Australo-Papuan treecreepers (Passeriformes: Climacteridae ...
Introduction: Linguistic challenges of the Papuan region
12
Language Documentation & Conservation Special Publication No. 5 (December 2012) Melanesian Languages on the Edge of Asia: Challenges for the 21st Century, ed. by Nicholas Evans and Marian Klamer, pp. 1–12 http://nflrc.hawaii.edu/ldc/sp05/ http://hdl.handle.net/10125/4558 Creative Commons Attribution Non-Commercial No Derivatives Licence 1 cc Marian Klamer The Australian National University Nicholas Evans Introduction: Linguistic challenges of the Papuan region The region where Papuan languages are spoken – centred on the Island of New Guinea, with extensions westward into Timor and the islands of eastern Indonesia, and eastward into the Solomon Islands – is at the same time the most linguistically diverse zone of the planet and the part of the logosphere. 1 It packs around 20% of the world’s languages into less than 1% of its surface area and less than 0.1% of its population. The absolute level of linguistic diversity – whether measured in sheer numbers of languages, or in terms of ‘maximal clades’ of unrelatable units – is comparable to the whole of Eurasia. Getting the right term to describe the region of interest in this collection is a famously difficult problem. Melanesia is a little too broad – extending out to Fiji, Vanuatu and New Caledonia to the east, a little beyond the scope of the present collection, and on the other 1 We gratefully acknowledge the support of the various bodies which supported the original confer- ence in Manokwari from 8-10 February, 2010, under the title ‘Melanesian Languages on the Edge of Asia’: the Australia-Netherlands Research Collaboration, the Australian National University (Department of Linguistics, College of Asia and the Pacific, and the Stephen and Helen Wurm Endowment), the Max Planck Institute for Evolutionary Anthropology, and the Universitas Negeri Papua (Unipa) and the Centre for Endangered Languages Documentation (CELD) for so gener- ously hosting the conference. We would further like to thank the various colleagues who acted as referees for the present collection (in alphabetical order): Sander Adelaar, Wayan Arka, Johan van der Auwera, Matthew Baerman, Rene van den Berg, Sonia Cristofaro, Mary Dalrymple, Philippe Grangé, Simon Greenhill, Harald Hammarström, Andy Pawley, Ger Reesink, Malcolm Ross, Alan Rumsey, Ruth Singer, Lourens de Vries, as well as to Susan Ford and Aung Si for their editorial assistance. Evans would further like to thank the Alexander von Humboldt Stiftung, through an Anneliese-Maier Forschungspreis, for financial support which assisted in some of the final production of this collection. Universiteit Leiden
Transcript of Introduction: Linguistic challenges of the Papuan region
Language Documentation & Conservation Special Publication No. 5 (December 2012)Melanesian Languages on the Edge of Asia: Challenges for the 21st Century,
ed. by Nicholas Evans and Marian Klamer, pp. 1–12http://nflrc.hawaii.edu/ldc/sp05/
http://hdl.handle.net/10125/4558
Creative Commons Attribution Non-Commercial No Derivatives Licence
1
cc
Marian Klamer
The Australian National UniversityNicholas Evans
Introduction: Linguistic challenges of the Papuan region
TheregionwherePapuanlanguagesarespoken–centredontheIslandofNewGuinea,withextensionswestwardintoTimorandtheislandsofeasternIndonesia,andeastwardintotheSolomonIslands–isatthesametimethemostlinguisticallydiversezoneoftheplanetandthepartofthelogosphere.1 Itpacksaround20%oftheworld’slanguagesintolessthan1%ofitssurfaceareaandlessthan0.1%ofitspopulation.Theabsoluteleveloflinguisticdiversity–whethermeasuredinsheernumbersoflanguages,orintermsof‘maximalclades’ofunrelatableunits–iscomparabletothewholeofEurasia. Gettingtherighttermtodescribetheregionofinterestinthiscollectionisafamouslydifficultproblem.Melanesiaisalittletoobroad–extendingouttoFiji,VanuatuandNewCaledoniatotheeast,alittlebeyondthescopeofthepresentcollection,andontheother
1 Wegratefullyacknowledgethesupportofthevariousbodieswhichsupportedtheoriginalconfer-enceinManokwarifrom8-10February,2010,underthetitle‘MelanesianLanguagesontheEdgeofAsia’: theAustralia-NetherlandsResearchCollaboration, theAustralianNationalUniversity(DepartmentofLinguistics,CollegeofAsiaandthePacific,andtheStephenandHelenWurmEndowment),theMaxPlanckInstituteforEvolutionaryAnthropology,andtheUniversitasNegeriPapua(Unipa)andtheCentreforEndangeredLanguagesDocumentation(CELD)forsogener-ouslyhostingtheconference.Wewouldfurtherliketothankthevariouscolleagueswhoactedasrefereesforthepresentcollection(inalphabeticalorder):SanderAdelaar,WayanArka,JohanvanderAuwera,MatthewBaerman,RenevandenBerg,SoniaCristofaro,MaryDalrymple,PhilippeGrangé, SimonGreenhill, HaraldHammarström,Andy Pawley,Ger Reesink,MalcolmRoss,AlanRumsey,RuthSinger,LourensdeVries,aswellas toSusanFordandAungSi for theireditorial assistance.Evanswould further like to thank theAlexander vonHumboldt Stiftung,throughanAnneliese-MaierForschungspreis,forfinancialsupportwhichassistedinsomeofthefinalproductionofthiscollection.
Universiteit Leiden
2Introduction
Melanesian languages on the edge of asia: Challenges for the 21st Century
handnotgenerallyincludingtheLesserSundaislandsintheIndonesianarchipelago. Noraredefinitionsintermsoflanguagefamilieseasytomakecleanly.TheAustronesianlanguages have wrapped NewGuinea and its surrounding islands in a three thousandyearembracethatisstillbeingplayedoutinintimatelanguagecontactwithalltheotherlanguagesoftheregion.SomeofthepapershereconcerneitherAustronesianlanguageswithsignificantstructuralresemblancestonon-Austronesianlanguagesoftheregion–seeExter’spaperonWogeo–orvarioustypesofhistoricalandtypologicalinteractionbetweenAustronesianandnon-Austronesianlanguages–seethepapersbyReesink&DunnandbyKlamer. For the non-Austronesian languages of Melanesia and its surrounds (excludingAustralia), the collective name ‘Papuan’ has been widely used and we continue thatpracticehere.Thisuse,basedondefinitionbyexclusion,hashungonforwantofabettertermlongaftercomparabletermslike‘Palaeosiberian’havebeenabandoned,butincludesupwardsoffortydistinctfamiliesandisolates.Togetanideaofhowdistortingatermlikethis is,considerhowunsatisfactoryitwouldbetousea termlike‘Eurasian’for thesetof languages includingBasque,Finnish,Georgian, Ingush,Chinese,Tamil,Cambodian,Japanese,Kurdish,Japanese,Hmong,Ket,Chukchi,Burushaskiandalltheothernon-Indo-EuropeanlanguagesofEurasia(wherewepartitionoffIndo-Europeanlanguagesonly,inthesamewaythatwepartitionoffAustronesianlanguages).YetthatisarguablythelevelofgeneticandtypologicaldiversitywhichwefacewhenconfrontedwiththefullrangeofPapuanlanguages.Despitetheseproblems,wecurrentlyhavenobetterterm,sothereaderissimplycautionedtokeepallthesecaveatsinmindeachtimetheword‘Papuan’isused. Our knowledge of this exuberant linguistic cornucopia lags behindwhatwe knowaboutanyotherregionoftheglobe.Itislikelythatthelinguist-to-languageratioislowerherethananywhereelse,anditiscertainthattherelativeleveloflanguagedocumentationislowerherethananywhereelse(seeHammarströmandNordhoff’spaper).Theinchoatestate of Papuan linguistic studies stems frommany reasons.These include the recencyof linguistic research in the area, the inaccessibilityofmany sites, the lackof relevanttraining organisations in the countries concerned, the fragmentation of research acrossnationalboundariesandacross theacademicvsmissionarydivide,and thegeneral lackofprioritisationthatlargepartsofthelinguisticprofessionhaveuntilrecentlyassignedtothedocumentationoflinguisticdiversity.Wehaveputtogetherthiscollectionofpapersasasampleofjustsomeoftheresearchquestions,languagesandapproachesthatcurrentlyseemparticularlyexciting,withthegoalofraisinginterestinthisfascinatingpartofthelogosphere,andencouraginglinguistsfromaroundtheworldtogetinvolvedinresearchinanareawherethereissomuchjustwaitingtobediscovered,. ThepresentcollectiongrowsoutofaconferenceheldattheCentreforEndangeredLanguagesDocumentation(CELD)atUniversitasNegeriPapua(Unipa)inManokwari,Indonesia, in February 2010, with the support of theAustralian Netherlands ResearchCouncil,theAustralianNationalUniversity,andtheMaxPlanckInstituteforEvolutionaryAnthropology inLeipzig, all ofwhomwe thank for their financial assistance. It is notsimplyaconferenceproceedings,however–itrepresentsmerelyaselectionofpapersfromthatconference,supplementedbyanadditionalpaper tofill ingapswe thoughtneededcoverage. Webeganthisintroductionbycontinuingthewell-establishedtraditionofstating,in
3Introduction
Melanesian languages on the edge of asia: Challenges for the 21st Century
anapproximateandratherunquantifiedway,thatMelanesia,andinparticulartheislandofNewGuinea,containsboththegreatestconcentrationoflinguisticdiversityanywhereonearth,andthelowestlevelofdocumentation.Thefirstpaperinthisvolume,The languages of Melanesia: Quantifying the level of coverage by Hammarström & Nordhoff, addsprecision to this statementbypresenting relevantfigures from theirLangDocdatabase.Thisdatabaseaims togivecomprehensiveglobal listingofallmaterialsexistingonalllanguages,alongwithaninitial,approximatemetricofdegreeofcoverage.Astheauthorspoint out, there are many shortcomings to their metric. It is relatively unambitious: alanguagepossessingalow-qualitygrammarof160pagesandnolexiconortextcollectionwouldalreadybeplacedatthehighestlevel—wellshortofthemoderngoldstandardofaBoasiantrilogysupplementedbyawidevarietyofannotatedmultimediafiles—andthereisnomeasureofqualityofanalysis.Despitetheseflaws,ithasthegreatvirtueofbeingoperationalisableandapplicabletoalltheworld’slanguagesinarelativelyautomaticway,and in theirpaper theyoutline their scheme indetailaswellascomparing the levelofcoverageforMelanesiawiththerestoftheworld. First, regarding the totalproportionof theworld’s languages spoken inMelanesia,theirfigurescount1347languagesthatare‘Melanesian’intheirsense(522Austronesian,825non-Austronesian)of thesub-regionofOceaniaextendingfromtheArafuraSea inthewest toFiji in the east (seefigure1 in their chapter).This is just over 20%of theworld totalof6496 (living) languageson their count (see their table7),with so-called‘Papuanlanguages’thenmaking12.7%oftheworld’stotal.Inabsoluteterms,thenumberof languages inMelanesia (1,347) is almost identical to those in thewhole ofEurasia(1,465),thesetwobeingsurpassedonlybyAfrica(1,986). Second,for theirassessmentof levelofdocumentation, theylumptogetherPapuanwithallAustronesian,sotheirfiguresalsoincludetherestofIndonesia,thePhilippines,Malaysia, and so on. Drawing on these figures, they draw some striking conclusions.First,inabsoluteterms,Papua+Austronesianhasthelargestnumberoflanguageswithonlyawordlisttotheirdocumentation(i.e.thelowestlevelofdocumentationwhichtheyrecognise).ThecomparisonwithAustralia,another regionwhereprofessional linguisticresearchisrelativelyrecent,issalutary:over42%ofAustralianlanguageshaveagrammaravailable, compare to half that number (20.48%) forAustronesian + Papuan. Second,inrelativeterms,Papua+Austronesianhasthelowestproportionoflanguageswiththehighestrankofdescription(i.e.agrammarof150pagesormore),thehighestproportionwithonlyawordlistorless,andthelowestaveragelevelofdocumentation.Third,whenAustronesianandnon-Austronesianlanguagesarecomparedwithintheabovecategories,thenon-Austronesianlanguageshavelowerlevelsofdocumentation,makingtheiroveralldocumentationstatusevenlowerthanthatforPapua+Austronesianasawhole(seetheirTable6). In theyears tocomeit is tobehopedthatLangDocwillbeextendedtogivemoreaccurate metrics in a number of ways – something which will be aided to the extentthatmore linguists heed the authors’ call to put their results in thepublic domain.Buttheirpaperalreadyprovidesaveryclearquantitativebasisforourclaimabove that theMelanesianregion–andparticularlythePapuanlanguageswithinit–isfarandawaythemostlinguisticallydiversepartoftheplanet,andthatconverselyitsuffersfromthelowestlevelof languagedocumentation found inanyquarterof theearth.Thecombinationof
4Introduction
Melanesian languages on the edge of asia: Challenges for the 21st Century
thesefactorsiswhatmakesthestudyofMelanesianlanguagesanenormouschallenge. Beforeleavingthispaper,wenotetwoimportantfuturedevelopments.FirstlyitwillbecrucialtolinksomeformofcomprehensivedatabaselikeLangDoctoactualdocumentsso that it is possible to inspect the actual materials listed there and gradually improve thequalitativeratingsthroughthecollectiveeffortsofworldscholarship.Secondly,itisdesirablethatthestructureoftheLangDocdatabaseallowsinspectionofdataatanumberofdifferentgeographiclevels.WhiletheirpresentarticlelargelytreatedMelanesiaasanundifferentiatedwhole,theirdiscussionofonegeographicalvariable(distancefromcoast)showshowmorefinelyarticulatedgeographicalcharacterisationscanbemade–sothatonecancompilecomparablereportsforgeographicalregionsliketheSepik,Bougainville,etc. The staggering linguistic complexity of Melanesia creates special problems forattemptstoclassifylanguagesintofamiliesandsubgroups,especiallyforeffortsthattrytoreducethelargenumberofindependentmaximalclades(overfortyonanyestimate)bygroupingsomeofthemtogether. Underthesecircumstances,thelanguagesofMelanesiahaveprovidedaparticularlyimportant testing-groundinrecentyearsfornewmethodswhichaimto‘breakthetimebarrier’ of the classical comparativemethod, by drawing inferences from the signal inassemblages of typological traits rather than simply in the sound-meaning pairings ofthelexiconandgrammaticalmorphology.Thoughcontroversialandstillsubjecttofiercecritique (see references in Reesink andDunn paper), it is likely that suchmethods asapplied to Melanesia are here to stay, at the very least as a supplement to the comparative method.Indeed,thesituationinMelanesiaforceshistoricallinguiststomakeavirtueofnecessitybydrivingthemtodevelopnewmethods. ThearticlebyReesinkandDunngivesanoverviewofthesemethods,focussingonthelanguagesofEasternIndonesia,spokenaroundtheBird’sHeadarea.Asintheirotherstudies,agraveproblemwiththemethodisthatresemblantsignalscansignaleithersharedphylogenyorarealconvergence.Inthecentralpartoftheirpaper,theyconsiderthecaseoftwoPapuanlanguagesoftheBird’sHead,HatamandMeyah,whichconsistentlyclusterwiththeAustronesianlanguageBiaknomatterhowmany‘foundinglineages’(Kvalues)areassumedonrunsofthe‘Structure’algorithm.Inthiscase,then,‘itthusappearsthatdiffusionoverridesphylogeny’,astheyputit. Buttheythentakeafurtherstep,teasingoutthefifteentypologicalfeatures(outof160 altogether) which align with phylogeny rather than areality, opposing the PapuanlanguagesHatamandMeyahagainsttheAustronesianlanguageBiak–seetheirtable6.Doesevidencelikethisholdthekeytorefiningtypological-suitebasedmodelssothattheycanfilteroutarealnoisetofindthephylogeneticsignal?Obviously,iftheargumentationproceedsjustfromasinglecase,ashere,itrisksbeingmerelyposthoc,butontheotherhanditwouldbepossibletoiteratethisprocedureoveranumberofareasandsmallgroups. Will iterations of this type, byfiltering out themore from the less diffusable overindependentcasesfromaroundtheworld,allowustofine-tuneanalgorithmlikeStructureby weighting the evidentiary value of different typological characters as regards tophylogenyvsareality?Thiswillbeacrucialquestionoverthenextdecadesofscholarshipas more extensive documentation of Melanesia’s languages provides us with moreinformationforfeedingintocomparativeenterpriseslikeReesinkandDunn’s.Atthesame
5Introduction
Melanesian languages on the edge of asia: Challenges for the 21st Century
time,theirworkremindsusthat,todrawmaximumbenefitfromresearchliketheirs,ourlanguage documentations need to ensure that matched typological data is obtained – this neednotentail‘questionnaire-style’grammars,butfeaturelistslikethoseintheAppendixtotheirarticledolaydownabasicchecklistoftypologicalpointswhichalldescriptionsshouldmakesuretocover. Thenext twopaperseachconsider regionsofMelanesia inwhich therehavebeencomplex interplaysbetween languagesbelonging toquitedifferent families, in a socialenvironmentwhere different types of contact appear to have played a role at differentpoints in the past. Marian Klamer’s Papuan-Austronesian language contact: Alorese from an areal perspectivefocusesonAlorese,anAustronesianlanguageabuttingthewesternmostgroupofextantPapuanlanguagesontheislandofPantar.Shededucesacomplexcontacthistorycomprisingatleasttwostagesplayedoutindifferentlocations. Thefirstphase,onhermodel,wouldhavetakenplaceontheislandofFloresornearby,atatimewhenPapuanlanguageswerestillspokenthere.ItisatthisstagethatthelanguageancestraltomodernLamoholotandAlorwouldhaveacquiredasuiteoftypologicalfeaturesthatareseenastypically‘Papuan’–or,moreprecisely,astypicalofthePapuanlanguagesoftheAlor-Pantarregion–includingpost-predicatenegation,themarkingofpossessors,noun-locationalorder in locativeconstructions, thepresenceofa focusparticleand theabsenceofapassiveverbform.This‘Papuanisation’ofproto-Lamoholotwouldhavetakenplaceunderconditionsoflong-termstablecontactinvolvingpreadolescentsacquiringthecomplexitiesofbothPapuanandAustronesian languagesandmelding them intoanewsystem. In a second phase, following themigration ofAlorese speakers to Pantar and theseparationthisentailedfromtheirLamoholotcousins,aseriesoffurtherchangeswouldhavetakenplace.AloresecontrastsdrasticallywithLamoholotintermsofmorphologicalcomplexity.Where Lamoholot has two sets of subject affixes to the verb (prefixes fortransitives,suffixesforintransitives),Aloresereliesonfreepronounswithallbutafewfrequentverbswhichretainfossilisedagentprefixes.AndwhereLamoholothasanumberofderivationalaffixes (someproductive, some lexicalised),Aloresehasnoderivationalmorphologyat all– reduplication is itsonlyproductiveword formationprocess.ThesedifferencessuggestaradicalprocessofmorphologicalsimplificationinthepassagefromLamoholottoAlorese.Klamerhypothesisesthat,intheinitialstagesofAloresesettlementsofPantarandAlor,Alorese-speakingmenwouldhavetakenastheirwiveswomenspeakinga number of differentPapuan languages of the inland.Entering the speech communityasadultstheywouldhavelearnedasimplifiedformofAlorese,jettisoningalmostallofitsmorphology.ThecontactbetweenAloreseandlocalPapuanlanguages,however,wasneitherprolongednorconsistentatthisstage.ThenumberofloanwordsfromlocalPapuanlanguagesisrelativelylow(only14Aloresetermsoutofa270word-listhaveaknownPapuansource)andismoreoverdistributedevenlyacrossthedifferentPapuanlanguagesofthelocality.Thissuggestsanumberofrelativelyweakcontactsandnostablepatternofbilingual contact. Thiscasestudyillustratesatypeofmulti-phasecontactscenariolikelytohavebeenplayedoutbetweenAustronesianandPapuan speakers in anumberofpartsofEasternIndonesia at different phases over the last two to three millennia. The very different
6Introduction
Melanesian languages on the edge of asia: Challenges for the 21st Century
outcomesofthetwophasespositedinKlamer’smodelareasalutaryreminderofthesocialandlinguisticcomplexitythatmusthavebeeninvolvedbetweentwogroupswhowouldhavebeendemographicallyequallypoisedandinterdependentinmanyways.Atthesametime,asKlamerpointsout,itisonlyareconstruction,andwewouldbeonmuchfirmergroundifwewereable todrawoncontemporarysociolinguisticstudiesof the typesofinteraction–socialandlinguistic–thatareoccurringbetweengroupsalongthePapuan-Austronesianinteractionzone.Aswithsomanyofthequestionsraisedinthisissue,thetimeforthissortofstudyisrunningoutfast,asthepresenceofanalternativelinguafranca(e.g.Indonesian)radicallyaltersthetypeoflinguisticinteractionbetweensuchgroups. FromNusaTenggarawe thenmove east to the SouthernNewGuinea region, thefocusofNicholasEvans’Even more diverse than we thought: The multiplicity of Trans-Fly languages.Incontrast totheAustronesian-Papuaninteractionsintheprecedingtwoarticles,heretheinteractionsarebetweenvariousunrelatedPapuangroups.SouthernNewGuineaisanintriguingzone,ofgreatdiversity,aboutwhichourlevelofknowledgedipsevenlowerthanthenormsforelsewhereinMelanesia. TheSouthernNewGuinearegionisessentiallyanucleusofseveralsmalllanguagefamiliessurroundedbyTrans-NewGuinealanguageswhichsignificantlyoutnumberthemdemographically,andwhichat thetimeoffirstcolonialdocumentationtendedtobefarmore expansive andmilitarised than their non-TNGcounterparts. It offers an excellentopportunityforhistoricallinguisticstostudythemechanismsbywhichTrans-NewGuinealanguages have expanded into areas previously characterised by greater levels of deepphylogenetic diversity. Nonetheless,itisclearthatalllanguagesoftheregionshareanumberoftypologicalcharacteristics–totheextentthatsomelanguages,whichhavebeenclassifiedasTNG,likeMarind,patterntypologicallywithotherSouthernNewGuinealanguages(aswellassomelanguagesfurtherafield,includingYeli-DnyeandInanwatan–seeReesinkandDunn,Fig.1, aswell asdiscussion in footnote4ofEvans’ article.)This suggests that, even if thepresenceofTNGlanguagesinSouthernNewGuinearesultsfromexpansionattheexpenseofothergroups,theremusthavebeenenoughstablelong-termbi-ormultilingualismforsignificant linguisticconvergence tooccur.The languagesof theSouthernNewGuineaexhibit high levels of morphological complexity allied with a host of highly unusualtypological features, and Evans’ paper gives short sketches of two neighbouring butunrelatedlanguages–NenandIdi–focussingparticularlyontheircomplexverbalandcasemorphology.Heshows that,despite thepresenceofsomeconvergent featuresandwidespreadbilingualismandcontactbetweenthespeakersofthesetwolanguaes,therearemajordifferencesinhowtheyorganisetheirgrammars.(NoteinpassingthattheReesinketal2009sampledidnotincludeanylanguagefromthePahoturiRiverfamilywhichIdibelongsto,soitisnotclearhowfaritwouldfareontheirtypologicalprofile). TheSouthernNewGuineacase– rootedas it is inasystemofmarriagebysister-exchangewhich favours comparable demographies, interdependence, and intermarriageandmultilingualism between neighbouring groups – is a clear case of how prolongedlanguagecontactcan lead toarealpatternscharacterisedbysharedcomplexification. Inillustrationofthis,EvansconsidersthewaydifferentlanguagesinSouthernNewGuineaderiveathree-valuednumbercontrast.(singular,dual,plural).ThisissomethingfoundinvirtuallyeverylanguageinSouthernNewGuineaexceptMarind(somelanguageshavean
7Introduction
Melanesian languages on the edge of asia: Challenges for the 21st Century
additionaltrialorpaucal).Buttheexactroutebywhichsuchsystemsarederivedvariessignificantlyfromlanguagetolanguage.Thissuggeststhatwhateverseriesofpathwaysleadstosharedarealfeaturesofthistypeisalongandtortuousone,probablybasedontheslowpatchworkemergenceofgrammaticalsolutionstoparticularsemantictargetssharedacrosslanguagesoftheregion. Staying inSouthernNewGuineaandsticking to the topicofgrammaticalnumber,WayanArka’spaperProjecting morphology and agreement in Marori, an isolate of southern New Guinea examines issuesofhow to represent systemsof ‘constructed’grammaticalnumber featurally, focussingon theTNG-level isolateMaroriandother languageswithcomparable phenomena. Marori,inlinewiththeSouthNewGuineapatterndescribedinEvans’article,constructsathree-valuednumbersystembycombiningtwobinaryvaluesinasystemofdistributedexponence.However,thebasefeaturesusedtoderivethisresultaredifferent:whereNencrossesasingularvsnon-singularwithadualvsnon-dualdistinction,Maroricrossesasingularvsnon-singularwithapluralvsnon-pluraldistinction.Arka’sarticleshowshowtheunificationofnumbervaluesinMarorimorphologycanbederivedwithinamodelinwhichthefeaturesarehierarchicallystructured,indifferentwaysindifferentlanguages.ThuswhereMaroritreatsdualasthenumbervaluethatisneithersingularnorplural–andhencerelegatesthedualtoaderivedcategory–theNenfeaturestructurebuildsindualasaprimaryspecifiedfeature,buttreatspluralasaderivedcategorythatisneithersingularnordual.Thismodelisanelegantillustrationofhowsomecross-linguisticvariabilityinfeaturestructurecanbebuiltintoarobustoverallarchitecture–thepresenceofanoverallfeaturestructure,andofaprimarysingularvsnon-singularcut,remainconstant,buttheinternalmakeupofthenon-singularsubspacediffersasbetweenMaroriandNen.Theavailabilityofdifferingfeaturearchitecturesthenmakesitpossibletomodelthedifferencesbetweenlanguageswithsimilarsetsofcontrasts,butderivedindifferentways,withinaformalismlikeLFG–wereferthereadertothatchapterfortheformaldetails. The difficulties involved in lining up language-specific descriptive categorieswithcomparative concepts are nicely illustrated, from a different theoretical perspective, inMats Exter’s article ‘Realis’ and ‘Irrealis’ in Wogeo: A valid category? Recall that, inReesinkandDunn’sarticle,oneoftheirquestions(50/87,aslistedintheirappendix)is‘Isadistinctionbetweenrealis/irrealismoodavailableasamorphologicalchoice(1:present,0: absent)’?But howdowe decidewhat ‘realis/irrealis’ actuallymeans?Wogeo offersinterestingdifficultiesinansweringthisquestion. Wogeoisa‘mood-dominated’Austronesianlanguage,spokenoffthenorthcoastofPNG,withacomplexverbalmorphology includingsixprefixalandeightsuffixalslots.Abasicopposition isbetween the ‘realis’and ‘irrealis’ formsof thepronominalprefix,illustratedbyapairlikeo-lako‘Igo,Iwent’vsgo-lako‘Imustgo,Iwanttogo,Iwillgo(now)’.Ifthiswerealltherewastotheoppositionthecharacterisationwouldbefairlystraightforward,butonceweconsidermoresemanticallyprecisecombinationsproblemsarise.Wogeo has additional prefixal combinations expressing suchmeanings as future,tentative (‘try doingX’), counterfactual (‘would have doneX’), proximal imperfective(‘am/wasdoingX,nearby’)’anddistalimperfective(‘am/wasdoingX(furtheraway)’),whicharefollowedbyeithertherealisorirrealisprefix,e.g.m-o-lako[FUT-1sg.realis-go]‘Iwillgo,Icango,Imaygo’.
8Introduction
Melanesian languages on the edge of asia: Challenges for the 21st Century
Forsomeofthese,thechoiceofprefixmakessenseintermsofnormally-characterisedpropertiesofthisopposition,e.g.therealisisusedwiththetwoimperfectiveseries.Butforothers,notablythecounterfactual,itistherealisseriesthatischosenratherthantheexpected irrealis.Exter then goes on to considerwhat such casesmean for the overallenterprise of trying to define terms like realis and irrealis in cross-linguistic terms.Heendsuparguingagainsttheusefulnessofatermwithasbroadarangeastherealis-irrealiscontrast–which,ifaccepted,raisesthepossibilitythattypologicalcomparisonsmaybemore successful if theywork atmuchmore semantically-specified levelswhere cross-linguistic comparison can be more precise. The next two papers examine the embedding of language in its sociocultural andpsychologicalcontexts. Darja Hoenigman’s paper From mountain talk to hidden Talk: Continuity and change in Awiakay registersexaminesthediachronicsociolinguisticsofspecialregistersinAwiakay,alanguageofEastSepikprovince,andintheprocessthrowsafascinatinglightonhowideologiesoftheneedforlinguisticdifferenceintersectwithhighlevelsofmetalinguisticawarenesstodriveadynamicoflexicalinnovation.Particularlynoteworthyisthecontinuity–intermsofutilisingspecialregisters–thatholdsinthefaceofsignificantchange– in theformofChristianstricturesagainst theongoinguseofsometraditionalregisters. Traditionally, Awiakay people used a special register, known as ‘mountain talk’,to protect themselves frommountain spiritswhen travelling up intomountain regions;this involved the substitutionoravoidanceofanumberof lexical items.ThearrivalofChristianityhasarrested theuseof ‘mountain talk’,with the recognition itgives to thepowerofpaganspirits,andknowledgeanduseofthistraditionalregisterisindecline.Butat the same time, another special register has come into use, kay menda or‘hiddentalk’.Travellingoutside the village to regional centres such asWewak, especiallywhen it isforcommercialpurposeswhichleavesthetravellersvulnerabletotheftandpredation,isregardedasariskybusinessandspeakingalanguageimpenetrabletooutsidersprovidesgood security. ThoughAwiakay is traditionally spoken in just one village, and would thereforenormally have been incomprehensible to outsiders, the recent arrival of Tok Pisinloanwords creates chink in the armor of linguistic impenetrability. It is precisely theseloanwordswhichgetreplacedinkay menda,throughingeniousnativecoinagessomeofwhichhavealreadywonfullacceptanceinthecommunityandothersofwhichstillincluderival coinages. Hoenigman’spaperincludessubtitledvideofootageofajourneyfromthevillagetotheregionalcentre,duringwhichwecanwitnessthecamouflagingprocessesof‘hiddentalk’atwork,aswellaswatchingtherehearsalandinductionoflessexperiencedmembersofthepartywhiletravellingtowardsthedestination.Thisisofinterestnotjustforthetopicofspecialregisters,butmoregenerallyforourunderstandingofhowatleastsomeoftheprocessesoflinguisticdiversificationinMelanesiaaredrivenalongbyveryconsciousandnegotiatedprocessesofchangeaimedatdifferentiatingone’slanguagefromthatofothergroups. Herpaperconcludesbysurveyingtheparallelsanddifferencesbetweenthenewregisterofhiddentalkandthefadingoldregisterof‘mountaintalk’.Bothareusedinunfamiliar,
9Introduction
Melanesian languages on the edge of asia: Challenges for the 21st Century
perilousterritorywhereonegoestoobtainvalueditems,encounteringdangerousentities(mountain spirits before, rascals now) and dangers (sickness before, robbery and theftnow),preventing thesedangers throughjudiciousout-of-the-ordinary languageuse,andpredominantlyinvolvingmenwhoaretheonestravellingtothedangerousdestinations.Themost interestingdifference,onhercomparison,has todowithwhoisheldtohavecreated the special register. In the caseofmountain talk this is attributed to ‘mountainspirits’deepinthepast,whereasinthecaseof‘hiddentalk’theprocessofcreationisstilltakingplace,involvescontemporaryAwiakayindividuals,andisthereforeaprocessthatis amenable to direct research on such questions as how rival innovations are selectedbetween,whichitemsarechosenforcamouflaging,andhowchangesarepropagatedfrominnovative individuals to the community. MichaelFrank’spaper,Cross-cultural differences in representations and routines for exact number,leadsusfromtheknowndiversityofMelanesianlanguagestothepresumedbutuntestedcognitivediversitythissubtends.Beller&Bender(2008),whomhequotesinhisarticle,observethat‘theremaybenootherdomaininthefieldofcognitivescienceswhereitissoobviousthatlanguage(i.e.,theverbalnumerationsystem)affectscognition(i.e.,mentalarithmetic).’CombiningthiswiththelikelihoodthatMelanesiandiversityinnumeral systems (Lean1992) is perhaps evengreater, in relative andabsolutive terms,thaninotheraspectsofthelanguagesystems,wehavehereafascinatingdomainfortheinvestigationofhowlinguisticdiversityshapescognitivediversity–aswellashowculturalpracticeslikedifferentcountingroutinesthemselvesselectfortheemergenceofdifferenttypesofnumeralsystem. Frank’s paper does not in itself begin the exciting project of investigating howMelanesiandiversityinnumeralsystemsproduces(ordoesn’t)significantdifferencesincognition.Rather,itsgoalistoclarifytherelationsbetween,ontheonehand,howexactnumberisrepresentedlinguisticallyorthroughothertypesofrepresentationaltoolsuchasthetheMentalAbacus,andnumericalcognitionontheother.Frankadducesexperimentalevidencethatlinguisticsystemsintheformofnumerals,butalsonon-linguisticsystemsintheformoftheMentalAbacus,bothprovidea‘cognitivetechnology’enablingtheonlineencodingandmanipulationofquantity information.Frankshows thatculturalexposurealonedoesnotscaffoldthemanipulationofexactnumber,thatthelackofexactnumeralterms in a language impacts negatively on arithmetical manipulations, and that it is not enough topossess a languagewith appropriate termsbut thatonemust alsobeable toaccessitonlineinordertosuccessfullycarryoutarithmeticalcalculations. Different numeral systems, such as the use of different bases (2, 5, 6, 10, 20) canbeexpectedtofurnishdifferentcognitivestrategiesinthissense;andsowoulddifferentnumeralsetsfordifferenttypesofcountedobjects.Thissuggeststhatcollaborativeresearchontheimpactofnumeralsystemsonnumericalcognitionwillyieldrichresults.Yetthechallengesofinvestigatingthisinteractionaregreat,andrequireatypeofcollaborationbetweenlinguistsandpsychologiststhathavebeenalltooraresofar:
The data that lead to this conclusion could not have been gathered by the standard methodsofcognitivepsychology,norbythestandardmethodsoffieldlinguistics.Manyoftheresultscitedherecomefromcarefullycontrolledstudiesperformedinthefieldwithpopulationsthatpossessculturally,linguistically,orcognitively
10Introduction
Melanesian languages on the edge of asia: Challenges for the 21st Century
interestingnumericalrepresentations.Thisgeneralizationsuggeststhebenefitsofpsycholinguisticfieldworkthatcombinesexperimentaldesignwithcross-culturalorcross-linguisticpopulations(Frank,thisvolume:234).
Infact,therelevantnumericalsystemsareratherfragile,insomecasessignificantlymoresothanotherpartsofthelanguagesystem:‘therearemanycaseswherealanguageisnotendangered,ornotparticularlyendangered,butwhosenumeralsystemsareendangered’(Comrie2005).OksapminisasalutaryMelanesiancase,investigatedbySaxe(1982)andthenSaxe&Esmonde(2005).Thoughthelanguageisstillhealthy(Loughnane2007)itsdistinctivebase-27bodycountsystemisgivingwaytoanEnglish-styledecimalsysteminamodernsettingwherecountingismostcommonlyappliedtomoney.ThisgreatfragilityofnumeralsystemsmeansthatthereisanexceedinglynarrowtimewindowforcarryingoutthesortofcollaborativeworkontheimpactofnumeralsystemsonnumericalcognitionwhichisoutlinedinFrank’sarticle. Themanyfascinatingquestionsandresearchthrownupbytheprecedingarticles–andevenmoreso, thefutureresearchwhichwehope theywillstimulate–generateanenormousamountofprimarydataasfieldworkersofarangeofinterestsandnationalitiesrecordmaterialsonnumerous languagevarieties, in increasinglydata-rich formats.Butendangereddataisaproblemweneedtotakealmostasseriouslyasendangeredlanguagesthemselves–fieldnotesand recordingsmayendup lost,uncatalogued,unlocatable,ordegenerateonoldtapesorothermaterials.Equallyproblematicarequestionsaboutwheredatashouldbehousedandwhoshouldgetaccesstoit.Moderndigitalarchivesaregivingusthepowertoaddresstheseissuesinanefficientway.Theyhavethepotentialtopreservehugeamountsofdatafarintothefuturewhileallowingthemtobeaccessibletoresearchersandcommunitymembersfromalllocations.Itiswiththedesignandrunningofonesucharchive, Paradisec, that the last article, byNicholasThieberger and Linda Barwick, isconcerned: The Pacific and Regional Archive for Digital Sources in Endangered Cultures (PARADISEC): A resource for Melanesian linguistics. PARADISEC was established inAustralia in 2003, by a team of researchers ledbyThieberger andBarwick. Itwas born as a response to the challenges set out in theprecedingparagraph,fromanawarenessthatavastbodyofhard-wonfielddatawasatriskofvanishingaltogether–partlyasaresultofpoorfacilitiesinlocalarchives(e.g.lackofair-conditioning to maintain tapes in good condition), partly through technological changes (e.g. thedisappearanceofmachinesabletoreadoldrecordingsonwaxcylinders,wire-recordersetc.),partlythroughalackofemphasisinthefieldoflinguisticsontheprimacyofdocumentationasopposedtotheoreticaldebateorgrammar-writing,andpartlythroughthe reluctance of individuals tomake theirmaterials available to others until they hadanalysedthemthemselves–amomentwhichsometimesgetsovertakenbyAlzheimer’sordeath. Part of their article is devoted to showing how PARADISEC works, in terms ofequipment setup, backup, access, workflow and data ingestion. But Thieberger andBarwickalsodiscussanumberofotherimportantconcernsraisedindiscussionsofwherearchiveslikePARADISECfitinaregioncharacterisedbysuchvastdiscrepanciesbetweencountriesintermsoflivingstandards,technology,accesstodigitaldata,andthepotentialvalueoflocalinformation.Acentralissueisthemoraltensioncentredontheforeseeable
11Introduction
Melanesian languages on the edge of asia: Challenges for the 21st Century
andunforeseeableusesofarchived information (e.g. in establishingclanownershipoflands,rightstoroyaltiesetc.)whichrequiregraduatedlevelsofaccess,onthephilosophythatthereshouldbegeneralcommitmenttopermanentarchiving,forfuturesafety’ssake,butthatcommunitiesandresearchersshouldbesuppliedwithtechnologiesforregulatingaccesswherethisiswarranted. Lookingintheotherdirection,thepotentialforharnessingthecollectiveknowledgeofvariouskindsofexpertthroughcumulativeannotationofarchivedmaterialbydifferentarchiveusersatdifferentlocationsisagoalthathasgreatpotentialtogalvaniseamorecollective and interdisciplinary approach to adding commentary and interpretation to primary material through time. TheninecontributionswehaveoutlinedcandonomorethangiveatantalisingglimpseofthechallengesraisedbythelanguagesofMelanesia–forlinguistsandscholarsinalliedfields, but also for educators, communication technologists, and development agencieswantingtofocusonlocalknowledgeandexpertise.Mostimportantly,thisisachallengeofutmostinteresttocommunitymemberswantingtomaintaintheintellectualwealthheldin their linguistic heritage. For them, collaborativeworkwith linguists and others canoffernewwaysofintegratingthatheritagewithothersortsoflanguageproductssuchasorthographies,dictionaries,grammars,textcollections,anddigitalethno-encyclopaedias. Forevenafractionofthesechallengestobemet,manythingsmusthappen.Weneedto attract a newgeneration of adventurous and capable young scholars towork in thisfascinating,diverseandhospitablepartof theworld.Weneed tobuildcapacityamonglocallinguistsandlanguageworkersinthecountrieswheretheselanguagesarespoken,soas toreverse thedrasticcurrent imbalancebetweenwhereMelanesian languagesarespokenandwherefutureresearcherscanreceiveadvancedtraininginhowtostudythem.CELD,theCentreforEndangeredLanguageDocumentationinManokwari,whichhostedtheconferencewheremostofthepapersherewerepresented,isapromisingstepinthisdirection. There need to be many other developments like this, and international fundingagencies need to be convinced that language diversity is a resource, not a handicap. This isparticularlyrelevantata juncturewhenkeysourcesof internationalresearchfundingoverthelasttwodecades(theVolkswagenstiftung’sDoBeSprogram,theHansRausingEndangered Languages Program, the NWO Bedreigde Talen program and the ESFEuroBABEL program) are drawing to a close, or have already.There is vast potentialinsuchnewapproachesasBOLDorBasicOralLanguageDocumentation(http://www.boldpng.info/iwlp) andmobile-phonebased crowd-sourcing to assist thedata-gatheringprocess.But theneed for long-term traditionalfieldworkdrawingon theknowledgeoflinguistswholearnthelanguagesandcultureson-sitewillremainfundamental.Finally,whiletherewillalwaysbesomedivergenceofinterestbetweenmissionaryorganisationsand academically-motivated researchers, the vast extent of missionary enterprisesthroughMelanesiameansthatthepotentialforfruitfulcollaborativeworkisvast,givengoodwillonbothsides.AnimportantrecentinitiativeisthereestablishmentofthejournalLanguage and Linguistics in Melanesia, now as an open-access on-line journal (http://www.langlxmelanesia.com/),asaforumforpublishingpeer-reviewedresearchandbookreviewsonthelanguagesofMelanesia.
12Introduction
Melanesian languages on the edge of asia: Challenges for the 21st Century
As can be seen from these considerations, and the fact that almost every paper inthiscollectionisanearlystepinanewresearchpath,thestudyofMelanesia’slanguagesoffersabundantopportunitiestomakenewdiscoveriesWehopethatinthecollectionofpapersgatheredhereyouwillfindmaterialthatinvitesyouintoanengagedanddiverseinternational community of scholars dedicated to advancing our understanding of alinguistic territory that is arguably the least charted on earth.
RefeRences
Beller, S. & A. Bender. 2008. The limits of counting: Numerical cognition betweenevolution and culture. Science319.213-215.
Comrie, Bernard. 2005. Endangered numeral systems. In Jan Wohlgemuth & TykoDirksmeyer (eds.), Bedrohte Vielfalt: Aspekte des Sprach(en)tods [Endangered Diversity: Aspects of Language Death],203-230.Berlin:WeissenseeVerlag.
Frank,MichaelC.Thisvolume.Cross-culturaldifferencesinrepresentationsandroutinesforexactnumber.
Haspelmath, Martin. 2009. Comparative concepts and descriptive categories in cross-linguistic studies. Language86(3).663-687.
Lean,G.1992.CountingsystemsofPapuaNewGuineaandOceania.Lae:PapuaNewGuineaUniversityofTechnology doctoral dissertation.
Loughnane,Robyn.2007.AgrammarofOksapmin.Melbourne:UniversityofMelbournePhD thesis.
Saxe, G. B., 1982. Developing forms of arithmetical thought among theOksapmin ofPapuaNewGuinea.Developmental Psychology18(4).583-594.
Saxe,G.B.&Esmonde,I.2005.Studyingcognitioninflux:Ahistoricaltreatmentoffu in theshiftingstructureofOksapminmathematics.Mind, Culture, and Activity 12(3/4). 171-225.
Nicholas [email protected]
Marian [email protected]
