Mapping the aesthetic space of literature “ from below ”
Christine A. Knoop
a,*
,1, Valentin Wagner
a,1, Thomas Jacobsen
b, Winfried Menninghaus
aaMaxPlanckInstituteforEmpiricalAesthetics,Grüneburgweg14,60322Frankfurt,Germany
bHelmutSchmidtUniversity,UniversityoftheFederalArmedForcesHamburg,AllgemeineundBiologischePsychologie,Holstenhofweg85, 22043Hamburg,Germany
ARTICLE INFO Articlehistory:
Received28June2015
Receivedinrevisedform31January2016 Accepted3February2016
Availableonline1April2016 Keywords:
Aestheticappealdimensions Genre
Literaryform
Terminologicalmapping
ABSTRACT
Thepresentpaperaimstoelucidatetheconceptualstructureoftheaestheticsofliterature.
FollowingFechner’s“aestheticsfrombelow”(1876)andadoptingamethodintroducedby Jacobsen,Buchta,Kohler,andSchroeger(2004),weasked1544German-speakingresearch participantstolistadjectivesthattheyusetolabelaestheticdimensionsofliteraturein generalandofindividualliteraryformsandgenresinparticular (novels,shortstories, poems,plays,comedies).Accordingtoouranalysesoffrequency,meanlistrank,andthe CognitiveSalienceIndex,beautifulandsuspensefulrankhighestacrossalltargetcategories.
Forplays/comedies,funnyandsadturnedouttobethemostrelevantterms;fornovelsand shortstories,suspenseful,interestingandromantic;andforpoetryromantic,alongwiththe music-relatedtermsharmonious,rhythmic,andmelodious.Acomparisonofourresultswith analogousstudiesforvisualaestheticsandmusicyieldedacomprehensivemapofthe distribution of aesthetic appeal dimensions across sensory modalities and aesthetic domains,withpoetryandmusicshowingthegreatestoverlap.
ã2016TheAuthors.PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCC BY-NC-NDlicense(http://creativecommons.org/licenses/by-nc-nd/4.0/).
1.Introduction
Readingliterature,justlikeviewingvisualartworks(Leder,Belke,Oeberst,&Augustin,2004)orlisteningtomusic(Brattico, Bogert,&Jacobsen, 2013),involvesperceptual, cognitive,affective andevaluativeprocessing dimensions(Jacobs,2015).
Aestheticappreciationintegratesalltheseprocessingdimensionswithafocusonevaluatingtheperceivedaestheticappealofa workofliterature,andthefelthedonicrewardoftheexposure(cf.Kant’stheoryofaesthetic“judgment”and“liking”;Kant, 2007).Whencommunicatingtheperceivedaestheticappealofaworkofliteraturetoothers,weprojectourexperienceontothe spaceofavailableverbalconceptsandselectthosewefindmostappropriatetocapturethenatureofourimpression.The resultingattributionsofparticularaestheticappealdimensionstoparticularworksorgenresofliteraturereflecttheconceptual representations(1)ofouraestheticappreciationoftheseworksor(2)oftheexpectationswehaveofcertaingenres.These attributions,i.e.theselectiveuseofspecificverbalterms,allowforstraightforwardempiricalinvestigation.Theyalsorepresent validoptionstobeconsideredwhenselectingratingscaleitemsforempiricalresearchonliterature.
Dating backas faras the18thcentury, literaryhistory recordsa number ofattempts toestablishratingscalesfor evaluatingtheperceivedaestheticappealoftheworkofindividualauthors(see Spoerhase,2014,fora review).These historicaleffortshaveanumberofanalogsinrecentliteraryscholarshipthatattempt,inlightofhistorical,psychologicaland
*Correspondingauthor.Fax:+49698300479599.
E-mailaddress:christine.knoop@aesthetics.mpg.de(C.A.Knoop).
1ChristineA.KnoopandValentinWagnerarejointfirstauthors.
http://dx.doi.org/10.1016/j.poetic.2016.02.001
0304-422X/ã2016TheAuthors.PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBY-NC-NDlicense(http://creativecommons.org/
licenses/by-nc-nd/4.0/).
ContentslistsavailableatScienceDirect
Poetics
j o u r n al h o m e p a g e : w w w . el s e v i e r . c o m / l o c a t e / p o et i c
sociologicalfindingsandbasedontheoreticalreflection,toestablishmodelsofaestheticappreciationofliterature;such modelsseektoidentifydimensionsofevaluationthatareeithercharacteristicforindividualworksofartorthatrecurinthe historyofliteraryreception(cf.,e.g.,vonHeydebrand&Winko,1996),sometimeswithaspecificfocusonindividualgenres (cf.,e.g.,Ribeiro,2012).Intheempiricalandexperimentalresearchonartreception,anumberofscaleshavebeendeveloped tocapturespecificaspectsofaestheticexperienceinreadingliterature(see,forinstance,Green&Brock,2000;Koopman, 2013, 2015; Kuijpers, Hakemulder, Tan, & Doicaru, 2014; Kuiken, Campbell, & Sopcák,9 2012; Miall & Kuiken,1995).
Nevertheless,empiricalstudiesofliteraturehavenotyetsystematicallyinvestigatedtheverbalconceptsreadersusefor communicatingtheiraestheticperceptionsandevaluationsofliteratureandtherangeanddistributionoftheseconceptsas dependentonspecificgenresofliterature.
Foranaesthetics“frombelow”asproposedbyFechner(1876),itiscrucialtocollectdataonhowreadersthemselves verballyrepresenttheiraestheticexperiencesandexpectationsratherthantorelyonexpertevaluationsandexperimenter- selectedratingscalesonly.Importantly,however,thesedifferentapproachesarebynomeansmutuallyexclusive.Infact, previousresearchinmusicreception(Zentner,Grandjean,&Scherer,2008)hasshownthatacombinedbottom-upandtop- downapproachwasbestsuitedtocapturemusicalexperiences.Similarly,arecentstudyonfilmsrevealedthatparticipants facedwithexperimenter-selectedscalesgaveconsistentlyhighratingsforitemswhichonlyfewofthemhadmentioned unpromptedwhenaskedtoverbalizetheirresponsesintheirownterms(Wassiliwizky,Wagner,Jacobsen,&Menninghaus, 2015).Thus,aselectionofdimensionsofaestheticappreciationthatexclusivelyreliesonabottom-upretrievalprocessmay besubject to serious limitations. However, collectingspontaneous bottom-up conceptualizations of the aesthetics of literaturedoeshavepotentialnotonlyforselectingitemsforratingscalesinfuturestudies,butalsoforprovidingadeeper comparativeunderstandingoftheindividualdomainsofaestheticsthroughthelensoftheirrespectiveconceptualizations.
Similartootherartdomains,aestheticperceptionsandevaluationsofworksofliteraturetendtobeshapedby,andentail, comparisons topast experiences with particular literary forms or genres (see, e.g., Picon, 1953; Schmidt, 2007; von Heydebrand&Winko,1996; Weninger,1994; Zwaan,1994).Inturn,theexpectationsderivedfromthesecomparisons influencethechoiceoftermswithwhichreadersverballydescribetheiraestheticperceptionsandevaluationsofliterature.
Thesameholdsforexpectationsderivedfromknowledgereadershaveacquiredacademicallyorviaparatextorexcerpts(cf.
Dixon,Bortolussi&Sopcak,2015).Basedontheseassumptions,thegoalofthepresentpaperistoelucidatetheconceptual structureoftheaestheticsofliterature.Thisefforthasthepotentialtorevealbothform-andgenre-specificvariationsinthe relevantvocabulary and interdependenciesand similaritiesbetween genre-specific expressions.Our approach is well establishedasalanguage-basedwayofcollectingassociationstogaininsightintothementalrepresentationofconceptual structure (cf. Fehr & Russell, 1984; Kuehnast, Wagner, Wassiliwizky, Jacobsen, & Menninghaus, 2014). Moreover, by comparingourdatawiththosefromstudiesconductedinotheraestheticdomains(Augustin,Wagemansetal.,2012;Istók etal.,2009;Jacobsenetal.,2004),wewish(1)todeterminetermsthatareusedspecificallyforliteratureand(2)toidentify overlapsbetweenaestheticappealdimensionsmentionedforcertainliterarygenresandotherdomains,includingfashion, architecture,design,physicalattractiveness,andsoon(seeAugustin,Carbonetal.,2012,forasimilarapproachcomparing visualarts,film,andmusic).Thisapproachallowsforacomparativemappingofthedomain-andgenre-specificaesthetic expectationsofnon-expertsacrossthebroaderfieldofaestheticappreciation.
2.Capturingaestheticperceptionsandevaluations
Jacobsenetal.(2004)successfullyusedafreelistingtasktocollecttermsusedfordesignatingaestheticallyrelevant dimensionsofobjectsatlarge.Usingthesamemethodology,Istóketal.(2009)conductedastudyonmusic,whileAugustin, Wagemansetal.(2012)collectedtermsdescribingaestheticappealdimensionsregardingeightdifferentgroupsofvisual objects(buildings,cars,clothing,faces,interiordesigns,landscapes,geometricshapesandpatterns,andvisualart).
Ourstudyadoptsthebottom-up,exploratory“aestheticsfrombelow”approachusedinthesethreeprecedingstudies onotheraestheticdomains.Thisplacesusinapositionnotonlytoaddthemissingdataregardingliterature,butalsotobe thefirsttocomparetheavailabledataforthedifferentaestheticdomains.Literatureistraditionallycomprisedofthree majorforms:poetry,prose,anddrama(cf.Hegel,1975).Weincludedalloftheseformsinourstudy,yettreatedthem differentlyforavarietyofreasons.Ourgeneralsubsampleforplaysencompassedthedramaticformasawhole,butthe datawerefurtherspecifiedbytheinclusionofacomedies-onlysubsample.Forprose,wecollectedsubsampledatafortwo ofthemostpopularprosegenres:novelsandshortstories.Wedidnotaskfordimensionsofaestheticappealassociated withproseasageneralcategorybecause“prose”alsoencompassesscientificandessayisticwriting,whiletheinterestof ourstudyislimitedtothemorenarrowlydefinedfieldofliterature.Incontrast,weincludedpoetryexclusivelyasabroad category, without further generic subdistinctions. We did soanticipating that (1) unlike prose, “poetry” is a fairly meaningfulcategoryevenfornon-readersofpoeticgenres,and(2)termsdesignatingspecialgenresofpoetry(suchasthe elegy,theode,orthehymn)mayseemtoospecifictothebroaderpublic.Regardingplays,weintendedtocollectdatafor theall-inclusivecategory“play/drama”,whichincludesthepolargenres “tragedy” and“comedy”; however,duetoa communicationerrorbetweentheteamswhocollectedthedata,oneteamgathereddataforthegenreofcomedyonly.
(Fortunately,thedatawedidobtainatleastpartlycompensatedforthiserror,sincethedifferencebetweentheentriesfor thegeneralcategory“play/drama”andthoseforthesubgenre“comedy”largelyappeartoreflecttheroleoftragedy.See thediscussionsectionfordetails.)
3.Terminologicalmappingofliterature
3.1.Method 3.1.1.Participants
One thousand five hundred and forty-four students of different academic disciplines participated in our study (994womenand542men,8undisclosed);themeanagewas23.5years(SD=7.28,min=17,max=84).Studentswerechosen asrespondentsforthepragmaticreasonofavailability,butalsotokeepthedesignascomparableaspossibletothestudies conductedinotheraestheticdomains(Augustin,Wagemansetal.,2012;Istóketal.,2009;Jacobsenetal.,2004),whichwere alsobasedonstudentsamples.Oftheparticipants,875weretestedinHamburgand669inBerlin,withthesubsamplesizes varyingbetween134and423participants.2
3.1.2.Procedure
The data were collected in several lecture classes of different academic disciplines, including biology, cultural anthropology,economics,law,pedagogy,history,linguistics,literarystudies,andpsychology,toaddvariancetothesamples andreducegroup-specificeffects.3Toobtaintheaestheticperceptionandevaluationtermsusedfordifferentformsand genresofliterature,wecollecteddataforthesixsubsamplesLiterature(ingeneral),Poems,Novels,Shortstories,Plays(in general),andComedies.Thestudywasconductedasapaper-and-pencilsurvey;theparticipatingstudentsreceivedthe following instruction:“Pleasewritedownterms thatcouldbeusedtodescribetheaestheticsof literature.Pleaseuse adjectivesonly.Younowhave2minutes.”Dependingonthesubsample,theword“literature”wasreplacedby“poems”,
“novels”,“shortstories”,“plays(comediesortragedies)”or“comedies”.
Thewordingoftheinstructionwaschosenforthefollowingreasons.First,wewantedtoavoidtheintricaciesofcoding andevaluatingqualitativedataofafreeresponseformatandprofitfromthestraightforwardquantitativeanalysisofword frequencyandlistrank.Thereforetheinstructionruledoutfull-sentencedescriptionsofparticipants’conceptionsregarding aestheticallyrelevantfeaturesoftherespectivegenres.Atthesametime,theinstructioninvolvedthenotionthatfull-length accountsofaestheticexpectationsarelikelytoincludemultipledimensionsratherthanjustone.Therefore,whileextracting thetermsparticipantsconsiderrelevantfordesignatingaestheticappealdimensionsdoesnotamounttoanyconcrete descriptionofanaestheticexperience,itprovidesuswithalistofwordstobeusedaspartofsuchdescriptions.Moreover, thewordingofourinstructionwaskeptassimilaraspossibletotheinstructionusedinthestudiesonotheraesthetic domains(Augustin,Wagemansetal.,2012;Istóketal.,2009;Jacobsenetal.,2004)toensurethatwewouldbeableto compareourdatawiththeirsinameaningfulway.4
Ontopofcollectingthewordlistsweloggedtheparticipants’gender,age,andacademicdiscipline,andaskedthemto indicatewhethertheyconsideredthemselvesexpertsonliterature.Allparticipantsremainedentirelyanonymous.Thedata werecollectedinGerman;theresultsdisplayedinthefiguresand tablesinthisarticlearetranslationsoftheoriginal material(seeAppendixA,TableA1).
3.2.Results
3.2.1.Samplestatistics
Weprocessedandanalyzedthedatainseveralsteps.Inthefirststep,wecorrectedspellingmistakesandcomputedthe absoluteandmeannumbersofentries.Ourparticipantsgenerated9669answersintotal,correspondingto2131different words.Thetotalfigurealsoincludesillegibleentriesandtermslistedtwicebythesameparticipant.Thenumberofanswers perparticipantrangedbetween1and21(M=6.26;SD=3.76).Regardingthenumberofentries,aneffectofthespecificityof thereferencecategoriesbecameevident(F(5,1538)=11.49,p<.001):themeannumberofentriesforliteratureingeneral (4.95)wassmallerthanthoseforthemorespecificreferencecategories,i.e.,novels(7.18),poems(6.50),plays(6.71),and comedies(6.79).Moreover,thenumberofmeanentriesforshortstories(5.78)wassmallerthanfornovelsandcomedies (thesepost-hoccomparisonswerebasedonTukey’s“HonestlySignificantDifference”method).5
2ThestudywasconductedinBerlinandHamburgbecausetheparticipatingresearchershadtheirrespectiveinstitutionalaffiliationsinthesetwocities, andthuscouldenlistthehelpofbothresearchassistantsandlecturerswillingtoletusconductourstudyintheirclasses.
3Tosecurevarietyinoursamples,wepreselectedclassesfromdifferentresearchfields(naturalsciences,socialsciences,humanities)and,withinthese fields,differentdisciplines.Thefinalselectionofclasseswascontingentupontheagreementoftherespectivelecturerstoallocatetimeforourstudyduring class.
4Fromatheoreticalpointofview,itwouldhavebeenmoreprecisetoaskfortermsdesignating“perceivedaestheticappealdimensions”ratherthanjust
“aesthetics”.However,wechosethesimpleroptionbecauseweanticipatedthatthisproxywouldbeeasiertounderstandforourparticipants.
5Differencesofthe15comparisonswith95%CIandadjustedp-values:poems-literature=1.55[0.74;2.37],p<.001;novels-literature=2.23[1.12;3.34], p<.001;shortstories-literature=0.83[ 0.12;1.78],p=.130;plays-literature=1.76[0.73;2.79],p<.001;comedies-literature=1.84[0.97;2.71],p<.001;
novels-poems=0.68[ 0.37;1.72],p=.437;shortstories-poems= 0.73[- 1.60;0.15],p=.164;plays-poems=0.21[ 0.75;1.16],p=.990;comedies- poems=0.28[ 0.50; 1.07],p=.908; short stories-novels= 1.40 [ 2.56; 0.25], p=.007; plays-novels= 0.47 [ 1.68; 0.75], p=.880; comedies- novels= 0.39[ 1.48;0.69],p=.907;plays-shortstories=0.93[ 0.14;2.00],p=.128;comedies-shortstories=1.01[0.09;1.93],p=.022;comedies- plays=0.08[ 0.92;1.08],p=1.0.
3.2.2.Resultsfortheindividualsubsamples
Inthesecondstep,termsmentionedbyfewerthan5%oftheparticipantsintheindividualsubsampleswereexcludedfrom theanalyses(forasimilarcut-offprocedure,seeIstóketal.,2009;Jacobsenetal.,2004;vanGoozen&Frijda,1993).This procedurereducesvariabilitythatmightreflectidiosyncraticuses.Fortheremaining52terms(outoftheoriginal2131)we computed relative frequency (i.e., the number of participants mentioning a given term divided by subsample size), mean list rank (basedontherawlists,beforeexcludingterms),andtheCognitiveSalienceIndex(CSI;Sutrop,2001)foreachindividual subsample.TheCSIisthequotientofrelativefrequencyandmeanlistrankandisboundedbetween0and1,withhighervalues reflectingmoresalientterms.TheresultsfortheindividualsubsamplesaredepictedinFig.S1,withthetermsrankedforeach subsample.
Overall,beautifulandsuspensefulturnedouttobelistedmostfrequently(423and269;27.4%and17.4%,respectively).
Beautifulandboringaretheonlytermsmentionedineachsamplebyatleast5%;boringalsoturnedouttobetheonlytermof unambiguouslynegativevalence.Thetermssuspensefulandinterestingrankparticularlyhighlyinthenarrative(plot-based) genres;theseinclude dramatic plot,inaccordance withAristotle’sconceptofmythos (Aristotle,2005)as avariant of narrativeplot.
3.2.3.Comparingthesubsamples
Tofurtherexaminethecommonalitiesand differencesbetweenthesubsamples,wepreprocessed thedataslightly differentlyand retainedin all subsamplesall entriesthat werementioned byat least 10% of theparticipantsin one subsample.Thisleaves22terms,asopposedtothe52termsforthe5%cut-off.Thenewfrequencypatternswereexamined bycross-tabulation(usingthe
x
2-test)totestfordifferencesbetweenthesamples.Forall15comparisonsofthefrequency patterns,thex
2-testrevealedsignificantdifferencesbetweenthesubsamples(p<.001,Bonferroni-correctedformultiple comparisons).Inthenextstep,wecalculatedtheoverlappingcoefficient(OVL;Inman&Bradley,1989;Marx,1976a,1976b) foreachpairofsubsamplesbyaddingupthelowerrelativefrequenciesinthetwodistributionsforeachofthementioned words.TheOVLranged between.32and.66, withthesmallestoverlapbetweenpoemsand comediesand thelargest betweennovelsandshortstories(seeTable1).WefurtheranalyzedtheOVLmatrixusingclassicalmultidimensionalscaling (MDS;alsocalledprincipalcoordinateanalysis;Gower,1966)andhierarchicalclusteranalysis(HCA,employingWard’s criterion).Thus,weidentifiedapatternofclustersinbothaclusteranalysisandanMDSbasedontheOVLmatrix(see Figs.S2AandB),withliteraturein generalandpoetryinparticularshowingvastsimilarities:theyarebothfrequently associatedwithtermssuchaspoetic,harmonic,rhythmical,andharmonious(seeFig.1).Thesamplesfornovelsandshortstoriesshowsubstantialoverlap.Twoimportantexceptionsstandout:notsurprisingly, shortandsuccinctrankhighlyonlyintheshortstorysample,whereasromantic,thrilling,riveting,andexcitingarelistedfar morefrequentlywithreferencetonovels.Playsand comediesalsoshowaconsiderableoverlap;nevertheless,positive emotiontermsaremorefrequentinthecomedies-onlysample,whereasnegativeemotiontermsplayalargerroleinthe generalcategoryofplays,mostlikelyreflectingthetraditionalhegemonyofthetragedyoverthecomedy.
Termsusedtodenotedimensionsofaestheticappealcanbeexpectedtohaveanevaluativeconnotation(cf.Juslin,2013);
accordingly,mostofthetermswhichwerelistedbyourparticipantsareevaluativeinnature(Jacobsenetal.,2004,reportthe samefortheirdata).However,afewofthetermsobtainedinourstudyarelessobviouslyevaluative.Rhythmical,short,and sadinparticularatfirstappeartobenothingbutpurelydescriptiveattributes.Nevertheless,ratingsofsadnesshavebeen foundtocorrelatepositivelywithratingsforaestheticappreciationinseveralstudies(Hanich,Wagner,Shah,Jacobsen,&
Menninghaus,2014;Taruffi&Koelsch,2014);similarly,ratingsconfirmthatrhythmicallyregularpoemscorrelatepositively withaestheticliking(Obermeieretal.,2013).Andeventhetermshorthas,infact,atraditionasadistinctaestheticmerit terminrhetoric(seeSection3.3.4).Therefore,therearereasonstoassumethattheuseofseeminglypurelydescriptiveterms mayalso,inthecontextofaesthetics,includeanevaluativedimension.
Table1
Descriptivestatisticsofthesubsamples.
# Subsamples Location Numberof participants
Number of answers
Number ofterms
Mean answers per participant
Numberofterms listedbymore than5%ofthe subsample
Numberofterms listedbymore than10%ofthe subsample
1 2 3 4 5
1 Literature Berlin 274 1356 591 4.95 8 7
2 Poems Hamburg 423 2751 826 6.50 18 9 0.64
3 Novels Hamburg 134 962 411 7.18 23 11 0.63 0.42
4 Short stories
Berlin 223 1288 570 5.78 16 10 0.54 0.34 0.66
5 Plays Berlin 172 1154 505 6.71 17 11 0.42 0.36 0.58 0.50
6 Comedies Hamburg 318 2158 791 6.79 16 10 0.50 0.32 0.60 0.57 0.58
overall 1544 9669 2131 6.26
ironic succinct sad witty dramatic humorous short tragic entertaining exciting funny thrilling rhythmic melodious harmonious riveting boring romantic interesting poetic suspenseful beautiful
0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5
Literature Poems Novels Short stories Plays Comedies
Fig.1.Relativefrequenciesplottedfortermsmentionedbyatleast10%oftheparticipantsinonesubsample,orderedbyfrequencyintheliterature subsample.Thegreyscaleindicateswhetherthetermwasmentioned/listedbylessthan5%(lightgrey),5%to10%(grey),ormorethan10%(darkgrey)ofthe participantsofeachindividualsample.
1.2 1.0 0.8 0.6 0.4 0.2 0.0
HCA (based on Jaccard Dissimilarity)
harmonious melodious rhythmic poetic beautiful romantic entertaining funny sad riveting thrilling exciting boring interesting suspenseful humorous ironic witty dramatic tragic short succinct
Fig.2.DendrogramoftheHierarchicalClusterAnalysisforthe22adjectives.
3.2.4.Overallmapping
Inordertointegrateallgenre-specificfindingsintoasemantic“map”ofthefieldofliteraryaesthetics,wecalculateda dissimilaritymatrixbasedontheco-occurrenceofthetermsonthelistsofallparticipantsandanalyzedthismatrixusing HCAandMDS.AsadissimilaritymeasureweusedtheJaccardIndex(Real&Vargas,1996),whichtakesintoaccountonly positivematchesandnon-matches.
Threedistinctclustersemergedfromouranalysis:onecomprisingthetermbeautifulalongwithsound-andprosody- relatedterms(e.g.,melodious,rhythmical),asecondcomprisingplot-andemotion-relatedterms(e.g.,suspenseful,thrilling, sad), andathirdthat seemedtobemoreheterogeneousatfirst,butthatcouldinasecondstepbedividedintothree subclustersthatappeartoberelatedtothespecificnaturesofshortstories,comedies,andplaysingeneral(seeFig.2).
ThedissimilaritymatrixwasalsofedintoaclassicalMDSprocedure;Fig.3depictsthetwo-dimensionalMDSsolution.6 Furthermore,basedontheMDScoordinatesofthetermsandtheirfrequenciesinthesamples,wecalculatedpointsthat representthelocalizationofthesubsamplesinthistwo-dimensionalaestheticappreciationspaceandplottedthemas vectorarrowsintotheMDSplot.Thefirst(horizontal)dimensionshowsaclearbipolarstructurewithtermsdesignating
“poetic”qualitiesandmusicalityontheoneside(withtheterm“poetic”linguisticallynotbeingrestrictedtotheverbal featuresofpoetrybutalsoincluding“poetic”feelings,atmospheres,moods,andrelatedphenomenainabroadersensethat extends to other art forms) and narrative and conceptual structure on the other. The second (vertical) dimension distinguishesbetween terms referringto emotional and potentially immersivestates at one end and humorous and intellectualstatesattheother.
−0.4 −0.2 0.0 0.2 0.4
−0.4
−0.2 0.0 0.2 0.4
literature
poems novels
short stories plays
comedies
beautiful
boring
entertaining
excitingharmonious
humorous
interesting
ironic
melodious
poetic rhythmic
rivetingromantic sad
short
succinct
suspenseful
thrillingtragic
witty
dramatic
funny
Fig.3.Two-dimensionalMultidimensionalScalingsolution.Thefiveclustersarecodedbycolor,thelocationofthesixsubsamplesinthisspaceisindicated bytheblackvectors.
6 TheGoodness-of-Fit(GOF)forthetwo-dimensionalsolutionisonly.17,andthusremainsbelowconventionalcriteria.YetinthespecificMDS-procedure weapplied,addingfurtherdimensions–whileincreasingtheGOF–leftthefirsttwodimensionsunchanged(cf.Gower,1966).Furthermore,onlythefirst twodimensions,whichbydefinitioncapturemorevariancethanthesubsequentones,appeartoreflectmoregeneralaspectsoftheaestheticspaceof literature,whereastheotherdimensionsshowastrongerfocusonmoreparticularaspects.Inanyevent,thetwo-dimensionalsolutionpresentedinFig.3 capturesonlyalimitedamountofthevarianceincludedinourdata.
3.3.Discussion
Theresultsofthisstudyhighlightthevarianceandcomplexityofhowtheperceivedaestheticappealdimensionsof literaturearemappedontoverbalconceptsbyreaders.Manyofthetermslisteddidnotmakeitpastthecut-offprocedures describedabove;byimplication,thetermsthatpassedthecut-offprocedurebynomeansrepresenttheentirevariancein aestheticappealdimensionsassociatedwiththecategoriesunderscrutiny.Nonetheless,thetermsthatendedupbeing namedbyatleast10%(22)or5%(52)oftheparticipantsinatleastonesamplestillofferasubstantial–and,aswillbecome evidentinstudy2,highlydistinctive–rangeofaestheticappealdimensionsregardingliterature.Inthefollowing,wewill individuallydiscussthemosthigh-rankingterms.
3.3.1.Beautiful
Thetermbeautiful–beinglistedmostfrequentlyoverall–provesitspreeminenceinaestheticsinthefieldofliteratureno lessthaninotherdomainsthathavebeenstudied,despitenumerousclaimsmadesincethe18thcenturythatliteratureno longerprimarilyaimsatbeauty(cf.Jauss, 1991;Lessing,1984).Tobesure,theGermanwordschönhasaconsiderablybroader application and higherusagefrequency thantheEnglish termbeautiful(Wortschatzprojekt UniversitätLeipzig, 2015).
However,contrarytoanindiscriminatebiastowardbeauty,ourdatarevealdifferencesregardingtherole oftheterm beautifulinthedifferentaestheticdomains,withthepatternsofco-occurrenceprovidingmorespecificinsightintowhat readersmightmeanwhenlabelingliterarytextsasbeautiful(seeFig.3):thetermclustersmoststronglywithromantic, followedbypoetic,rhythmic,melodious,andharmonious.Apparently,theattributionofbeautytoaliterarytextreliesmore stronglyontheprosodicqualitiesoflanguageandon“romantic”or“poetic”feelings,moods,andatmospheresthanon narrativeplotinthestrictersense.Inlinewiththispatternofco-occurrence,beautyseemstobepreeminentlyassociated withqualitiesofpoetry.Themorecolloquialmeaningsofpoetic,i.e.,soulfulandenchanting,hintattheaffectivequalitiesof thetermbeautiful.Romanticistheonlyoneoftheclusteringtermsthatcouldbereadasreferringtocertaincontentfeatures (typicaloftheRomanticperiod),butitmayequallypertaintoacharacteristicallyRomanticform,aswellas(intheeveryday senseoftheword)topowerfullysentimental,evocative,oratmosphericeffectsthatcouldbeduetobothformandcontent.
3.3.2.Suspenseful
Thetermsuspensefulisthesecondmostfrequentlymentionedtermoverall;itshighestscoresoccurintheplot-based samples.Thefactthatmostoftheformsandgenresweexaminedareplot-basedmaythereforeserveasanexplanationforthe term’soverallhighsalience.Inthesamplesfornovels,shortstories,andcomedies,suspensefulislistedmoreoftenthan beautiful.
Narrative suspensemakes readersfear ordesireparticular plotoutcomes,which canvaryin likelihood duringthe differentstagesofanarrative;thiscreatesaneedforresolutionthataccompaniestheentiresuspensefultrajectory(cf.Anz, 1998;Berlyne,1960;Carroll,1996;Fill,2007;Lehne&Koelsch,2015;Löker,1976;Wulff,1996;Zillmann,1980).Importantly, thistensefeedbackloopbetweenchangingdegreesofuncertaintyandanticipationofthepotentialresolutionstrongly dependsonthewaythecontentisnarrativelyarranged.Infact,artisticnarrativearrangementscanevenbuildsuspense whenreadersalreadyknowtheoutcomeoftheplot(cf.Gerrig, 1989;Hoeken&vanVilet,2000;Lehne&Koelsch,2015;Yanal, 1996).Suspensealsoservestoincreasethegeneralemotionalsusceptibilityoftheaudience(cf.Oatley, 1999;Vorderer,Wulff,
&Friedrichsen,1996).Consistentwiththesetheoreticalassumptions,thetermsuspense clustersstronglywithemotion termssuchasthrillingandrivetinginoursample.
3.3.3.Interesting
Theterminterestingisthefourthhighestscoringterminourliterature,novel,andshortstorysubsamples.Intheoretical aesthetics,theattribute“interesting”hasbeenacknowledgedasanessentialcategorysincethe18thcentury(cf.Diderot, 1995; Garve, 1974; Ostermann,1997; Sulzer,1967). Friedrich Schlegel in particular developed a concept of aesthetic evaluationinwhichtheinteresting,notthebeautiful,isthemainreferencepoint(Schlegel,2001).Morerecentempirical researchsupportsthenotionoftheinterestingasaprimecategoryofaestheticappreciation(Silvia,2005,2010).
Ourpoetrysubsamplescoresverylowonthedimensioninteresting,whichmaybeduetovariousreasons.Firstly,since the18thcenturytheinterestinghasbeenconsideredtoappealtotheintellectratherthantheheart,whichiswhyitwas primarilyappliedtonovelsandplays;poetry,bycontrast,longcontinuedtobelargelyconceivedasfocusingonhighly personalfeelingsandonbeautifulpoeticdiction(Hegel,1975;Lukács,1971;Schlegel,2001).Secondly,inaccordancewith Schlegel’s(2001)theoreticalreflections,Silvia(2010)suggeststhatahighdegreeofnoveltyisanimportantfactorforfinding somethinginteresting.However,thepoetrywithwhichaveragereaderstypicallycomeintocontact(inschool,atuniversity, inprintedanthologies,ongreetingcards,atweddingorbirthdaycelebrations)isoftenrepresentativeoftraditionallyrical forms and motifs; contactwithlyrical texts thata contemporary readerwould deem innovative,and therefore more interesting,israrebycomparison(cf.Shetley,1993).
3.3.4.Otherterms
Overall,wefoundahighnumberofemotion-relatedterms.Needlesstosay,emotiongenerallyplaysanenormousrolein theaestheticexperiencewithandtheaestheticevaluationofliterature.Theparticularlyhighscoresofemotiontermsinthe
listsforplaysingeneralandcomediesinparticularconfirmthenotionthatdramaasagenreaimsprimarilyatevokingstrong affectiveresponsesinaudiences(Aristotle,2005;Zillmann,1980,1983).
Twofrequentlyusedqualityjudgmentsregardingworksofliterature,namely,attributionsofgoodnessorbadness(“a goodnovel”,a“badplay”),areentirelylackingfromourdata.Attributionsoftheadjectivesgoodandbadreflectaparticularly abstractandunspecifictypeofaestheticjudgmentthatfailstoprovidehintsregardingeithertextualfeaturesorexperiential responsedimensions;theexplicitinstructiontolistaesthetics-specificadjectivesmayhavepreventedourparticipantsfrom listingthissimpledichotomy.Theabsenceofthesetermsmaythuswellbeataskeffect;confirmingthisassumption,the terms“good”and“bad”werelikewisenotlistedinthestudiesonotheraestheticdomainsthatusedsimilarinstructions (Augustin,Wagemansetal.,2012;Istóketal.,2009;Jacobsenetal.,2004).Similarly,thetermpleasantdoesnotappearinour results,despitebeingoneofthemorefrequentlyemployeditemsinaestheticratingscales.Thismight,again,beduetothe relativelyunspecificnatureoftheterm,designatingasitdoesabroadrangeofpositiveexperiencesthatcouldbemediated byagreatvarietyofemotionsandotherresponsedimensions,aswellasaestheticallyrelevanttextfeatures.
Theanalysisof theterms listedfor poetryreveals a strongbiasfor music-related terms, mostnotablyrhythmical, melodious,andharmonious.Theseresultscoincidewithalong-standingtraditionofequatingpoetswithsingersandpoems withsongs(e.g.,Hegel,1975;Herder,1998;Nietzsche,1993),whichbecameacornerstoneoftheRomanticconceptofpoetry andhasshapedprototypicalnotionsofpoetryeversince.Ourbottom-updatademonstratetheenduranceofthisconceptfar beyondascholarlyawarenessoftheunderlyingintellectualtraditions.Atthesametime,termsdesignatingnarrativeand conceptualproperties,suchasthrilling,riveting, andsuspenseful,which prove very importantfor thenovel,are rarely mentionedinthepoetrysubsample;thissupportsthenotionthatbothformalandcontent-relatedcharacteristicsofliterary genresclusterwithspecificaestheticevaluations(cf.Ribeiro,2012).
Asweexpected,thedataobtainedfornovelsandshortstoriesclusterclosely.Yetwealsofoundsomedifferencesthatare worthacloserlook.First,thefrequenciesfortheadjectiveshortsetthegenresneatlyapart.Onemightdismissthisasan indicationthatourparticipantsconfoundedthenameofthegenrewithitspotentialaestheticmerits;however,shortness (gr.brachytes,lat.brevitas)isinfactawell-establishedcategorydesignatingtherhetoricalachievementofcondensinga messageandmakingitunusuallycompact(Quintilian,1953;foranempiricalstudyoneffectsofrhetoricalbrevityinthe processingofproverbs,cf.Menninghausetal.,2015).Infact,ashortstorythatisnotshortwouldviolategenreexpectations, includingaestheticrewardexpectations.Thus,thehighscoresforshortmaynot,afterall,beameretaskeffect.
Second,thetermsromantic,thrilling,riveting,andexcitingwerelistedveryfrequentlyfornovelsbutdonotseemtobe expectedpropertiesofshortstories—whichmightbafflereadersofPoe,Maupassant,orChekhov.Onepossibleexplanation forthiscouldbefoundintheGermancurricula:theGermanshortstoryafter1945isfeaturedprominentlyinGerman secondaryschools;therefore,mostorallofourparticipantswillhavebeenexposednotonlytoexamplesofthegenreitself, butalsotoacatalogofitsspecificfeatures:thistypeofshortstorycommitstostraightforward,openstorylineswhile forgoingcomplicatedchainsofcauseandeffect,highlyemotionalsuspense–resolutionsequences,andevenunambiguous endings(Weyrauch,1989).Knowledgeofthisparticulartraditionandresultinggenreexpectationsmightbereflectedinthe lowerratingsforexciting,thrilling,andrivetinginourshortstorysubsamplecomparedtothenovelsubsample.
3.3.5.ThetwodimensionsoftheMDSsolution
Termsreferringtopoeticqualitiesofatextandtermsreferringtoitsnarrativeand/or thematicstructureseemto form the opposite ends of the horizontal axis of our MDS solution (see Fig. 3). These poles are most markedly representedbythetermsbeautiful,whichrefersmostlytothepoeticandformalappealoftexts,andsuspenseful,which primarilyreferstoplottrajectories.However,thisresultshouldnotbereadassupportingaform-contentdissociation.In fact,theplotsideofthespectrumactuallyrepresentsnotjustcontentalone,butaparticulartypeofinteractionbetween formand content.
TheverticalaxisofourMDSgraph(seeFig.3)showstermsthatrefertomoreintellectualaspectsofreading(witty, humorous,ironic)ontheonehand,andterms thatpertaintothetext’semotional contentortheaffectiveresponsesit mayevoke(e.g.,suspenseful, sad, boring)ontheother (fora basis forthis distinction betweenemotions perceivedas beingrepresentedbyaworkofartandtheactualfeltemotionsoftheaudience,cf.Juslin,2013;seealsoRegel,Gunter,&
Friederici,2011;Nagelsetal.,2013).Notably,inthis graphcomediesandplaysingeneral (whichbydefinitioninclude comedies)areclearlyseparated:comparedtocomedies,playsingeneral arealotfarther removedfromthepolethat entailswitty,humorous,andironic,and farclosertothepolethatentailssuspenseful,sad,boring.Wesurmise thatthe differencesbetweentheresponsesgivenforcomediesonlyandforplaysingeneralprimarilyreflecttheroleoftragedy.
Inaccordance withthis interpretation,thecategoryof playsingeneral includesthetermtragic,whichis listedmuch morefrequently than humorous and witty(although less frequently thanfunny), and also thetermsad,which is the highest-scoringtermintheentiresubsample. Inalllikelihood, separatedatafor tragedywould havefurtherenlarged this distance betweenplaysin general and comedies, because theinclusion of comedies in thecategory of plays in generalshouldkeepthisdistanceatamoremoderatelevel.Thus,thelackofaseparatedatasetfortragediesisatleast partly compensated by the differences between our data for comedies and for plays in general. In light of this interpretationofthedata,itappearsthatfromanaffectiveperspective,thetwodramaticgenresofcomedyandtragedy areperceivedtohavetwoentirelydifferentaffectivesignatures,notsimplyinvalence(positive negative)andaffective content(funny sad),butalsoregardingthequalityof theevokedaffects andtheextenttowhich theyarelinkedto cognitive processes(cf.,e.g., Taylor,1988; Morreall,1983).
Whilethetermsonthemoreintellectualendofthespectrumhaveaprimarysemanticfocusonappealdimensions ascribedtothetext(e.g.,witty,humorous,ironic),theaffectiveendofthespectrumentailsalargenumberoftermsthatplace aprimaryfocusonthetext’s(potential)effectonthereader(e.g.,suspenseful,exciting,thrilling).
4.Comparingliteraturewithotherdomains
4.1.Studiesanddata
Inordertocompareourresultsregardingthedomainofliteraturewithother(aesthetic)domains,weconductedanalyses includingtheresultsofourstudyandtheresultspublishedbyJacobsenetal.(2004)onaestheticobjectsatlarge,Istóketal.
(2009)onmusic,andAugustin,Wagemansetal.(2012)oneightdifferentclassesofvisualobjects:buildings,cars,clothing, faces,interiordesign,landscapes,geometricshapesandpatterns,andvisualart.Althoughthestudiesallemployedafree listingmethod,there areseveral differencesbetweenthem:theydifferedregarding thelanguagein whichtheywere conducted(German,Finnish,orDutch),restrictionsontheanswerformat(onlyadjectivesvs.unconstrained),andthetime accordedforthetask(2min,5min,orunconstrained);foranoverview,seeTable2.
Thetermsmentioned byatleast5%in therespective samples,along withtheirfrequencies,werethebasis forthe followinganalyses(thesedataareavailableaspartofthepublishedstudies;forourdata,seeSection3.2.2).Basedonthese frequencydata,wecalculatedtheOVLforeachsamplepairing(seeTableS1),which wefedintoanHCAand anMDS procedure.
4.2.Results
Unsurprisingly, individualsamples tend toshow thegreatestoverlapwithothersamples fromthesame domain;
however,weidentifiedafewinterestingexceptions.Literatureingeneralprimarilyintersectswithnovelsandpoems,but whencomparedtononliterarydomains,itshowsthegreatestoverlapwithvisualarts.However,theliterarysubdomainof poems–again,whencomparedtononliterarydomains–showsthegreatestoverlapwithmusic,followedbythevisual artsandvisualaestheticobjectsingeneral.Thetermsmentionedformusichavethegreatestoverlapwiththoselistedfor poetry, followedbyvisualart.Visualart,inturn,showedthegreatestoverlapwithclothing,followedbyliteraturein general, andthen visualaesthetic objects in general,faces, andcars. Hence,our dataalso reveal someconsiderable similaritiesbetweenthetermsusedtoconceptualizetheaestheticappealdimensionsofvisualart,music,poetry,and literatureingeneral.
Moreover,wetransformedtheOVLmatrixintoadissimilaritiesmatrixandfeditintoanHCAprocedure,usingWard’s method.Theresultshowstwoclearlyseparateclusters,oneofwhichcomprisesmostsamplesofthevisualdomain.The otherclusterincludestheliterarysamples,aswellasthemusicandlandscapesamples.Bothclusterscanbefurtherdivided intotwosubclusters:withinthevisualcluster,manufacturedartifactsaregroupedtogether,andsoarefaces,geometrical Table2
Overviewofthecomparisonbetweenstudies.
Study Year Objectclass N Language Answer
format
Design Procedure Time
Jacobsenetal.(2004) 2004 Objects 311 German Adjectives – Paper&
pencil
2min
Istóketal.(2009) 2009 Music 300 Finnish Adjectives – Paper&
pencil
5min Augustin,Wagemansetal.
(2012)
2012 Buildings 178 Dutch No
restriction
Within participant
Online No
restriction
Cars 177
Clothing 175
Faces 175
Geometricshapes&
patterns
173
Interiordesign 175
Landscapes 177
Visualart 177
Ourstudy Literature 274 German Adjectives Between
participant
Paper&
pencil
2min
Poems 423
Novels 134
Shortstories 223
Plays(tragediesor comedies)
172
Comedies 318
patterns,andvisualaestheticobjects.Withintheliterarycluster,onesubclustercomprisesthenarrativegenres,andthe othercomprisesliteratureingeneral,poems,music,andlandscapes(Fig.4).
AnMDSbasedonthedissimilaritiesmatrixlargelymirrorstheresultoftheclusteranalysis(seeFig.5).7Again,wefinda clustercomprisedoftheplot-basedliterarygenres,andonecomprisingliteratureingeneral,poems,andmusic,thelastthree clusteringsomewhatmoreloosely.Thereareslightdifferencesinthetwonon-literature-relatedclusters:onecomprises manufacturedartifactsofpracticaluse,includinggeometricalpatterns,butnotthevisualarts;theothercontainsthevisual arts,landscapes,faces,andvisualaestheticobjectsingeneral.AninterpretationofthetwodimensionsofthisMDSisnot straightforward.
4.3.Discussion
Aparticularlyinterestingresultofthecomparisonofourdatawiththoseofpreviousstudiesisthestrongoverlapbetween thetermsusedformusicandpoetry;thisunderscoresourfindingthatthephonological(includingprosodic)propertiesof poetry,alongwithitsemotionalaspects,aredeemeditsmostdistinctivecharacteristics.Moreover,thecontentofpoetry– consisting,asittypicallydoes,ofsituationalminiaturesanddepictionsofmomentarymoodstates(cf.Schlaffer,2012)– usuallydoesnotinvolveanythingresemblingafull-blownnarrativeplot.Asaresult,poetryappearstofavoracomparison withmusicoveracomparisonwithotherliterarygenresthatarebasedonplot.Ofcourse,anumberofsubgenresmight renderthisclear-cutsolutionsomewhatcomplicated:thenarrativelystructuredballad,themeteredandsometimeseven rhymedtragedy,freeverse.However,whileourdataforpoetrydonotcovertheseandothersubgenres,theydoreflectnot merelytheexpectationsofourparticipants,butalsoimportant–ifnotpredominant–featuresofpoetry.
1.5 1.0 0.5 0.0
visual arts clothing interior design cars
buildings patte rns objects faces plays comedies novels short sto ries literature poems music landscapes
Fig.4. DendrogramoftheHierarchicalClusterAnalysisforthe16objectclassesinthecomparedstudies.
7 Thetwo-dimensionalMDSsolutiononlyyieldedaGOFof.41.However,aswasthecaseforthefirstMDSreportedinthispaper,thepatternfoundforthe twofirstdimensionsdidnotchangeasweaddedmoredimensionstotheMDSsolution.WeherereportthisMDSsolutiononlyasanadditionalvisualization ofthedataunderlyingtheclusteranalysis.
Generally, theartisticdomains (literature,visual arts,music)sharemany ofthe termsused toconceptualizetheir aestheticappeal;thiscorrespondstothewayinwhich18thcenturyaestheticscomprisedallarts,whichuntilthenhadbeen consideredexclusivelyasindividualartes,underthenewall-encompassingsingularofart(cf.vonSchelling, 1989).Moreover, ourfindingthatliteratureingeneral,whencomparedtononliterarydomains,hasthegreatestoverlapwithvisualartscanbe readassupportingthelong-standingutpicturapoesistradition(cf.Horace,2011;Lessing,1984).
Insomerespects,however,wefoundcleardomain-baseddifferences:ugly,forinstance,islistedonlyforvisualobjects andmusic.Itappearsthatuglinessinliterature,ifitcanbefoundatall,isamatterofsemanticcontentratherthanaesthetics (Eco,2007).Theclearestantonymtobeautifulintheliteraturesampleisthetermboring(cf.Lorand,1994)–whichinturnis barelyrelevantfortheotherdomains.
5.Generaldiscussion
5.1.Keyresults
(1) Inallstudiestodate(includingours)thatanalyzethelinguistictermsusedtodesignateexpectedorperceived aestheticappealdimensions,theconceptofbeautyhasbeenfoundtobeprevalentinalltesteddomains.Atthesametime, beautifulisbynomeansthemostfrequentlymentionedtermineachofthedomainsinourstudy.Itscoredhighestinthe subsamplesfor poetryandliterature ingeneral;yetintheplot-basedsubsamples,othertermsturnedouttobemore significant:suspensefulinthecaseofthenovelaswellastheshortstory,funnyandsadinthesubsampleofplaysingeneral, andsuspenseful,funny,witty,andironicinthecomedysubsample.Nevertheless,countingalltermslistedintheentirestudy, beautifulismentionedmostfrequentlyoverall.Regardingonlythecategoriesofliterature,whichwereinvestigatedinthe presentstudy,attributionsofthetermbeautifulclusterstronglywithtermsreferringtothe“poetic”qualitiesofdictionand emotionaltonality,andfarlesswithtermsthataremorecloselyassociatedwithplottrajectories.Overall,theselectionof verbalconceptswegatheredrevealsthat,inmanyinstances,conceptswithalong-standingtraditioninaesthetictheoryplay animportantroleinnon-expertconceptualizationsoftheaestheticsofliterature;thisappearstoapplyfarbeyondany scholarlyawarenessofthistradition.
−0.4 −0.2 0.0 0.2 0.4
−0.4
−0.2 0.0 0.2 0.4
literature
poems novels
short sto ries plays comedies
music
objects visual a rts landscapes
faces patterns
cars clothing interior design
buildings
Fig.5.Two-dimensionalMultidimensionalScalingsolutionforthe16objectclassesofthecomparedstudies.
(2)Thecomparisonofpreviousstudieswithourownfindingsunderscoresthecommonalitiesbetweenmusicandpoetry thatcanalreadybeextractedfromourowndataalone.Theconsistentdatapatternconfirmsthatthesecommonalitiesare notonlyseenbyexpertsfamiliarwiththelongstanding“Romantic”traditionoftheorizingpoetry,butareseeminglyalsopart ofthewayreadersconceiveofpoetrytothisveryday.
(3)Regardingtheaestheticsofliterature,ourdatashowtwopreeminentfactors:plot-basedappealdimensions,which areprimarilyimportantfornovels,shortstories,andplays,and“poetic”appealdimensions,whicharemostlybasedon prosody-andmusic-relatedaspectsaswellasonspecialemotionaltonalities.Thissecondgroupofappealdimensionsisby nomeansexclusivelyrelatedtopoetry(whereitisparticularlystrong),butextendsintothebroadersampleofliteraturein general.
(4) Our approach “frombelow” identifies verbalconcepts which, together, forma conceptual map of prototypical aestheticperceptionsandevaluationsregardingparticularliterarygenres.Theparticipantsinourstudydidnotevaluate individual works of literature, but listed expected dimensions of aesthetic appreciation regarding literary genres;
furthermore,theychosetheirownwordingratherthanrespondingtoexpert-selecteditems.Nodataofthistypehasbeen collectedinpreviousempiricalresearchonliterature.
5.2.Limitations
First,itisimportanttonotethattheappealdimensionsextractedinourstudyrelyonabstractrepresentationsofgenre concepts.Inalllikelihood,theseabstractconceptsareidealizedentitiesthatdonotcoverthefullhistoricalvarianceofthe concreteworksofliteraturefoundacrossmultiplelanguagesandcultures.Moreover,ourdataexclusivelyreflectappeal dimensionsidentifiedbyasampleoftoday’sreadersandarelikelytobesubjecttoongoingchange.Therefore,therangeof dimensionsofappealidentifiedheremustbeconsideredasafirst,rathergeneralguidepost.
Second,thefourstudiesontheconceptualstructureofspecificaestheticdomainsthatwecomparedwereconductedin differentlanguages:German,Finnish,andDutch(seeTable2).Alldata,includingourown,havebeentranslatedintoEnglish.
Therefore,anycomparisonmustbearinmindlanguage-specificdifferencesinusagefrequency,emotionalvalence,habitual connotations,and,notleast,semanticsbetweentheselinguisticgroups(forotherproceduraldifferencesseeTable2).
Third,collectingfreelistingdatabydefinitiondoesnotprovideanyinformationabouttheparticipants’previousgenre exposureandpreferences,thoughthesemayhaveinfluencedtheresultswereceived(regardingthepotentialimpactof genreexposure,seeFong,Mullin,&Mar,2013).Thetaskisalsonotsuitedtocapturepreviousexperiencesthatmayhave promptedourparticipantstoincludeparticularitemsintheirlists.Moreover,wecannotextrapolatewhichrepliesmayhave beenprimedbytheoreticalconceptsoftherelevantgenres(e.g.,conceptsourparticipantsacquiredinschool)andwhich oneswerearesultofownreadingexperiences.
5.3.Futureresearch
Toacquireadeeperunderstandingofhowrecipientsperceiveandconceptualizethedimensionsofaestheticappealfor differentartforms,andtoprovideamapofthedimensionsofaestheticappreciation,additionaldataforotherartistic domainswillbeindispensable.Thisincludesmultimodaldomainslikefilm,opera,performanceart,anddance,whichin manywaysbridgethegapbetweenliteratureandmusic,literatureandtheatricalperformance,andliteratureandthevisual arts.Attheotherendof thescale,itwouldbebeneficialtomakefinerdistinctionswithintheindividualdomainsand subdomains–forinstance,asubdomainlikethenovelcouldbedividedintosmallerandmorespecificsubgenres,suchas sciencefictionnovels,detectivenovels,romancenovels,andsoon.Furthermore,cross-linguisticaswellascross-cultural comparisonsforthesameartisticdomainwouldbeveryusefulinordertoestablishtheextentofvarietyweshouldaccept whentranslatingitems.Lastly,astudysuchasours,whichisbasedongeneral,openquestions,couldbecomplementedwith acorpus-basedapproachbyextractingadjectivesappraisingliteraturefromcontemporaryandhistoricalnewspapersand journals,onlineandprintreviews,blurbs,andothertypesofparatext.Anotherpossibilitywouldbeempiricalresearchthat collectsandanalyzesmorecomprehensivecommentsonthetopicofliteraturefromresearchparticipants(forastudyalong theselines,cf.Dixon&Bortolussi,2009).Inthisway,themoregeneralaestheticappealdimensionswerecordedwouldbe complementedbytermsthatreadersusetoconceptualizetheirimpressionsregardingavarietyofspecificliterarytexts.
Acknowledgments
WewishtothankMichaelBerners,JohannesBohn,NoamEshel,SabineKüster,MartinLohr,AquilesLuna-Rodriguez, AlexanderNolte,FlorianSchefenacker,AnnaWeinbrecht,andMikeWendtfortheirhelpindatacollectionandprocessing, FelixBernoullyforassistingwiththedesignofourfigures,andMariaKraxenbergerforherhelpfulcommentsonanearlier versionofthemanuscript.Apartofthisresearchwasconductedattheresearchcluster“LanguagesofEmotion”,fundedby theGermanResearchFoundation(DFG).
AppendixA.
OriginalGermaninstruction
“BitteschreibenSieWörterauf,diemanzurBeschreibungderÄsthetikvonLiteraturverwendenkann.BittebenutzenSie nurAdjektive(Eigenschaftswörter).Siehabenabjetzt2MinutenZeit.”
TableA1
Termslistedbyatleast5%oftheparticipantsinonesubsample.
English(translation) German(original)
Aesthetic ästhetisch
Amusing amüsant
Beautiful schön
Boring langweilig
Brief knapp
Charming lieblich
Comical komisch
Creative kreativ
Creepy gruselig
Deeplymoving ergreifend
Diverse abwechslungsreich
Diverting kurzweilig
Dramatic dramatisch
Emotional emotional
Entertaining unterhaltend
Entertaining unterhaltsam
Euphonious wohlklingend
Exciting aufregend
Fantastic phantastisch
Fluent fließend
Funny lustig
Harmonious harmonisch
Humorous humorvoll
Imaginative phantasievoll
Instructive lehrreich
Interesting interessant
Ironic ironisch
Long lang
Long-winded langatmig
Melancholic melancholisch
Melodic melodisch
Melodious klangvoll
Metaphorical metaphorisch
Moving bewegend
Open offen
Passionate leidenschaftlich
Poetic poetisch
Pointed pointiert
Profound tiefgründig
Realistic realistisch
Rhyming reimend
Rhythmic rhythmisch
Riveting fesselnd
Romantic romantisch
Sad traurig
Sarcastic sarkastisch
Satirical satirisch
Sentimental gefühlvoll
Short kurz
Stimulating anregend
Succinct prägnant
Surprising überraschend
Suspenseful spannend
Thrilling mitreißend
Touching berührend
Tragic tragisch
Witty witzig
