Voice pitch modulation in human mate choice

Article (Accepted Version)

http://sro.sussex.ac.uk

Pisanski, Katarzyna, Oleszkiewicz, Anna, Plachetka, Justyna, Gmiterek, Marzena and Reby, David (2018) Voice pitch modulation in human mate choice. Proceedings of the Royal Society B: Biological Sciences, 285 (201816). pp. 1-8. ISSN 0962-8452

This version is available from Sussex Research Online: http://sro.sussex.ac.uk/id/eprint/80518/

This document is made available in accordance with publisher policies and may differ from the published version or from the version of record. If you wish to cite this item you are advised to consult the publisher’s version. Please see the URL above for details on accessing the published version.

Copyright and reuse: Sussex Research Online is a digital repository of the research output of the University.

Copyright and all moral rights to the version of the paper presented here belong to the individual author(s) and/or other copyright owners. To the extent reasonable and practicable, the material made available in SRO has been checked for eligibility before being made available.

Copies of full text items generally can be reproduced, displayed or performed and given to third parties in any format or medium for personal research or study, educational, or not-for-profit purposes without prior permission or charge, provided that the authors, title and full bibliographic details are credited, a hyperlink and/or URL is given for the original metadata page and the content is not changed in any way.

Voicepitchmodulationinhumanmatechoice

Katarzyna Pisanski1,2*, Anna Oleszkiewicz2,3, Justyna Plachetka4, Marzena Gmiterek2,

& David Reby1

1 – School of Psychology, University of Sussex, UK

2 – Institute of Psychology, University of Wroclaw, Poland

3 – Taste and Smell Centre, Department of Otorhinolaryngology, Technische Universität

Dresden, Germany

4 –SWPS University of Social Sciences and Humanities, Wroclaw, Poland

* Corresponding author: K. Pisanski (e-mail: kasiapisanski@gmail.com)

Running head: Voicemodulationinhumanmatechoice

Word count: 5702 (incl. everything except Tables)

Abstract

Inter-individualdifferencesinhumanfundamentalfrequency(F0,perceivedasvoice

pitch)predictmatequality,reproductivesuccess,andaffectlisteners’social

attributions.Althoughhumanscanreadilyandvolitionallymanipulatetheirvocal

apparatusandresultantvoicepitch,forinstanceintheproductionofspeechsounds

andsinging,littleisknownaboutwhetherhumansexploitthiscapacitytoadjustthe

nonverbaldimensionsoftheirvoicesduringsocial(includingsexual)interactions.

Here,werecordedfull-lengthconversationsofthirtyadultmenandwomentakingpart

inrealspeeddatingevents,andtestedwhethertheirvoicepitch(mean,range,and

variability)changedwiththeirpersonalmatechoicepreferencesandtheoverall

desirabilityofeachdatingpartner.Within-individualanalysesindicatedthatmen

loweredtheminimumpitchoftheirvoiceswheninteractingwithwomenwhowere

overallhighlydesiredbyothermen.Menalsoloweredtheirmeanvoicepitchondates

withwomentheyselectedaspotentialmates,particularlythosewhoindicateda

mutualpreference(matches).Interestingly,althoughwomenspokewithahigherand

morevariablevoicepitchtowardmentheyselectedaspotentialmates,women

loweredbothvoicepitchparameterstowardmenwhoweremostdesiredbyother

womenandwhomtheyalsopersonallypreferred.Between-individualanalyses

indicatedthatmeninturnpreferredwomenwithlower-pitchedvoices,wherein

women’sminimumvoicepitchexplainedupto55%ofthevarianceinmen’smate

preferences.Theseresults,derivedinanecologicallyvalidsetting,showthatindividual

andgroup-levelmatepreferencescaninteracttoaffectvocalbehaviour,andsupport

thehypothesisthathumanvoicemodulationfunctionsinnonverbalcommunicationto

elicitfavourablejudgmentsandbehavioursfromothers,includingpotentialmates.

Keywords:matechoice,sexualselection,speeddating,nonverbalcommunication,

fundamentalfrequency,vocalcontrol

Thehumanvoiceconveysevolutionarilyandsociallyrelevantinformationthataffects

theoutcomesofmatechoiceandintrasexualcompetition.Todate,mostresearchinthis

areahasfocusedonmen,whosefundamentalfrequency(F0,perceivedaspitch)

appearstohavebeenshapedbysexualselectiontocommunicatemasculinity,physical

formidability,andgeneticquality[1].Menwithlow-pitchedvoicestypicallyhave

higherlevelsofpubertal[2]andcirculating[3,4]testosterone,andareperceivedas

moremasculineanddominantthanaremenwithhigher-pitchedvoices[5forreview].

Lowpitchvariability,perceivedasamonotonevoicequality,alsopredictsmen’s

formidability[6].Inturn,alargeliteraturedemonstratesthatwomengenerallyprefer

menwithrelativelylow-pitchedvoices,particularlyinthecontextofashort-term

relationship,andthatsuchmenreportmoresexualpartnersandhavemoreoffspring

thantheirhigher-pitchedcounterparts[7].Infact,menwithlow-pitchedvoicestendto

havehighersuccessinarangeofsocialcontexts,frommatingtosocioeconomicand

political[5,7].

Amongwomen,between-individualdifferencesinvoicepitchareunderstudied,but

appeartoindicatereproductivestatus(e.g.,pre-pubertalandpost-menopausalstages

offertility)andsexualreceptivity[7,8].Womenwithrelativelyhigh-pitchedvoicesare

typicallyperceivedasmorefeminine,younger,andmoreattractivethanarewomen

withlow-pitchedvoices[9–12].However,severalstudiessuggestthatalowerpitchin

womenisperceivedasattractiveor‘sexy’[13–16],whileothersfailtoidentifyany

relationshipbetweenfemalevoicepitchandjudgmentsofattractiveness[6].In

contrast,womenwithlow-pitchedvoicesareconsistentlyjudgedasmoredominant,

competent,andmature,andasbetterleadersthanwomenwithahighervoicepitch

[12,17,18].Thetrade-offimpliedbythisdichotomysuggeststhatwomenmay

volitionallyraisetheirvoicepitchtosignalyouthandfemininity,butlowertheirpitch

incontextswheretheywishtobetakenseriously,ortoindicatesexualinteresttoa

listener(seee.g.,[13]).

Lowfrequencyvocalisationssignaldominance,notonlyinhumans,butinawiderange

ofanimals.Manyspeciesvocalisewithalowervoicepitchduringagonisticinteractions,

therebycommunicatingaggressionandthreat,andinsomespecies(e.g.,reddeer

Cervuselaphus,andkoalasPhascolarctoscinereus),malesextendthelengthoftheir

vocaltractbyloweringtheirlarynx(andthustheirformantfrequencies),thereby

portrayingalargerbodysizetoothermalesincompetitivecontexts,aswellasto

potentialfemalemates[19,20].Importantly,suchfrequencychangesobservedinthe

callsofmostnonhumanmammals,includingprimates,typicallyoccurinresponseto

physiologicalorenvironmentaltriggers.Humans,incontrast,arereadilycapableof

volitionally(i.e.,voluntarily)manipulatingtheirvocalapparatus,owingtoincreased

neuralcontrol[21,22].Perhapsthemostcommonformofvolitionalvoicemodulation

inhumansinvolvesmanipulatingthearticulators,includingthelips,tongueandjaw,to

producedifferentvowelsoundsinarticulatedspeech[23].Wecanalsoactively

modulatethepitchofourvoicesbyadjustingthetensionandtheeffectivelengthofthe

vocalfolds,orbyincreasingsub-glottalpressure(i.e.,airflowfromthe

lungs)[24].Pitchmodulationisimportantforintonation,prosody,andemotional

expression,andisalsoobservedinactingandsinging[22,25].

Althoughthecapacityforvoicepitchmodulationinspeechproductioniswell

documented,whetherhumansexploitthiscapacitytoadjustthenonverbaldimensions

oftheirvoicesduringsocialinteractionshasnotyetbeensystematicallyinvestigated.

Theabilitytovolitionallychangethepitchofourvoicemaybeevolutionarily

advantageous.Duringsocialandsexualcommunication,suchvocalmodulationcould

functiontohonestlycommunicateone’smotivationsandemotions,butmayalsobe

usedtofavourablymanipulatetheperceptionsandbehavioursofothers,including

potentialmates[22].Indeed,menandwomencanvolitionallylowertheirvoicepitch

wheninstructedtosoundmoremasculine[26]orphysicallylarger[27],andbothsexes

modulatetheirpitchwheninstructedtosoundconfident,dominantandintelligent[14].

Thus,notonlyarehumanscapableofvocalcontrol‘ondemand’,butalsooffine-tuning

voicemodulationstomimicrealorperceivedassociationsbetweenphysicaland

psychologicaltraitsofaspeakerandtheirvoice.

Studiesexaminingvoicemodulationinmatingcontextshave,todate,onlybeen

conductedinthelaboratoryusingwidelyvariedmethodologies,andhaveproduced

mixedresults[13,28–31].Somestudieshavefoundthatmenspeakwithalowerand

lessvariablevoicepitch[13,29]oralowerminimumvoicepitch[31],whereasothers

reportanincreaseinmen’smeanpitchandpitchvariability[28,31]whenspeaking

withahypotheticalfemalemate.Likewise,womenhavebeenshowntoeitherincrease

[30],decrease[13,14],ornotaltertheirvoicepitch[31]whenspeakingtohypothetical

malemates.Inthesestudies,datingcontextsweremocked(e.g.,participantswere

askedtoleaveavoicemessageforanunknownpersonwhosephotoorvideothey

viewed),andtheattractivenessofhypotheticaldatingpartnerswaseithernot

controlledorwaspre-ratedbyanothergroupofparticipants.

Here,werecordedthevoicesofthirtyadultmenandwomen(aged20-40)takingpart

inrealspeeddatingevents,andexaminedwithin-individualchangesintheirvoicepitch

parametersasafunctionoftheirdate’soveralldesirability(i.e.numberofsuccessful

speeddates),andimportantly,theirownpersonalmatechoicepreferenceforeach

datingpartner.Thepresentstudyisthefirsttoexaminehumanvoicemodulationina

real-worldmatechoicescenario.Withthishighdegreeofecologicalvalidityand

multiplematepreferencemeasures(individualandgroup-level),thisstudyaddresses

limitationsofpastworkthatlikelycontributedtoinconsistentfindings.

METHODS

Participants

Thirtyparticipantswererecruitedbyaprofessionalspeeddatingcompanyviaposters

andonlineadvertsannouncinglocalspeeddatingeventsforsinglemenandwomen

aged20to40.Participantswerethenassignedtooneoftwoagegroupstolimitage

differencesbetweendatingpartners,asistypicalinspeeddating.Agesrangedfrom20

to33intheyoungergroup(mean22±1.7inwomen,28.3±1.0inmen)and25to40in

theoldergroup(mean29±4.1inwomen,29.0±5.6inmen).Althoughmenwereon

average3.5yearsolderthanwomen(28.5vs24.9years),theagedifferencebetween

pairsinspeeddatesaveragedlessthanayear(0.07±6.9years)andwasnot

significantlydifferentfromzero(t213=-.137,p=.89).Beforeconfirmingtheir

participation,interestedindividualswereinformedthattheirvoiceswouldberecorded

duringtheeventforsubsequentacousticanalysisaspartofaresearchstudy.All

participantsprovidedsignedconsentbeforetakingpartinthestudy.

Questionnairescompletedaftertheeventconfirmedthatallparticipantsweresingle.

Thereportedamountoftimesinceapreviousrelationshiprangedfrom1to60months

anddidnotdiffersignificantlybetweenthesexes(mean19.3monthsamongwomen,

11.9monthsamongmen;t24=1.08,p=.29).Halfofthewomen,andone-quarterofthe

men,reportednotpreviouslyhavingparticipatedinaspeeddatingevent.Amongthose

whohad,participationratesrangedfrom1-2(women)and1-15(men).Meninour

samplethereforereportedgreaterexperiencewithspeeddatingthandidwomen(t26=-

2.34,p=.03,equalvariancesnotassumed).

Procedure

Weheldtwospeeddatingeventsco-organisedandco-directedbyanexperienced

speeddatinghost,heldinadedicatedroomatalocalcafé.Participantsarrived

individuallytothecaféwheretheyweregreetedbythehostandresearchers.Upon

arrival,thehostexplainedthespeeddatingprocedurestoparticipantsinsmallgroups,

andeachparticipantwasgivenanametagandbookletinwhichtomarktheirpersonal

matechoicepreferencesaftereachspeeddate(i.e.,‘yes’or‘no’).Afterproviding

informedconsent,participantswerefittedwithaportablevoicerecorderandheadset

andseatedatoneofseveraldesignatedtables.

Followingtypicalspeeddatingprocedures,eachdatingroundlasted6minutesafter

whichmenrotatedtoanadjacenttable.Duringthebriefinterludebetweendates,both

sexesindicatedtheirpreferencefortheirpreviousdateinapersonalizedbookletby

marking‘yes’or‘no’besidethatdate’sID.Thisprocesscontinueduntilallmenand

womenhaddatedoneanother.Followingthefinalround,participantsweregiven

additionaltimetoindicatetheirmatechoicepreferences.Asiscustomaryatspeed

datingevents,participantswhowere‘matched’(i.e.,whoindicatedamutualpreference

foroneanother)wereinformedbyemailwithin24hoursandgivenoneanother’s

contactinformation.Participantsalsocompletedashortquestionnaireaftertheevent,

inwhichtheyindicatedtheirage,sex,andwereaskedtoprovidedemographic

informationanddetailsregardingpreviousspeeddatingexperience.

Acousticrecordingandanalysis

Participants’voiceswererecordedthroughouttheentiredurationofthespeeddating

eventusingportableTascamDR-05recordersandlightweight(12g),discreetcardioid

condenserheadsetmicrophonesatasamplingrateof96kHzand24-bitamplitude

quantization.RecordingswerestoredontomicroSDHCmediacardsasuncompressed

WAVfilesandlatertransferredtoalaptopcomputerforeditingandanalysis.This

methodallowedustoobtainhighquality,directionalvoicerecordingsthatwould

otherwisebedifficulttoobtaininanoisyenvironmentusingastationarymicrophone.

AcousticeditingandanalysiswereperformedinPraatv.6.0.21[32].Fragmentsof

silence,acutenoise,nonverbalvocalizations(e.g.,laughter)andmulti-voicing(e.g.,the

voiceofthedatingpartner)werefirstmanuallyremovedfromaudiofiles.Recordings

werethensegmentedintomultiplepartseachcorrespondingtoagivenparticipantand

asinglespeeddate.Wefurtherspliteachsoundfileintothreeequaltimesegments

(beginning,middleandendofthedate;meansegmentduration50.6±23s),resultingin

atotalof726voiceclipsforacousticanalysis.

Weusedabatch-processingscripttomeasurefiveparametersoffundamental

frequency:mean(F0mean),range(F0minandF0max),variabilityandcontour,

includingstandarddeviation(F0sd)andthecoefficientofvariation(F0CV;givenbyF0

sd/F0mean[33]).AllF0parametersweremeasuredusingPraat’sautocorrelation

algorithmwithasearchrangeof60-600Hzandatimestepof0.01s.Spuriousoctave

jumpsweremanuallycorrected(see[34]).Perceptually,F0mean,minandmax

representtheaverage,lowerandupperrangesofaspeaker’svoicepitch,respectively,

withrelativelylowervaluessounding‘deeper’.Incontrast,F0sdrepresentsthe

absolutedegreeofvoicepitchvariabilityaroundthemeanacrossanutterance,andF0

CVadjuststhisvariationtothemagnitudeofF0therebycontrollingforthenonlinear

perceptionofvoicepitch.Thus,F0CVmorereliablyrepresentstheperceptualsalience

ofthisF0variability.

Preferencescores

Wecomputedthreetypesofpreferencescoreforeachparticipant.Desirabilityscores

representhow‘desired’theparticipantwasbyothers,andwerecomputedbydividing

thenumberofdateswhomarkedtheparticipantas‘yes’bythetotalnumberofdates,

givingtheproportionofdatingpartnerswhoindicatedapreferencefortheparticipant.

Choosinessscoresrepresenthow‘choosy’theparticipantwas,andwerecomputedby

dividingthenumberofdateswhomtheparticipantmarkedas‘yes’bythetotalnumber

ofdates.Finally,matchscoresindicatethenumberoftwo-waypreferences,thatis,the

numberofdatesonwhichbothparticipantsindicatedamutualpreferenceforone

another.

StatisticalAnalysis

Weusedlinearmixedmodels(LMMs)withmaximum-likelihoodestimationtotestfor

within-individualvariationinF0parameters(voicepitchmodulation).Weranseparate

modelsforeachsexduetonon-independenceinfemaleandmaledata(pairinginspeed

dating),andbecausevocalparametersandpreferencesdifferbetweenthesexes.

WeexaminedF0modulationasafunctionofthepersonalmatechoicepreferencesof

bothdatingpartners(i.e.,indicatingoneanotherasa‘yes’or‘no’),thespeaker’soverall

choosinessscore,andthedate’soveralldesirabilityscore.Preferencevariableswere

includedasfixedfactors(chosedate,theparticipantmarkedthatrespectivedateasa

‘yes’;chosenbydate,thedatemarkedtheparticipantasa‘yes’)orfixedcovariates

(choosinessscoreofspeaker;desirabilityscoreofdate).Participantidentitywasincluded

asarandomsubjectvariableinallmodels,andtheagedifferencebetweeneachman

andwomanoneachspeeddatewasincludedasarandomcovariate.Weadditionally

includedtimesegmentasafixedfactortoexaminewhethervoicechangesweremore

likelytooccuratthebeginning,middleorendofadate.Thesefactorswerefirst

examinedinafullyfactorialmodel.Therewerenomainorinteractioneffectsof

choosinessscoreofspeakeronvocalparametersforeithersex,thereforethisvariable

wasexcludedfromthefinalmodels.Timesegment,althoughnotsignificant,was

retainedinfinalmodelstoavoidpseudo-replication;itsinclusiondidnotaffectthe

patternofresults.Thefinalmodelcanbedescribedwiththeequation:

𝑦"# = 𝑏& + 𝑢&# +𝑏*𝑋"# +𝑏,𝑋"# + 𝑏-𝑋"# + 𝑏.𝑋"# + 𝑏/𝑋"# + 𝑢/# 𝑋"# + 𝜖"#

Note:b1–chosedate;b2–chosenbydate;b3–timesegment;b4–date’soveralldesirability;b5–agedifference

betweeneachmanandwomanoneachspeeddate;u0j–acomponenttotheinterceptmeasuringvariabilityin

intercepts;u5j–acomponenttotheslopeoftheoverallmodelthatmeasuresthevariabilityinslopes

Followingthis,weexaminedsignificantmaineffectsofcategoricalvariablesintheLMM

usingpairwisetestswithŠidákcorrectionformultiplecomparisons,andexamined

significantmaineffectsof,orinteractionswith,continuousvariables(i.e.,date’s

desirability)usinglinearregressiontoillustratethedirectionandstrengthofthese

relationships.Forinteractionsbetweencontinuouscovariatesandfixedfactors,we

averagedvocalparameterswithineachrelevantfixedcategory(e.g.,chosedate,didnot

choosedate,)andplottedseparatelinesofbestfitforeachfixedfactor.

Sexdifferences,effectsofparticipantage,oreffectsoftheagedifferencebetween

datingpairsonchoosiness,desirabilityandmatchscoresoronpersonalmatechoice

preferencesweretestedusingone-wayanalysesofvarianceorlinearregression.All

testsweretwo-tailedatanalphalevelof.05.

RESULTS

Desirability,choosiness,andmatches

Wefoundthatwomenwereonaveragesignificantlychoosierthanmen(F1,29=6.74,

p=.01):whilewomenindicatedapreferencefor30%(range0-67%)oftheirdates,men

indicatedapreferencefor51%(range13-88%).Inturn,women’sdesirabilityscores

weresignificantlyhigherthanmen’s(F1,29=6.72,p=.01),averaging52%inwomen

(range0-91%)comparedto28%inmen(range0-67%).Bothpartnersindicateda

mutualpersonalpreferenceforoneanotheron14%ofdates;individualsuccessrates

formatchesrangedfrom0%(nomatch)to33%(women)and38%(men),anddidnot

differsignificantlybetweenthesexes(F1,29=0.61,p=.81).Desirabilityandchoosiness

scorescorrelatednegativelyinbothsexes,howeverthisrelationshiponlyreached

statisticalsignificanceamongwomen(r=-.60,p=.03;men:r=-.24,p=.35),

indicatingthatwomenwhoweremoredesirablewerealsochoosier.

Participantagedidnotsignificantlypredictdesirabilityscores(men:F1,16=1.7,p=.21;

women:F1,11=3.8,p=.08),choosinessscores(men:F1,16=.05,p=.82;women:F1,11=4.4,

p=.06),ormatchscores(men:F1,16=.38,p=.55;women:F1,11=.004,p=.95)ineithersex,

thougholderwomenshowednonsignificanttrendstowardlowerdesirabilityscores,

andhigherchoosinessscores,comparedtoyoungerwomen.

Voicepitchmodulation

Linearmixedmodels(LMMs)testedforwithin-individualchangesinvoicepitch

parametersacrossspeeddates(seeStatisticalAnalysis).Table1reportssignificant

effectsofmodelsconductedseparatelyforeachsex.

Table1.Within-individualF0modulation.LinearMixedModelsexaminingrelationships

amongspeeddatingpreferences,date’soveralldesirability,andwithin-individualmodulation

ofvoicepitchparametersinwomen’sandmen’svoices,controllingfortheagedifference

betweendatingpartners.Onlysignificant(p<.05)effectsarereportedhere,forfullmodel

outputsandestimatedmarginalmeansseeTablesS1andS2.Modelsyntaxisprovidedin

supplementalmaterials.SeealsoFigure1.

Voice

parameterModelsource df1,df2 F p

Women

F0mean Intercept 1,321 9411.2 <.001

Chosedate 1,321 18.7 <.001

Date’sdesirability 1,321 19.1 <.001

Chosedate*Date’sdesirability 1,321 9.8 .002

F0min Intercept 1,321 595.2 <.001

Chosenbydate 1,321 4.6 .034

F0max Intercept 1,321 6237.0 <.001

Chosedate 1,321 4.0 .047

F0sd Intercept 1,321 1571.9 <.001

Chosedate 1,321 16.4 <.001

Date’sdesirability 1,321 14.8 <.001

Chosedate*Date’sdesirability 1,321 6.9 .009

F0CV Intercept 1,321 2013.1 <.001

Chosedate 1,321 7.0 .009

Date’sdesirability 1,321 5.5 .02

Chosedate*Chosenbydate*

Date’sdesirability

1,321 3.9 .048

Men

F0mean Intercept 1,321 1089.6 <.001Chosedate

1,321 4.6 .033

Chosedate*Chosenbydate*

Date’sdesirability

1,321 4.4 .036

F0min Intercept 1,321 633.0 <.001

Date’sdesirability 1,321 6.6 .01F0max Intercept 1,321 340.2 <.001F0sd Intercept 1,321 173.2 <.001F0CV Intercept 1,321 257.9 <.001

Womenmodulatedtheiraveragevoicepitch(F0mean),maximumvoicepitch(F0

max),andpitchvariability(F0sdandCV)asafunctionoftheirpersonalmatechoice

preferencesacrossspeeddates(Fig1a).PairwisetestswithŠidákcorrectionshowed

that,overall,womenspokewithasignificantlyhigherF0mean(p<.001,df=321,95%

CIfordifference=6.5to19.3)F0max(p=.01,df=321,95%CI=4.2to36.5),andF0

variability(F0sd,p<.001,df=321,95%CI=4.0to11.5;F0CV,p=.006,df=321,95%

CI=.005to.029)towardmentheychoseaspotentialmates(markedas“yes”)than

towardmentheydidnotchoose(markedas“no”).Theaveragemagnitudeofwomen’s

meanvoicepitchmodulation(13Hz)exceedsperceptualdiscriminationthresholdsby

almostthree-fold[9].

However,thesemaineffectswerequalifiedbyinteractionsbetweenwomen’spersonal

matechoicepreferencesandmen’soveralldesirabilityscores.AsillustratedinFigure

1b,womenraisedtheF0meanandabsolutevariability(F0sd)oftheirvoicesondates

withmentheypersonallypreferred,butonlytowardmenwithdesirabilityscores

below0.50(i.e.,menchosenbyfewerthanhalfoftheirdatingpartners),forwhom

womenraisedtheirpitchbymorethan20Hzonaverage.Incontrast,women

marginallyloweredthesepitchparameterstowardmentheypreferredandwhose

desirabilityscoreswereamongthehighest.Ondateswithmentheypreferred,14%

and12%ofthemodulationinwomen’sF0meanandF0sdwasexplainedbytheman’s

desirabilityscore,respectively,whereasthedesirabilityofnon-preferredmendidnot

predictwomen’svoicepitchmodulation(Fig1b).Whilewomenspokewithaless

monotonevoice(higherF0CV)towardmentheypreferred,themagnitudeofthisvoice

changealsodecreasedasmen’soveralldesirabilityincreased,andwasabsentondates

inwhichneitherthewomannorthemanshowedapersonalpreferenceforone

another.

Variationinwomen’sminimumpitchwasnotqualifiedbythedate’soveralldesirability

norwomen’sownpreferences,butratherwaspredictedsolelybymen’spreferences.

WomenspokewithalowerF0minondatesinwhichtheirdatingpartnerchosethem,

loweringtheirF0minbyanaverageof15.8Hzcomparedtodatesonwhichtheywere

notchosen(p=.007,df=321.95%CI=-27.3to-4.3;Fig1a).

Menalsomodulatedtheirvoicesacrossspeeddates.AlthoughtheLMMindicatedthat

menspokewithalowerF0meantowardwomentheychose(markedas“yes”)than

towardthosetheydidnotchoose(Table1),pairwisetestsshowedthatthismaineffect

didnotreachsignificancefollowingŠidákcorrection(p=.26,df=321,95%CIfor

difference=-11.0to-2.9;Fig1c).Indeed,likewomen,thiseffectwasqualifiedbya

significantinteraction,indicatingthatmenloweredtheirF0meanmoretowardwomen

withlowthanhighdesirabilityscores,andonlyondateswithwomentheychoseas

potentialmates,particularlyifthosewomenchosetheminreturn(Fig1d).Indeed,men

loweredtheirF0meanbymorethan20Hz,orapproximatelyfourtimesthejust-

noticeabledifference,toward‘matched’womenwhosedesirabilityscoreswereamong

thelowest.Likefemaleparticipants,thedesirabilityofnon-preferredwomendidnot

predictmen’smeanvoicepitchmodulation(Fig1d).Linearregressionfurthershowed

thatmenloweredtheirF0minwhenspeakingtowomenwithrelativelyhigheroverall

desirabilityscores,regardlessofpersonalpreference(Fig1c).

InadditiontocontrollingforagedifferenceinLMMs,weranadditionalanalysesof

variancethatconfirmedthattheagedifferencebetweensexesineachspeeddatingpair

didnotsignificantlypredictthematechoicesofeithersexineitheragegroup(younger

women:F1,76=0.02,p=.90,olderwomen:F1,29=1.18,p=.29;youngermen:F1,76=3.74,

p=.06,oldermen:F1,29=0.51,p=.48).Moreover,ourLMMshowednosignificanteffect

oftimesegmentonvoicepitchmodulation(TableS1),indicatingthatpitchmodulation

emergedearlyinthespeeddateandpersistedatasimilarmagnitudethroughout.

Figure1.Women’sandmen’sF0modulationacrossspeeddates.Panels(a)and(c)showmaineffectsofpersonalmatechoicepreferencesonvoicepitchmodulation,whereaspanels(b)and(d)showinteractionsbetweenpersonalpreferencesanddate’soveralldesirability.ColumnsinbargraphsrepresentestimatedmarginalmeansfromLMMs(seeTables1andS2),whereerrorbarsrepresentstandarderrorsofthemeans.Markersinscatterplotsrepresentindividualspeeddates;markersfordistinctcategoriesareminimallyoffsetalongthex-axistoavoidoverplottingandtoimprovevisualisation.Asterisks’indicatestatisticalsignificanceofpairwisecomparisonsfollowingŠidákcorrection,where***p<.001,**p<.01,*p<.05,andnsp>.05.Correlationcoefficients(r)aregivenbesideeachlinearregressionline.Seeembeddedlegendforadditionaldetails.

Relationshipsbetweenvoicepitchanddesirabilityorchoosiness

Inadditiontoexaminingwithin-individualfluctuationsinvoicepitch(thatis,how

participantsmodulatedtheirvoicesfromdate-to-date),wetestedwhetherbetween-

individualdifferencesinpitchpredicteddesirabilityandchoosinessscores.Average

pitchparameterswerecomputedforeachparticipantbyaveragingF0parameters

acrossalldates,within-individuals,andregressedagainsttheiroveralldesirabilityand

choosinessscores.

Wefoundthatwomen’sF0minexplained43.7%to55%ofthevarianceintheir

desirabilityscores(r=-.66,p=.014,n=13),indicatingthatwomenwithlowerpitch

minimaweremoredesiredbymen.Thestrengthofthisrelationshipincreasedafter

controllingforwomen’schoosinessscores(rp=-.74,p=.006,df=10).Although

relationshipsbetweenwomen’sdesirabilityandtheirF0mean(rp=-.45)andF0CV(rp

=.37)weremoderateinstrength,noothervoicepitchparameterexplaineda

significantamountofvarianceinthedesirabilityorchoosinessscoresofeithersex

(TableS3).

DISCUSSION

Theresultsofthisstudysupportthehypothesisthatwomenandmenmodulatetheir

voicesinreal-lifematechoicecontextsbasedonpersonalmatechoicepreferences,and

thematequalityofapotentialpartner.Womenspokeinahigher-pitchedandless

monotonevoiceonspeeddateswithmentheychoseaspotentialmates,howeveronly

ifthosemenalsohadarelativelyloweroveralldesirabilityscore(i.e.,werepreferred

byfewerthan50%ofotherwomen).Menloweredtheirmeanpitchondateswith

womentheypersonallypreferred,butheretoo,themagnitudeofmen’svoicepitch

modulationdecreasedastheirdates’overalldesirabilityincreased.

Previousstudiesexaminingvoicemodulationinmockdatingcontextshaveproduced

conflictingresults[13,28–31],possiblybecausedatingpartnerswerehypothetical,and

theirdesirabilitywasbasedsolelyonpre-ratedattractivenessofphotosorvideos

ratherthanonparticipants’personalpreferences.Here,weshowthatpersonalmate

choicepreferencescandeviatefrom,interactorevenconflictwithgroup-based

desirabilityscores.Indeed,althoughwomengenerallyspokewithahigherpitch(inline

with[30]),andmenwithalowerpitch(inlinewith[13,31]),towarddatingpartners

theymarkedas‘yes’,thiswasdrivenlargelybydatingpartnerswithlowerdesirability

scores,towardwhombothsexesalteredtheirvoicepitchbymorethan20Hz(1.5

semitonesinwomen,2.7inmen)onaverage.Infact,womenspokewithalowervoice

pitchtowardmenwhowerebothhighlydesiredbyotherwomenandwhomthey

personallypreferred.Thisfindingsupportsatleasttwootherstudies[13,14]that

reportthatwomenlower,ratherthanraise,theirvoicepitchinamatingcontext,and

thushighlightstheneedtoconsiderbothgroup-levelandindividualmatepreferences

infuturework.

Alargebodyofresearchindicatesthatmenwithrelativelylow-pitchedvoicesare

preferredbywomenasmatesandhavehighmatevalue[5,7].Thus,theobservation

thatmenloweredtheirvoicepitchinresponsetowomentheypreferred,particularlyif

thosewomenalsopreferredthem,suggeststhatvoicemodulationinmenmayfunction

toincreasetheirreproductivesuccess.Incontrast,ourresultscontradicttheprediction

thatwomenubiquitouslyfeminizetheirvoicestowardpreferredpotentialmates,as

thiswasnotthecaseondateswithhighlydesirablemen.Moreover,inoursampleof

speeddaters,womenspokewithlowerpitchminimatowardmenwhopreferredthem,

andwhencomparingacrosswomen,menpreferredwomenwhospokewithalower

minimumpitch,suchthatwomen’sminimumpitchexplainedupto55%ofthevariance

inhowdesiredtheywerebymen.

Thisapparentconflictbetweenthefunctionalrelevanceofhighversuslowvoicepitch

inwomenmayberesolvedbyconsideringthatindexicalcuestostaticspeakertraits

(e.g.,age,sex)mayfunctiondifferentlythanmoredynamic,socialorsexualcues.Thus,

whilearelativelyhighvoicepitchinwomencansignalyouth,femininityand

reproductivefecundity[7],bydynamicallyloweringhervoicepitchawomanmightbe

signallingsexualinterestandintimacytoaman[13–15].Alternatively,or

simultaneously,shemightlowerherpitchtocommunicatesocialdominanceora

confidentandmaturepersona,aspeoplewithlow-pitchedvoicesareoftenattributed

traitssuchascompetence,trustworthinessandleadership[5,17].Indeed,recent

studiessuggestthatyoungwomenareincreasinglyusingaverylowpitchregister,

resultinginvocalfryora‘creaky’quality,inprofessionalworkcontexts[35,36].

Dynamicvoicemodulationmaybeespeciallyfunctionalwhenmultiplemodalitiesare

availabletothereceiver,astheywereinourstudyandtypicallyareinsocial

interactions.Inamatingcontext,forinstance,amanmaygaugeawoman’sage,

femininity,andfecundityfromvisualandolfactorycues[37],whilesimultaneously

interpretingsocialandsexualinformationfromthedynamicpropertiesofthewoman’s

voice.Whileitispossiblethatawoman’sagemayinfluencethedirectionanddegreeto

whichshemodulateshervoicepitchtowardpotentialmates,ourresultsshowedno

effectofage,oragedifferencesbetweendatingpartners,onvoicemodulation.Agealso

didnotpredictchoosiness,desirabilityorsuccessfuldatingmatches,thoughitmustbe

notedthatthesebetween-subjectcomparisonswereunderpowered.

Inlinewithevolutionarymodelsthatimplicatewomenasthe‘choosier’sex[38],both

choosinessanddesirabilityscoreswerehigheramongwomenthanmen,withwomen

showingapersonalpreferenceforonlyone-thirdoftheirdates(comparedtoone-half

inmen).Womenalsomodulatedmorevocalparametersinresponsetopreferred

datingpartnersthandidmen,includingtheirmaximumpitch,andmostnotably,

variabilityintheirpitch,whichisknowntocommunicatemasculinityandphysical

formidabilityamongmen[6].Thissexdifferencecouldbetiedtoabilityoreffort.

Womenhavebeenshowntomoreeffectivelymodulatetheirvocalattractiveness‘on

demand’thanmen[14].Alternatively,forwomen,datesmarkedas‘yes’werelikelyto

includemenwhomwomenweremaximallyattractedtoandmayhavethusexerted

maximalvocalefforttoward,whereasmen’slongerlistofpotentialmateslikely

includedmarginallypreferredpartnerstowardwhommenmighthavedisplayedless

vocaleffort.

Studyingvoicemodulationinreal-lifedatingoffershighecologicalvalidity,yetthelack

ofexperimentalmanipulationdoesnotallowforcausalinferencesregardingtheroleof

variousvocalparameters,ordifferentmodalities,onmatechoicepreferences.Thismay

beinvestigatedinfutureworkusingresynthesizedspeechasuni-modalandwithin

multi-modalstimuli.Givenourresults,experimentsarenowalsoclearlyneededto

gaugetheroleofownversusgroup-levelmatepreferencesonvoicemodulation.Our

smallsamplesize,whileadequateforcapturingdynamicwithin-individualmodulation

acrossspeeddates,limitstheextenttowhichinferencescanbemadeaboutbetween-

individualdifferencesinvocalparametersandbehaviour.Individualdifferencefactors,

suchastheinfluenceofpastdatingexperience,shouldthusbeinvestigatedin

replicationstudiesutilizinglargersamples.Finally,futurestudiesmayexaminethe

influenceofstress(e.g.,duetoafirstdateortherealizationofbeingaudio-recorded)

andotheremotionsonvoicemodulationinadatingcontext.

Althoughthereissomerecentevidenceforbehavioralandcontextualflexibilityinthe

vocalizationsofothermammalsandgreatapes[22],thecapacityforvolitionalvocal

controlinhumansisunprecedentedinitscomplexityandthus,initspotentialbreadth

offunctionality.Indeed,thisstudyshowsthat,whileintegraltohumanspeech

production,voicemodulationalsoaffectsthenonverbaldimensionsofvocal

communicationduringmatechoice.Thecapacityforwomenandmentodynamically

altertheirvoicepitchthereforehasthepotentialtoaffectreproductivesuccess,but

beyondthis,mayfunctiontomanipulatetheperceptionsandbehavioursofothersina

widerangeofsocial,economicandpoliticalcontexts[22].

Funding

ThisworkwassupportedbytheEuropeanCommissionthroughaMarieSkłodowska-

CurieindividualfellowshiptoK.P.(H2020-MSCA-IF-2014-655859).

Authors’contributions

K.P.andD.R.designedtheinvestigation.K.P.,M.G.andA.O.collectedthedata.J.P

preparedaudiofilesandquestionnairedataforanalysis.K.Pperformedacoustic

analysis.K.P.,andD.R.performedstatisticalanalyses.K.P.wrotethemanuscriptand

createdthefigures.Themanuscriptwasreviewed,editedandapprovedbyallauthors,

whoagreetobeaccountableforthework.

Competinginterests

Theauthorsreportnocompetinginterests.

Ethics

ThestudywasreviewedandapprovedbytheSciencesandTechnologyCross-Schools

ResearchEthicsCommittee(C-REC)oftheUniversityofSussex(ER-REBY-3,ER-KP292-

3/4).

Dataavailability

Thedatasetsupportingthisarticlewillbeuploadedtofigshare.

FIGURELEGEND

Figure1.Women’sandmen’sF0modulationacrossspeeddates.Panels(a)and(c)

showmaineffectsofpersonalmatechoicepreferencesonvoicepitchmodulation,

whereaspanels(b)and(d)showinteractionsbetweenpersonalpreferencesanddate’s

overalldesirability.Columnsinbargraphsrepresentestimatedmarginalmeansfrom

LMMs(seeTables1andS2),whereerrorbarsrepresentstandarderrorsofthemeans.

Markersinscatterplotsrepresentindividualspeeddates;markersfordistinct

categoriesareminimallyoffsetalongthex-axistoavoidoverplottingandtoimprove

visualisation.Asterisks’indicatestatisticalsignificanceofpairwisecomparisons

followingŠidákcorrection,where***p<.001,**p<.01,*p<.05,andnsp>.05.Correlation

coefficients(r)aregivenbesideeachlinearregressionline.Seeembeddedlegendfor

additionaldetails.

REFERENCES

1. Puts D. 2016 Sexual selection on male vocal fundamental frequency in humans and other anthropoids. Proc. R. Soc. B Biol. Sci.

2. Hollien H, Green R, Massey K. 1994 Longitudinal research on adolescent voice change in males. J. Acoust. Soc. Am. 96, 2646–2654.

3. Dabbs JM, Mallinger A. 1999 High testosterone levels predict low voice pitch among men. Personal. Individ. Differ. 27, 801–804. (doi:10.1016/S0191-8869(98)00272-4)

4. Cartei V, Bond R, Reby D. 2014 What makes a voice masculine: Physiological and acoustical correlates of women’s ratings of men’s vocal masculinity. Horm. Behav. 66, 569–576. (doi:10.1016/j.yhbeh.2014.08.006)

5. Pisanski K, Bryant GA. 2018 The Evolution of Voice Perception. In The oxford handbook of voice studies (eds NS Eidsheim, KL Meizel), New York, NY, USA: Oxford University Press.

6. Puts DA, Apicella CL, Cárdenas RA. 2012 Masculine voices signal men’s threat potential in forager and industrial societies. Proc. R. Soc. Lond. B Biol. Sci. 279, 601–609. (doi:10.1098/rspb.2011.0829)

7. Pisanski K, Feinberg DR. 2018 Voice attractiveness. In Oxford Handbook of Voice Perception (ed P Belin), New York, NY, USA: Oxford University Press.

8. Abitbol J, Abitbol P, Abitbol B. 1999 Sex hormones and the female voice. J. Voice 13, 424–446. (doi:10.1016/S0892-1997(99)80048-4)

9. Pisanski K, Rendall D. 2011 The prioritization of voice fundamental frequency or formants in listeners’ assessments of speaker size, masculinity, and attractiveness. J. Acoust. Soc. Am. 129, 2201. (doi:10.1121/1.3552866)

10. Borkowska B, Pawlowski B. 2011 Female voice frequency in the context of dominance and attractiveness perception. Anim. Behav. 82, 55–59. (doi:10.1016/j.anbehav.2011.03.024)

11. Feinberg DR, DeBruine LM, Jones BC, Perrett DI. 2008 The role of femininity and averageness of voice pitch in aesthetic judgments of women’s voices. Perception 37, 615.

12. Fraccaro PJ, O’Connor JJ, Re DE, Jones BC, DeBruine LM, Feinberg DR. 2013 Faking it: deliberately altered voice pitch and vocal attractiveness. Anim. Behav. 85, 127–136.

13. Hughes SM, Farley SD, Rhodes BC. 2010 Vocal and physiological changes in response to the physical attractiveness of conversational partners. J. Nonverbal Behav. 34, 155–167.

14. Hughes SM, Mogilski JK, Harrison MA. 2014 The perception and parameters of intentional voice manipulation. J. Nonverbal Behav. 38, 107–127. (doi:10.1007/s10919-013-0163-z)

15. Tuomi SK, Fisher JE. 1979 Characteristics of simulated sexy voice. Folia Phoniat 31, 242–249.

16. Babel M, McGuire G, King J. 2014 Towards a more nuanced view of vocal attractiveness. PLoS ONE 9, e88616. (doi:10.1371/journal.pone.0088616)

17. Klofstad CA, Anderson RC, Nowicki S. 2015 Perceptions of competence, strength, and age influence voters to select leaders with lower-pitched voices. PLoS ONE 10, e0133779. (doi:10.1371/journal.pone.0133779)

18. Jones BC, Feinberg DR, DeBruine LM, Little AC, Vukovic J. 2010 A domain-specific opposite-sex bias in human preferences for manipulated voice pitch. Anim. Behav. 79, 57–62. (doi:10.1016/j.anbehav.2009.10.003)

19. Morton ES. 1977 On the occurrence and significance of motivation-structural rules in some bird and mammal sounds. Am. Nat. 111, 855–869.

20. Taylor AM, Charlton BD, Reby D. 2016 Vocal production by terrestrial mammals: source, filter, and function. In Vertebrate Sound Production and Acoustic Communication, pp. 229–259. Springer.

21. Belyk M, Pfordresher PQ, Liotti M, Brown S. 2015 The neural basis of vocal pitch Imitation in humans. J. Cogn. Neurosci. , 1–15. (doi:10.1162/jocn_a_00914)

22. Pisanski K, Cartei V, McGettigan C, Raine J, Reby D. 2016 Voice modulation: A window into the origins of human vocal control? Trends Cogn. Sci. 20, 304–318.

23. Titze IR. 1994 Principles of vocal production. Englewood Cliffs: Prentice-Hall.

24. Hollien H. 2014 Vocal fold dynamics for frequency change. J. Voice 28, 395–405. (doi:10.1016/j.jvoice.2013.12.005)

25. Kreiman J, Sidtis D. 2011 Foundations of voice studies: An interdisciplinary approach to voice production and perception. Wiley-Blackwell. See http://eu.wiley.com/WileyCDA/WileyTitle/productCd-0631222979.html.

26. Cartei V, Cowles HW, Reby D. 2012 Spontaneous voice gender imitation abilities in adult speakers. PLoS ONE 7, e31353. (doi:10.1371/journal.pone.0031353)

27. Pisanski K, Mora E, Pisanski A, Reby D, Sorokowski P, Franckowiak T, Feinberg D. 2016 Volitional exaggeration of body size through fundamental and formant frequency modulation in humans. Sci. Rep. 6, 34389. (doi:10.1038/srep34389)

28. Anolli L, Ciceri R. 2002 Analysis of the vocal profiles of male seduction: From exhibition to self-disclosure. J. Gen. Psychol. 129, 149–169.

29. Hodges-Simeon CR, Gaulin SJC, Puts DA. 2010 Different vocal parameters predict perceptions of dominance and attractiveness. Hum. Nat. 21, 406–427. (doi:10.1007/s12110-010-9101-5)

30. Fraccaro PJ, Jones BC, Vukovic J, Smith FG, Watkins CD, Feinberg DR, Little AC, Debruine LM. 2011 Experimental evidence that women speak in a higher voice pitch to men they find attractive. J. Evol. Psychol. 9, 57–67. (doi:10.1556/JEP.9.2011.33.1)

31. Leongómez JD, Binter J, Kubicová L, Stolařová P, Klapilová K, Havlíček J, Roberts SC. 2014 Vocal modulation during courtship increases proceptivity even in naive listeners. Evol. Hum. Behav. 35, 489–496. (doi:10.1016/j.evolhumbehav.2014.06.008)

32. Boersma P, Weenink D. 2016 Praat: Doing phonetics by computer v 6.0.21. See www.praat.org.

33. Eguchi S, Hirsh IJ. 1969 Development of speech sounds in children. Acta Oto-Laryngol. Suppl. 257, 1–51.

34. Reby D, Levréro F, Gustafsson E, Mathevon N. 2016 Sex stereotypes influence adults’ perception of babies’ cries. BMC Psychol. 4. (doi:10.1186/s40359-016-0123-6)

35. Yuasa IP. 2010 Creaky voice: a new feminine voice quality for young urbanoriented upwardly mobile American women? Am. Speech 85, 315–337. (doi:10.1215/00031283-2010-018)

36. Wolk L, Abdelli-Beruh NB, Slavin D. 2012 Habitual use of vocal fry in young adult female speakers. J. Voice 26, e111–e116.

37. Groyecka A, Pisanski K, Sorokowska A, Havlíček J, Karwowski M, Puts D, Roberts SC, Sorokowski P. 2017 Attractiveness is multimodal: Beauty is also in the Nose and Ear of the Beholder. Front. Psychol. 8.

38. Bleu J, Bessa-Gomes C, Laloi D. 2012 Evolution of female choosiness and mating frequency: effects of mating cost, density and sex ratio. Anim. Behav. 83, 131–136.