2007)-English‑Tigrigna dictionary: a dictionary of the Tigrinya language.2ndrevisededition.Asmara:SimonWallenbergPress.

RODRígUEzRAMOS,Jesús(2000)-Lalecturadelasinscripcionessudlusitano-tartesias.Faventia.Barcelona.22:1,pp.21–48.

RODRígUEzRAMOS,Jesús(2002)-Elorigendelaescriturasudlusitano-tartesiaylaformacióndealfabetosapartirdealefatos.Rivista di Studi Fenici.Roma.30:2,pp.187–216.

SASS,Benjamin(2005)-The alphabet at the turn of the millennium: the west Semitic alphabet ca. 1150–850 BCE; the antiquity of the Arabian, Greek and Phrygian alphabets.TelAviv:EmeryandClaireYassPublications.

SCHMOLL,Ulrich(1961)-Die südlusitanischen Inschriften.Wiesbaden:Harrassowitz.

SCHMOLL,Ulrich(1962)-zurEntzifferungdersüdhispanischenSchrift.Madrider Mitteilungen.Heidelberg.3,pp.85–100.

SCHULTEN,Adolf(1940)-LosTirsenosenEspaña.Ampurias.Barcelona.2,pp.33–53.

SILES,Jaime(1979)-ÜberdieSibilanteniniberischerSchrift.InTOVAR,Antonio;FAUST,Manfred;FISCHER,Franz;KOCH,Michael,eds.-Actas del II Coloquio sobre Lenguas y Culturas Prerromanas de la Península Ibérica (Tübingen, 17-19 de junio de 1976).Salamanca:Universidad,pp.81–99.

SILVA,ArmandoCoelhoFerreira;gOMES,MárioVarela(1992)-Proto‑História de Portugal.Lisboa:UniversidadeAberta.

STEPHENS,Laurence;JUSTESON,JohnS.(1978)-Reconstructing‘Minoan’phonology:theapproachfromuniversalsoflanguageanduniversalsofwritingsystems.Transactions of the American Philological Association.Baltimore,MD.108,pp.271–284.

TOVAR,Antonio(1951)-Sobrelasupervivenciadelsilabismomicénicoenibéricoyenotrosalfabetos.Minos.Salamanca.1,pp.61–70.

TOVAR,Antonio(1952a)-Observacionessobreescriturastartesias.Archivo de Prehistoria Levantina.Valencia. 3,pp.257–262.

TOVAR,Antonio(1952b)-Laescriturahispánicaylosorígenesdelalfabeto.Boletín del Seminario de Estudios de Arte y Arqueología.Valladolid.18,pp.15–19.

TOVAR,Antonio(1956)-Hispaniaenlahistoriadelaescritura.Paraladelimitacióngeográficadelconceptodelotartesio.Anales de Historia Antigua y Medieval.BuenosAires.7,pp.7–14.

UNTERMANN,Jürgen(1975)-Monumenta Linguarum Hispanicarum. Band I: Die Münzlegenden. 1. Text.Wiesbaden:Dr.LudwigReichert.

UNTERMANN,Jürgen(1985)-LenguasyunidadespolíticasdelSuroestehispánicoenépocaprerromana.InWENTzLAFF-EggEBERT,Christian,ed.-De Tartessos a Cervantes.Köln:Böhlau,pp.1–40.

UNTERMANN,Jürgen(1987)-Repertorioantroponímicoibérico.Archivo de Prehistoria Levantina.Valencia.17,pp.289–318.

UNTERMANN,Jürgen(1997)-NeueÜberlegungenundeineneueQuellezurEntstehungderalthispanischenSchriften.Madrider Mitteilungen.Mainz.38,pp.49–66.

VALéRIO,Miguel(2008)-LinearAduandCypriotsu:acaseofdiachronicacrophony?Kadmos.Berlin.47,pp.57–66.

WOODARD,RogerD.,ed.(2008a)-Ancient languages of Asia Minor.Cambridge:CambridgeUniversityPress.

WOODARD,RogerD.(2008b)-The ancient languages of Syria‑Palestine and Arabia.Cambridge:CambridgeUniversityPress.

YAKUBOVICH,Ilya(2007)-[Review]BÁDENASDELAPEÑA,Pedro;TORALLASTOVAR,Sofía;LUJÁN,EugenioRamón;gALLEgO,MaríaÁngeles,eds.-Lenguas en contacto: El testimonio escrito.Madrid:ConsejoSuperiordeInvestigacionesCientíficas.The Journal of the American Oriental Society.AnnArbor,MI.127:2,pp.217–218.

YAKUBOVICH,Ilya(2008)-Sociolinguistics of the Luvian language.UniversityofChicago,PhDDissertation.

YAKUBOVICH,Ilya(2009)-[Review]géRARD,Raphaël(2004)-Phonétique et morphologie de la langue lydienne.Leiden:Peeters.Journal of Near Eastern

Studies.Chicago,Il.68:1,pp.43–45.

origin and development of the paleohispanic scripts: the ... · revista portuguesa de arqueologia....

Documents