Treść książki
Przejdź do opcji czytnikaPrzejdź do nawigacjiPrzejdź do informacjiPrzejdź do stopki
24
1.Tearchitectureofthebilingualmentallexicon
simplecategories,suchascolourandshape.Featurelistsinturnarebettersuited
forqualifyingmorecomplexconcepts,e.g.acat.Sincecategorizationismotivated
byexperiencewhichisfragmentary,modality-dependentandsubjective,prototype
judgmentsshowbothinter-groupandintra-groupvariability,withdiferencesin
culture,geographicalsituationandtypeofexposureplayingacrucialrole.
Conceptsareorganizedintohierarchicalnetworksconceivedtofunctionat
threelevels:thesuperordinateandsubordinateone,andabasicorgenericlevelin
between.Itemsinthemiddleofthehierarchydisplaythemosttypicalandlorbasic
features.Teyarealsothemostinformativeandsalientperceptually.Infact,Crof
andCruse(2004)observethatitisrelativelyeasytoformaclearvisualimageforthe
medialitems,andthattheyalsohappentobethemostinclusiveandlinguistically
neutral.Basiclevelconceptsconstituteorratherinstantiatetheprototypes(Lakof
1987:32;Taylor1995;cf.Verbeemenetal.2007).
Overtheyears,Rosch(1973,1978),whopioneeredresearchinthisarea,and
others,e.g.Labov(1973),havecompiledconsistentevidenceinsupportofprototype
theory.Amongotherthings,theysucceededinestablishingthatprototypical
categoriesarementionedfirstandmostfrequentlyinexamplelistingtasks.Tey
arealsoratedasbestexamplesandgivetheshortestreactiontimesintrueorfalse
designs,e.g.arobinisabird(Geeraerts2010).Finally,theycomeupfirstin
primingtestsinresponsetosuperordinatecategoryprimes,suchasfruitorfish.
Anissuethatremainsintherealmsofspeculationistheoriginofprototypes.Here,
Roschpointstowardsanumberofpertinentfactors,includinggeneralfrequency
ofuse(nottobeconfusedwiththefrequencyofoccurrenceinlistingtests),the
orderoflearning,depictionofaveragefeaturesandsoon.Soundinmanyrespects,
Rosch’sspeculationsdidnotreceiveunanimousempiricalbacking.Inreferringto
thisissue,Taylor(1995)hypothesizesthatprototypesmostlikelyowetheiroriginto
thestabilityandadaptabilityofthecognitivedomain.Tatistosay,categoriesare
flexibleenoughtoaccommodatenewexemplars,perhapsasperipheralmembers,
withoutafectingthestabilityoftheentirecategoryinstantiatedbytheprototype.
AnexampleshowingthisprincipleatworkistheBelorussianherder,i.e.aperson
drivingaherdofcattlealongapublicroad.TeBelorussianHighwayCodeclassifies
suchindividualsasdrivers(Kiklewicz2006).
Tetheoryalsohasafewflaws.Forinstance,Goddard(1998)pointsoutthatthe
resultsofprototyperesearchmaybeartifactsofdatacollectiontoolsbecausepeople
willinglyproducetaxonomieswhenaskedtorankitemswithclear-cutboundaries.
Moreover,Aitchinson(1997:67)arguesthatpeoplerespondfastertocommonwords
andofenconfusebestexampleswith“favouriteorvaluedspecimen.”Consequently,
theirjudgmentsreflectpersonalpreferencesandculturalvaluesandbeliefs.Tere
arealsoproblemswithreliabilitycausedbythevaguenessofsomeoftheanalysed
notions.Tatis,attributeslikesimilarity(totheprototype)arenotoriouslydifcult
toqualifyobjectivelysince,firstandforemost,similarityisagradedconcept,i.e.
thingsmaybemoreorlesssimilar.Secondly,appraisalsofsimilaritytendtobe