Full Terms & Conditions of access and use can be found athttp://www.tandfonline.com/action/journalInformation?journalCode=pqje20

Download by: [University of La Laguna Vicerrectorado] Date: 19 November 2015, At: 09:16

The Quarterly Journal of Experimental Psychology

ISSN: 1747-0218 (Print) 1747-0226 (Online) Journal homepage: http://www.tandfonline.com/loi/pqje20

Negation in context: Evidence from the visualworld paradigm

Isabel Orenes, Linda Moxey, Christoph Scheepers & Carlos Santamaría

To cite this article: Isabel Orenes, Linda Moxey, Christoph Scheepers & Carlos Santamaría(2015): Negation in context: Evidence from the visual world paradigm, The Quarterly Journal ofExperimental Psychology, DOI: 10.1080/17470218.2015.1063675

To link to this article: http://dx.doi.org/10.1080/17470218.2015.1063675

Accepted author version posted online: 24Jul 2015.Published online: 28 Jul 2015.

Submit your article to this journal

Article views: 43

View related articles

View Crossmark data

Negation in context: Evidence from the visual worldparadigm

Isabel Orenes1, Linda Moxey2, Christoph Scheepers2, and Carlos Santamaría1

1Departamento de Psicología Cognitiva, Universidad de la Laguna, Tenerife, Spain2Department of Psychology, University of Glasgow, Glasgow, UK

(Received 23 April 2014; accepted 2 June 2015)

Literature assumes that negation is more difficult to understand than affirmation, but this might dependon the pragmatic context. The goal of this paper is to show that pragmatic knowledge modulates theunfolding processing of negation due to the previous activation of the negated situation. To test this,we used the visual world paradigm. In this task, we presented affirmative (e.g., her dad was rich) andnegative sentences (e.g., her dad was not poor) while viewing two images of the affirmed and denied enti-ties. The critical sentence in each item was preceded by one of three types of contexts: an inconsistentcontext (e.g., She supposed that her dad had little savings) that activates the negated situation (a poorman), a consistent context (e.g., She supposed that her dad had enough savings) that activates the actualsituation (a rich man), or a neutral context (e.g., her dad lived on the other side of town) that activatesneither of the two models previously suggested. The results corroborated our hypothesis. Pragmaticsis implicated in the unfolding processing of negation. We found an increase in fixations on thetarget compared to the baseline for negative sentences at 800 ms in the neutral context, 600 ms inthe inconsistent context, and 1450 ms in the consistent context. Thus, when the negated situationhas been previously introduced via an inconsistent context, negation is facilitated.

Keywords: Negation; Context; Visual world paradigm.

The comprehension of language is context depen-dent: Its content depends not only on the syntacticand semantic properties of the types of expressionsused, but also on facts about the situation in whichthe expressions are used (Stalnaker, 1998).Negative sentences do not point to the actual situ-ation but to what is not the case. As a consequence,in their comprehension we must figure out what thespeaker is referring to, and pragmatics becomes

crucial. If someone tells us that their car is notred there should have been a reason to presume(or speculate) that it is red. Otherwise the uttererwould have referred to the actual colour of thecar. In fact, individuals seem to use affirmation tocommunicate as default, but negation is oftenused in some specific contexts, and outside ofthese contexts negation is difficult to understand(Khemlani, Orenes, & Johnson-Laird, 2012).

Correspondence should be addressed to Isabel Orenes, Universidad de La Laguna, Departamento de Psicología Cognitiva,

Campus Guajara, sn. 38205 La Laguna, Tenerife, Spain. E-mail: iorenes@ull.es

We would like to thank the Language group of the University of Glasgow and David Beltrán for their suggestions and comments.

This research was supported by the Spanish Government [grant number PSI2011-26386], [grant number Consolider-Ingenio

2010 CSD2008-00048].

© 2015 The Experimental Psychology Society 1

THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2015

http://dx.doi.org/10.1080/17470218.2015.1063675

Dow

nloa

ded

by [

Uni

vers

ity o

f L

a L

agun

a V

icer

rect

orad

o] a

t 09:

16 1

9 N

ovem

ber

2015

The representation and processing of negation haveimportant implications for memory and languagetheories. Research has shown that negative sen-tences are more difficult to comprehend thantheir corresponding affirmatives: People takelonger to understand negation and are likely tomake more mistakes in processing negation (seeCarpenter & Just, 1975; Clark & Chase, 1972;Mayo, Schul, & Burnstein, 2004; Trabasso,Rollins, & Shaughnessy, 1971; Wason &Johnson-Laird, 1972; Wason & Jones, 1963).Furthermore, negated concepts are less well recalledthan affirmative ones (Fillenbaum, 1966; Lea &Mulligan, 2002; MacDonald & Just, 1989).These results would indicate that negative andaffirmative information is processed and rep-resented differently.

Among other explanations, Kaup, Lüdtke, andZwaan (2006) proposed a two-step simulationhypothesis to understand negation. For a negativesuch as the door is not open, for example, individualswould first represent the negated situation—anopen door—and then the actual situation—aclosed door. For an equivalent affirmative (e.g.,the door is closed), the actual situation is directlyavailable: a closed door (Kaup, Zwaan, & Lüdke,2007; Zwaan, Stanfield, & Yaxley, 2002). Thistwo-step processing has been found in metaphors(Hasson & Glucksberg, 2006), quantifiers(Moxey & Sanford, 1987), and counterfactuals(Byrne, 2005), among others. Within this frame-work, it has also been proposed that rather than atwo-step sequence of static images, negationwould be represented by the continuous-trajectoryderivative over time in a perceptual simulation(Anderson, Huette, Matlock, & Spivey, 2010;Huette & Anderson, 2012; Huette, Anderson,Matlock, & Spivey, 2011; Tian, Breheny, &Ferguson, 2010).

Despite this, many studies have also shownthat when negation is presented in an adequatecontext, it can be similar to affirmation (Dale &Duran, 2011; Givon, 1978; Glenberg,Robertson, Jansen, & Johnson-Glenberg, 1999;Johnson-Laird & Tridgell, 1972; Nieuwland &Kuperberg, 2008; Strawson, 1952; Wason,1972). The question that remains unsolved to

date is what makes a context “adequate” forunderstanding negation. Wason (1965) proposedthat negation is easier to understand when it isused to communicate deviations from expectations(see also Givon, 1978). In this case, negation maybe easier to understand because the negated situ-ation has been constructed already (Khemlaniet al., 2012). For example, Lüdtke and Kaup(2006) presented a context (e.g., that a boy’sT-shirt is dirty after he played outside in the back-yard) that activates the negated situation (e.g., adirty T-shirt) previously, and the processing ofnegation (e.g., the T-shirt was not dirty) was facili-tated. It seems then, that the difficulty of negationis due to the change of attention (or deviation)from the negated concept toward the actualsituation. However, when the negated concept ispreviously activated in an adequate context, itsprocessing is facilitated.

The visual world paradigm allows us to study theunfolding process of comprehending negation. In atypical visual world task, verbal and visual inputs arepresented simultaneously while the participants’ eyemovements are recorded. Eye movements providean index of real-time processing that most othermethodologies do not offer, and they are sensitiveto subtle aspects of language, attention, andmemory (Allopenna, Magnuson, & Tanenhaus,1998; Duñabeitia, Avilés, Afonso, Scheepers, &Carreiras, 2009; Rayner, 1998). From the litera-ture, it is expected that when something is heard,it is processed and attended to automatically. Atthe same time, if the object is visible, the eyesbegin to move towards the corresponding object(Cooper, 1974; Tanenhaus, Spivey-Knowlton,Eberhard, & Sedivy, 1995). It follows that whenpeople are exposed to negative assertions in avisual world paradigm, they will look more fre-quently at the most active information in workingmemory.

Our main goal in the present study is to test therole of context in the process of understandingnegation. Our hypothesis is that certain contextscan activate the negated concept before the nega-tion is encountered. This is consistent withKhemlani et al. (2012) who, based on Wason(1965), pointed out that negation is easier when

2 THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2015

ORENES ET AL.

Dow

nloa

ded

by [

Uni

vers

ity o

f L

a L

agun

a V

icer

rect

orad

o] a

t 09:

16 1

9 N

ovem

ber

2015

the negated situation has been constructed pre-viously within the context. First, we presented affir-mative (e.g., her dad was rich) and negativesentences (e.g., her dad was not poor) in a neutralcontext (e.g., her dad lived on the other side oftown), and we expected to find the classical effectof negation: Negative sentences take longer toprocess than affirmative sentences. This effectcould be reduced in a context in which thenegated situation has been constructed previously,what we have labelled an inconsistent context(e.g., She supposed that her dad had little savings).This context activates the negated situation (e.g.,her dad was poor). In a third condition, we pre-sented a consistent context sentence (e.g., She sup-posed that her dad had enough savings) that activatesthe actual situation (e.g., her dad was rich) followedby the affirmative (e.g., her dad was rich) or thenegative sentence (e.g., her dad was not poor).While the affirmative statement is directly under-stood after the context sentence, understanding ofthe negative statement requires listeners to rep-resent a rich man after hearing the context, andthen a poor man after hearing negation (notpoor), and finally the rich man again. These extrasteps required to understand negation will taketime compared to affirmation. Therefore, the con-sistent context is expected to be the most difficultone for negation (see Johnson-Laird & Tridgell,1972). The extra processing in the consistentcontext condition should be reflected in the eyemovement because people need to change theirfocus of attention twice from the rich man to thepoor man, and from the poor man to the richman in our example.

EXPERIMENTAL STUDY

Method

ParticipantsFifty native Spanish speakers from the University ofLa Laguna, Tenerife (Spain), participated in theexperiment in exchange for course credits. All ofthem had normal vision or wore soft contactlenses or glasses.

MaterialsForty-eight experimental vignettes of simple eventswere heard. The first two sentences in each vignettedescribed a general situation (e.g., Veronica needed anew car for work. She wondered whether her dad couldhelp her financially). The second sentence in eachvignette presented a context. This sentence wasalways stated as a presupposition by using reportedspeech forms (he supposed, she thought, etc.) so thatthere was no a priori contradiction with the sub-sequent affirmative or negative sentences. Thecontext could be consistent (e.g., She supposed thather dad had enough savings), inconsistent (e.g., Shesupposed that her dad had little savings), or neutral(e.g., her dad lived on the other side of town) withrespect to the target part of the story in the finalsentence that consisted of an affirmative (e.g., herdad was rich) or a negative sentence (e.g., her dadwas not poor).

For each trial, a different display with two imagesappeared on the screen—for example, an image of apoor man on the left and a rich man on the right (seeFigure 1). The position of each image was counter-balanced across conditions. The order of thematerials was randomized, and 12 versions of thefull set of materials were prepared: Each context(consistent, inconsistent, and neutral) was combinedwith the two types of sentences (affirmative andnegative) and the two poles of the adjective contra-ries (e.g., poor-rich), so that only one of the 12possible versions of each experimental item occurredin each presentation file (see Table 1). In total, therewere eight experimental trials in each of six con-ditions: consistent–affirmative, consistent–negative,inconsistent–affirmative, inconsistent–negative,neutral–affirmative, and neutral–negative. For alltrials, the analysis of fixations was time-locked tothe onset of the critical adjective in the target sen-tence—that is, “rich” in an affirmative sentencesuch as her dad was rich and “poor” in a negativesentence such as her dad was not poor.

Apparatus and procedureParticipants’ eye movements were recorded at a rateof 500 Hz using an SR Research EyeLink II head-mounted eye-tracker connected to a 21′′ colourCRT for visual stimulus presentation. Procedures

THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2015 3

NEGATION IN CONTEXT

Dow

nloa

ded

by [

Uni

vers

ity o

f L

a L

agun

a V

icer

rect

orad

o] a

t 09:

16 1

9 N

ovem

ber

2015

were implemented in SR Research ExperimentBuilder. Calibration and validation procedureswere carried out at the beginning of the experimentand were repeated several times per session. Trialsstarted with the presentation of a central fixationdot for drift correction while participants listenedto the general situation and the context (consistent,inconsistent, or neutral). After that, a display withtwo images appeared for 2 s, and then the story

finished with the presentation of an affirmative ornegative sentence that was the target. The trial con-cluded with the appearance of a written question(e.g., did Veronica need a new car?), which partici-pants had to answer by pressing either a “yes” or a“no” button. There was a practice block of fourtrials before the experiment proper started. Theentire experiment lasted approximately 30 min.

Results

The eye-movement data generated by the EyeLinksystem were analysed as follows. First, bitmap tem-plates were created for identifying regions of inter-est in each display (e.g., a rich and a poor man). Theobject regions were defined in terms of rectanglescontaining the relevant objects; fixations landingwithin the perimeters of these rectangles werecoded as fixations on the relevant objects. Theoutput of the eye-tracker included the x- andy-coordinates of participant fixations, which wereconverted into region of interest codes using thetemplates. Fixations shorter than 80 ms (account-ing for ∼0.2% of all fixations) were pooled withpreceding or following fixations if they werewithin 0.5° of visual angle. Times for blinks wereadded to the immediately preceding fixations.The analysed time period was from the onset ofthe target to 3000 ms after the target word (asestablished on a by-trial basis). This time windowwas chosen to guarantee that there was enoughtime to comprehend negations (e.g., Kaup et al.,2006; Lüdtke, Friedrich, De Filippis, & Kaup,2008). This period was further divided into 50-ms time slots. For each time slot, the number of fix-ations on the target image was divided by the sum

Table 1. Examples of experimental sentences

Veronica needed a new car for work. She wondered whether her dad

could help her financially.

Consistent–

affirmative

She supposed that her dad had little savings.

Her dad was poor. Version 1.

She supposed that her dad had enough

savings.

Her dad was rich. Version 2.

Consistent–

negative

She supposed that her dad had enough savings.

Her dad was not poor. Version 3.

She supposed that her dad had little savings.

Her dad was not rich. Version 4.

Inconsistent–

affirmative

She supposed that her dad had enough savings.

Her dad was poor. Version 5.

She supposed that her dad had little savings.

Her dad was rich. Version 6

Inconsistent–

negative

She supposed that her dad had little savings.

Her dad was not poor. Version 7.

She supposed that her dad had enough

savings.

Her dad was not rich. Version 8.

Neutral–

affirmative

Her dad lived on the other side of town.

Her dad was poor. Version 9.

Her dad lived on the other side of town.

Her dad was rich. Version 10.

Neutral–

negative

Her dad lived on the other side of town.

Her dad was not poor. Version 11.

Her dad lived on the other side of town.

Her dad was not rich. Version 12.

Figure 1. Example of a display with two images; one corresponds to a poor man and another to a rich man.

4 THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2015

ORENES ET AL.

Dow

nloa

ded

by [

Uni

vers

ity o

f L

a L

agun

a V

icer

rect

orad

o] a

t 09:

16 1

9 N

ovem

ber

2015

of the fixations on the target and competitor image,yielding fixation probabilities for the target image.Since eye-blinks, off-screen fixations, and empty-space fixations together accounted for no morethan 6% of the data at any given point in time,

proportions of looks to the competitor imagewere more or less complementary to proportionsof looks to the target image.

Figure 2 depicts probabilities of fixations on thetarget region of interest as a function of time and

Figure 2. Probabilities of fixations on the target interpretation image (the rich man) over time, separately for affirmative and negative

sentences within neutral, inconsistent, and consistent contexts. CW = critical word. (a) Neutral contexts. Before the onset of the target

word (t, 0), participants focused on both images about equally frequently because the context was neutral. From about 500 ms after target

word onset, probabilities of fixations on the target interpretation image started to increase for affirmative and negative sentences. Error

bars represent 95% confidence intervals within subjects such that no overlap between conditions indicates a significant difference (see Morey,

2008; O’Brien & Cousineau, 2014). (b) Inconsistent contexts. In this context, we observed that participants changed from initially

preferring the poor man (considerably less than 50% of fixations on the target interpretation image early on) to the rich man after hearing

the critical word. As the two curves indicate, this happened earlier in affirmative than in negative sentences. Error bars represent 95%

confidence intervals within subjects such that no overlap between conditions indicates a significant difference (see Morey, 2008; O’Brien &

Cousineau, 2014). (c) Consistent contexts. In this context, we observed that participants were already biased towards the target

interpretation image early on (around 70% of fixations were on the target interpretation image before the critical word was heard). In

both conditions, proportions of looks to the target first dropped to around 60% and then rose towards an asymptote of around 80% after

processing the critical word. This rise towards asymptote happened much earlier in the affirmative than in the negative sentence condition.

Error bars represent 95% confidence intervals within subjects such that no overlap between conditions indicates a significant difference (see

Morey, 2008; O’Brien & Cousineau, 2014). To view this figure in colour, please visit the online version of this Journal.

THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2015 5

NEGATION IN CONTEXT

Dow

nloa

ded

by [

Uni

vers

ity o

f L

a L

agun

a V

icer

rect

orad

o] a

t 09:

16 1

9 N

ovem

ber

2015

context. Figure 2a corresponds to the neutralcontext, where people heard a context that wasnot related to the critical sentence (or target)—forexample, “Her dad lived on the other side of town”.In this case, the context does not favour any ofthe images that are shown on the screen; thereforelisteners initially look at both images equally often.When participants listened to the affirmative sen-tence—for example, “Her dad was rich”—and thenegative sentence—for example, “Her dad was notpoor”—fixation probabilities on the targetinterpretation image increased (e.g., the richman), but this increase was earlier for affirmativesentences than for negative sentences.

Figure 2b shows the data for the inconsistentcontext condition, where participants heard thecritical affirmative or negative sentence followinga context that was inconsistent with what was men-tioned in the critical sentence. As can be seen,people started to look at the target interpretationimage earlier in the affirmative than in the negativesentence condition.

Figure 2c shows thedata for the consistent contextcondition. In this context, an affirmative (e.g., her dadwas poor after She supposed that her dad had littlesavings) and a negative sentence (e.g., Her dad wasnot poor after She supposed that her dad had enoughsavings) ultimately led to the samepatternoffixations.The difference is in the timing: Participants fixate onthe target more quickly following an affirmative thanfollowing a negative statement.

T-test against baselineTo avoid problems inherent to proportional data,participant and item averages were arcsin-trans-formed prior to t-test comparisons. Given that180–200 ms are usually assumed to account forsaccade programming (Martin, Shao, & Boff,1993), the mean of the first time-region (0–100ms) was considered to be the baseline and wasused to conduct statistical comparisons againstmeans on each time point until 3000 ms later (fora similar method, see Huettig & Altmann, 2011).This correction to baseline allowed us to controlfor any bias in the pattern of fixations on figurescaused by the type of context. A false discoveryrate (FDR) thresholding procedure was used to

effectively control for the Type 1 error due to mul-tiple comparisons (58 for each condition;Genovese, Lazar, & Nichols, 2002). We elimi-nated one subject due to problems in the recordingof her eye movements.

In the neutral context, there was an increase infixation on the target interpretation for affirmativesentences at 500 ms [pFDRcorr1= .011(subjectanalysis); pFDRcorr2= .031(item analysis)] and fornegative sentences at 800 ms (pFDRcorr1= .007;pFDRcorr2= .001). In the inconsistent context, thesignificant differences with respect to the baselinestarted at 500 ms for affirmative sentences(pFDRcorr1= .016; pFDRcorr2= .002) and at 600 msfor negative sentences (pFDRcorr1= .014;pFDRcorr2= .025). It is important to keep in mindthat while affirmative sentences are processed simi-larly in both contexts, negative sentences are facili-tated in the inconsistent context when the negatedsituation is already constructed. In the consistentcontext condition, there was an increase in fixationon the target interpretation starting at 650 ms(pFDRcorr1= .002; pFDRcorr2= .006) for affirmativesentences, while there was a decrease in fixation onthe target interpretation at the beginning, at thetemporal point of 800 ms, (pFDRcorr1= .048; andbetween 450 and 700 ms in the item analysis:pFDRcorr2= .028) and then there was anincrease in fixation starting at 1450–1550 ms(pFDRcorr1= .039; and from 2000 ms in the itemanalysis pFDRcorr2= .047) for negative sentences.In this last context, we can see that participantsneeded to build first the negated situation around800 ms because it has not been constructed pre-viously and then the actual situation around 1450ms. In conclusion, pragmatic information clearlyhas an influence on the unfolding processing ofnegation.

The growth-curve analysisThis analysis is a multilevel regression techniquedesigned for analysis of time course with orthog-onal power polynomials (see Mirman, Dixon, &Magnuson, 2008). First, we modelled the effectof polarity separately for each verbal context.With this method, we can examine whether thedifferences between affirmative and negative

6 THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2015

ORENES ET AL.

Dow

nloa

ded

by [

Uni

vers

ity o

f L

a L

agun

a V

icer

rect

orad

o] a

t 09:

16 1

9 N

ovem

ber

2015

sentences lay at the intercept (which would reflectdifferences between overall fixation ratios), at thelinear term (which would reflect differences at amonotonic change in fixation ratios), and at thequadratic term (which would reflect differences atthe symmetric rise and fall rate of fixation ratiosaround a central inflection point). Second, we mod-elled whether differences between affirmative andnegative sentences varied across contexts. Wereport a deviance statistic, often called –2LL(minus 2 times the log-likelihood), the changes indeviance (AD), and the p-value. The change indeviance allows us to test whether including theparameter increases the fit of the model (seeTable 2).

The results of Table 2 show that affirmative andnegative sentences differed significantly across con-texts. In the neutral context, there was a significanteffect at the intercept, which indicates that fixationon the target interpretation was higher for affirma-tive sentences than for negative sentences. There

was also a difference at the linear term (slope),reflecting a different increase in fixation for affirma-tive and negative sentences. That is, while affirma-tive sentences increased at 500 ms, negativesentences increased at 800 ms. The inconsistentcontext had significant effects on the linear andthe quadratic terms. These effects reflect that affir-mative sentences had a steeper slope; while affirma-tive sentences increased at 500 ms, negativesentences increased at 600 ms. The largest effectwas on the quadratic term. That is, the differencebetween affirmative and negative sentences was pri-marily in the rise and fall of the fixation probabilitycurves. Figure 2b shows that the temporal locationof the peak is around 1000 ms for affirmative sen-tences and 1500 ms for negative sentences. In theconsistent context, we only found differencesbetween affirmative and negative sentences in thequadratic term. This result reflects differences inthe rise and fall of fixation proportions betweenaffirmative and negative sentences. Figure 2cshows that participants decreased their attentionon the target around 400 ms (the peak), andthen they increased their attention, with the tem-poral location of the peak around 700 ms for affir-mative sentences. However, the proportions aredifferent for negative sentences, where the lowestpeak was around 700 ms, and the highest peakwas around 2500 ms. On the other hand, thedifferences between affirmative and negative sen-tences varied across contexts in the quadraticterm. This indicates that the main differencebetween contexts is in the distribution proportionrather than the intercept or slope. To sum up,these results indicate that affirmative and negativesentences result in different target fixation profilesacross all contexts.

Discussion

The results are consistent with the view that prag-matic knowledge influences the process of under-standing negation. We found an increase infixation compared to the baseline on the target fornegative sentences at 800 ms in the neutralcontext condition, 600 ms in the inconsistentcontext condition, and 1450 ms in the consistent

Table 2. GCA results of the comparisons between affirmative and

negative conditions within each context and of the comparison of the

polarity effect across contexts

Model –2LL AD p

Neutral context

Base 69,898.61

Intercept 69,894.01 4.60 .032

Linear 69,865.69 28.32 ,.001

Quadratic 69,862.95 2.74 .097

Inconsistent context

Base 70,355.92

Intercept 70,355.04 0.88 .348

Linear 70,345.93 9.11 .002

Quadratic 70,327.16 18.77 ,.001

Consistent context

Base 69,622.72

Intercept 69,622.72 0 .947

Linear 69,620.60 2.12 .146

Quadratic 69,575.04 45.56 ,.001

Contexts

Base 110,835.6

Intercept 110,835 0.6 .74

Linear 110,829 6.00 .0505

Quadratic 110,816.9 12.1 .002

Note: GCA = growth-curve analysis. Polarity effect: affirmative

minus negative.

THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2015 7

NEGATION IN CONTEXT

Dow

nloa

ded

by [

Uni

vers

ity o

f L

a L

agun

a V

icer

rect

orad

o] a

t 09:

16 1

9 N

ovem

ber

2015

context condition. These differences are smaller foraffirmative sentences, at around 500 ms in theneutral and inconsistent context condition, and ataround 650 ms in the consistent context condition.Therefore, it seems that pragmatic information, asmanipulated in this experiment, modulates theunfolding processing of negation more than ofaffirmation.

On the other hand, our results show that nega-tive sentences are always slower to process thanaffirmation regardless of context. This finding isat odds with many studies that indicate that nega-tion and affirmation can be processed in a similarway (Dale & Duran, 2011; Givon, 1978;Glenberg et al., 1999; Johnson-Laird & Tridgell,1972; Nieuwland & Kuperberg, 2008; Strawson,1952; Wason, 1972). That is, while our resultshave shown that pragmatic information canindeed aid the processing of negation, the latterstill remained more difficult to process than affir-mation regardless of context. This could suggestthat, compared to other methods, visual worldeye-tracking data may be more sensitive to subtle,context-independent processing differencesbetween affirmative and negative sentences.

In the neutral context, we found the classicaleffect of negation (see Carpenter & Just, 1975;Clark & Chase, 1972; Mayo et al., 2004;Trabasso et al., 1971; Wason & Johnson-Laird,1972; Wason & Jones, 1963). Growth-curve analy-sis revealed an effect at the intercept, indicating thatfixations on the target were more likely (within theconsidered time period) for affirmative sentencesthan for negative sentences. More importantly,there was also a difference in the linear term,reflecting differential increases in target fixationprobability over time for affirmative (at ∼500 ms)and negative sentences (at ∼800 ms). That is, par-ticipants were mainly slower to attend to the targetin negative than in affirmative sentences (see also:Orenes, Beltrán, & Santamaría, 2014). Thisfinding could be related to the fact that negatedconcepts are less readily recalled than affirmativeones (Fillenbaum, 1966; Lea & Mulligan, 2002;MacDonald & Just, 1989).

In the consistent context, participants heard thenegative sentenceHer dad was not poor after hearing

She supposed that her dad had enough savings. In thiscase, participants switched their attention from theimage of the rich man (after hearing the context)toward the image of the poor man (after hearingnegation) 800 ms after hearing the critical word(“poor”). After that, they switched their visualattention again from the poor man towards therich man (at around 1450 ms). These two compu-tations take extra time (see also Johnson-Laird &Tridgell, 1972). It is interesting to note that a com-parable increase in the number of fixations on thetarget started at around 650 ms from critical wordonset in affirmative sentences. This delay in com-parison to the other context conditions may bedue to the target not being very informative—thatis, it confirms information already suggested bythe context. As the theory of relevance purports,communication must be relevant (Grice, 1961;Wilson & Sperber, 2004); however, in the consist-ent context condition, after listening to, forexample, She supposed that her dad had littlesavings, listeners already inferred that her dadmust have been poor, thus when they actuallyheard the affirmative Her dad was poor, this infor-mation was less relevant in the sense that partici-pants already knew it. As a result, (re)inspectingthe target was somewhat delayed.

The process of understanding negation is facili-tated when the negated situation has been pre-sented previously (Givon, 1978; Khemlani et al.,2012; Lüdtke & Kaup, 2006; Wason, 1965), butit is still more difficult than understanding affirma-tion. This suggests that the differences betweennegative and affirmative sentences are not purelypragmatic. In the inconsistent context, in whichthe negated situation was presented previously, wefound differences in the linear term, reflecting adifferent increase in fixation for affirmative (ataround 500 ms) and negative sentences (at around600 ms). There was also a difference in the quadra-tic term, reflecting differences at rise and fall in fix-ation proportions between affirmative and negativesentences. That is, the highest peak happenedaround 1000 ms for affirmative sentences ms andaround 1500 ms for negative sentences.

The advantage of using the visual world para-digm is that this method allows us to see how

8 THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2015

ORENES ET AL.

Dow

nloa

ded

by [

Uni

vers

ity o

f L

a L

agun

a V

icer

rect

orad

o] a

t 09:

16 1

9 N

ovem

ber

2015

people update their current interpretations onlinewhile processing negation within different con-texts. For example, when participants werefocused on the poor man image after hearing theinconsistent context, and then they heard he wasnot poor, they switched toward the target image(the rich man) at 600 ms. However, as they werefocused on both images equally often in theneutral context, when they heard he was not poor,they switched toward the rich man target imagewith a delay compared to the inconsistentcontext, at 800 ms. The visual world paradigmwas therefore sensitive enough to detect the effectof pragmatic.

It is interesting to note that affirmative andnegative sentences were easier when they wereused to deny a false belief (inconsistent context)than when no prior beliefs were established(neutral context) because previous literaturesuggests that once beliefs are held, they should bedifficult to change. That is, people tend to maintaintheir beliefs even when faced with evidence to thecontrary (Khemlani & Johnson-Laird, 2012;Orenes, Navarrete, Beltrán, & Santamaría, 2012).Based on this premise, one could predict thatwhen people listen to a context that is inconsistentwith the target, the latter must be processed moreslowly (see Ferguson, Sanford, & Leuthold,2008). However, our results show that both affir-mative and negative sentences are processed morequickly in the inconsistent context (when they areused to deny or change a belief) than in a neutralcontext. This could probably be due to a primingeffect. In our experiment, we presented a contextthat led to a prediction—for example, whetherthe critical protagonist (“her dad”) was rich orpoor. Given the context She supposed that her dadhad little savings, for example, there is a predictionthat “her dad” is poor. In the inconsistent con-dition, this was then rejected with “her dad wasnot poor”. The neutral context Her dad lived onthe other side of town makes no such predictionabout the critical protagonist’s financial status. AsKhemlani et al. (2012) pointed out, individualsshould find it easier to understand a negation ifthey have already constructed the models of thecorresponding affirmative assertion.

In conclusion, pragmatic information influencesthe processing of negation more than affirmation,especially when the negated situation is alreadyconstructed, but it does not explain all of the differ-ence in difficulty between negative and affirmativesentences. Negation is, clearly, a complex phenom-enon, and further research should be conducted toits apprehension.

REFERENCES

Allopenna, P. D., Magnuson, J. S., & Tanenhaus, M. K.(1998). Tracking the time course of spoken word rec-ognition using eye movements: Evidence for continu-ous mapping models. Journal of Memory and

Language, 38, 419–439.Anderson, S., Huette, S., Matlock, T., & Spivey, M.

(2010). On the temporal dynamics of negated percep-tual simulations. In F. Parrill, V. Tobin, &M. Turner(Eds.), Meaning form and body (pp. 1–20). StanfordUniversity, CA: CSLI Publications.

Byrne, R. M. J. (2005). The rational imagination: How

people create alternatives to reality. Cambridge, MA:MIT Press.

Carpenter, P. A., & Just, M. A. (1975). Sentence com-prehension: A psycholinguistic processing model ofverification. Psychological Review, 82, 45–73.

Clark, H. H., & Chase, W. G. (1972). On the process ofcomparing sentences against pictures. Cognitive

Psychology, 3, 472–517.Cooper, R. M. (1974). The control of eye fixation by the

meaning of spoken language: A new methodology forthe real-time investigation of speech perception,memory, and language processing. Cognitive

Psychology, 6, 84–107.Dale, R., & Duran, N. D. (2011). The cognitive

dynamics of negated sentence verification. CognitiveScience, 35, 983–996.

Duñabeitia, J. A., Avilés, A., Afonso, O., Scheepers, C.,& Carreiras, M. (2009). Qualitative differences in therepresentation of abstract versus concrete words:Evidence from the visual-world paradigm.Cognition, 110, 284–292.

Ferguson, H. J., Sanford, A. J., & Leuthold, H. (2008).Eye-movements and ERPs reveal the time course ofprocessing negation and remitting counterfactualworlds. Brain Research, 1236, 113–125.

Fillenbaum, S. (1966). Memory for gist: Some relevantvariables. Language and Speech, 9, 217–227.

THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2015 9

NEGATION IN CONTEXT

Dow

nloa

ded

by [

Uni

vers

ity o

f L

a L

agun

a V

icer

rect

orad

o] a

t 09:

16 1

9 N

ovem

ber

2015

Genovese, C. R., Lazar, N. A., & Nichols, T. (2002).Thresholding of statistical maps in functional neuroi-maging using the false discovery rate. Neuroimage,15(4), 870–878.

Givón, T. (1978). Negation in language: Pragmatics,function, ontology. Syntax and Semantics, 9, 69–105.

Glenberg, A. M., Robertson, D. A., Jansen, J. L., &Johnson-Glenberg, M. C. (1999). Not propositions.Journal of Cognitive Systems Research, 1, 19–33.

Grice, H. P. (1961). The causal theory of perception.Proceedings of the Aristotelian Society, Supp., XXXV,121–153.

Hasson, U., & Glucksberg, S. (2006). Does understand-ing negation entail affirmation? An examination ofnegated metaphors. Journal of Pragmatics, 38, 1015–1032.

Huette, S., & Anderson, S. (2012). Negation withoutsymbols: The importance of recurrence and contextin linguistic negation. Journal of Integrative

Neuroscience, 11, 295–312.Huette, S., Anderson, S., Matlock, T., & Spivey, M.

(2011, July 20–23). A one-stage distributed processing

account of linguistic negation. Proceedings of the 33rdannual meeting of the Cognitive Science Society,Boston, MA, pp. 2037–2042.

Huettig, F., & Altmann, G. T. M. (2011). Looking atanything that is green when hearing “frog”: Howobject surface colour and stored object colour knowl-edge influence language-mediated overt attention.The Quarterly Journal of Experimental Psychology,64(1), 122–145.

Johnson-Laird, P. N., & Tridgell, J. (1972). When nega-tion is easier than affirmation. Quarterly Journal of

Experimental Psychology, 24, 87–91.Kaup, B., Lüdtke, J., & Zwaan, R. A. (2006). Processing

negated sentences with contradictory predicates: Is adoor that is not open mentally closed? Journal of

Pragmatics, 38, 1033–1050.Kaup, B., Zwaan, R. A., & Lüdke, J. (2007). The experi-

ential view of language comprehension. How is nega-tion represented? In F. Schmalhofer & C. A. Perfetti(Eds.),Higher language processes in the brain (pp. 255–289). Mahwah, NJ: Lawrence Erlbaum Associates.

Khemlani, S., & Johnson-Laird, P. N. (2012). Hiddenconflicts: Explanations make inconsistencies harderto detect. Acta Psychologica, 139, 486–491.

Khemlani, S., Orenes, I., & Johnson-Laird, P. N.(2012). Negation: A theory of its meaning, represen-tation, and use. Journal of Cognitive Psychology, 24,541–559.

Lea, R. B., & Mulligan, E. J. (2002). The effect of nega-tion on deductive inferences. Journal of ExperimentalPsychology: Learning, Memory, and Cognition, 28(2),303–317.

Lüdtke, J., Friedrich, C. K., De Filippis, M., & Kaup, B.(2008). ERP correlates of negation in a sentence-picture-verification paradigm. Journal of Cognitive

Neuroscience, 20(8), 1355–1370.Lüdtke, J., & Kaup, B. (2006, July 26–29). Context effects

when reading negative and affirmative sentences

(poster). 28th annual conference of the CognitiveScience Society, Sheraton Vancouver Wall Centre,Vancouver, BC, Canada.

MacDonald, M. C., & Just, M. A. (1989). Changes inactivation levels with negation. Journal of

Experimental Psychology: Learning, Memory, and

Cognition, 15(4), 633–642.Martin,E.,Shao,K.C.,&Boff,K.R. (1993).Saccadicover-

head: Information-processing time with and withoutsaccades. Perception & Psychophysics, 53, 372–380.

Mayo, R., Schul, Y., & Burnstein, E. (2004). “I amguilty” vs. “I am innocent”: Successful negation maydepend on the schema used for its encoding. Journalor Experimental Social Psychology, 40, 443–449.

Mirman, D., Dixon, J. A., & Magnuson, J. S. (2008).Statistical and computational models of the visualworld paradigm: Growth curves and individualdifferences. Journal of Memory and Language, 59,475–494.

Morey, R. D. (2008). Confidence Intervals from nor-malized data: A correction to Cousineau (2005).Tutorial in Quantitative Methods for Psychology, 4

(2), 61–64.Moxey, L. M., & Sanford, A. J. (1987). Quantifiers and

focus. Journal of Semantics, 5, 189–206.Nieuwland, M. S. & Kuperberg, G. R. (2008). When the

truth is not too hard to handle: An event-relatedpotential study on the pragmatics of negation.Psychological Science, 19(12), 1213–1218.

O’Brien, F., & Cousineau, D. (2014). Representing errorbars in within-subject designs in typical softwarepackages. Tutorial in Quantitative Methods for

Psychology, 10(1), 56–67.Orenes, I., Beltrán, D., & Santamaría, C. (2014). How

negation is understood: Evidence from the visualworld paradigm. Journal of Memory and Language,74, 36–45.

Orenes, I., Navarrete, G., Beltrán, D., & Santamaría, C.(2012). Schizotypal people stick longer to their firstchoices. Psychiatry Research, 200, 620–628.

10 THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2015

ORENES ET AL.

Dow

nloa

ded

by [

Uni

vers

ity o

f L

a L

agun

a V

icer

rect

orad

o] a

t 09:

16 1

9 N

ovem

ber

2015

Rayner, K. (1998). Eye movements in reading and infor-mation processing: 20 years of research. PsychologicalBulletin, 124(3), 372–422.

Stalnaker, R. (1998). On the representation of context.Journal of Logic, Language and Information, 7, 3–19.

Strawson, P. (1952). Introduction to logical theory.London: Methuen and CO.

Tanenhaus, M. K., Spivey-Knowlton, M. J., Eberhard,K. M., & Sedivy, J. C. (1995). Integration of visualand linguistic information in spoken language com-prehension. Science, 268(5217), 1632–1634.

Tian, Y., Breheny, R., & Ferguson, H. J. (2010). Whywe stimulate negated information: A dynamic prag-matic account. Quarterly Journal of Experimental

Psychology, 63, 2305–2312.Trabasso, T., Rollins, H., & Shaughnessy, E. (1971).

Storage and verification stages in processing concepts.Cognitive Psychology, 2, 239–289.

Wason, P. C. (1965). The contexts of plausible denial.Journal of Verbal Learning and Verbal Behavior, 4,7–11.

Wason, P. C. (1972). In real life negatives are false.Logique et Analyse, 15, 17–38.

Wason, P. C., & Johnson-Laird, P. N. (1972). Psychologyof reasoning: Structure and content. Cambridge, MA:Harvard University Press.

Wason, P. C., & Jones, S. (1963). Negatives: Denotationand connation. British Journal of Psychology, 54,299–307.

Wilson, D., & Sperber, D. (2004). Relevance theory. InL. R. Horn & G. Ward (Eds.), The handbook of prag-matics (pp. 607–632). Oxford: Blackwell.

Zwaan, R. A., Stanfield, R. A., & Yaxley, R. H.(2002). Language comprehenders mentally representthe shape of objects. Psychological Science, 13, 168–171.

THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2015 11

NEGATION IN CONTEXT

Dow

nloa

ded

by [

Uni

vers

ity o

f L

a L

agun

a V

icer

rect

orad

o] a

t 09:

16 1

9 N

ovem

ber

2015