Journal of Memory and Language 70 (2014) 68–84

Contents lists available at ScienceDirect

Journal of Memory and Language

journal homepage: www.elsevier .com/locate / jml

Not only the apples: Focus sensitive particles improve memoryfor information-structural alternatives

0749-596X/$ - see front matter � 2013 Elsevier Inc. All rights reserved.http://dx.doi.org/10.1016/j.jml.2013.09.001

⇑ Corresponding author. Address: Humboldt-Universität zu Berlin,Institut für deutsche Sprache und Linguistik, Unter den Linden 6, 10099Berlin, Germany.

E-mail address: katharina.spalek@staff.hu-berlin.de (K. Spalek).

Katharina Spalek a,c,⇑, Nicole Gotzner a,c, Isabell Wartenburger b,c

a Department of German Language and Linguistics, Humboldt-Universität zu Berlin, Germanyb Department of Linguistics, Universität Potsdam, Germanyc Collaborative Research Centre SFB 632 ‘‘Information Structure’’, Germany

a r t i c l e i n f o a b s t r a c t

Article history:Received 1 October 2012Received in revised form 18 June 2013Available online 6 November 2013

Keywords:Information structureFocus particlesAlternative setDelayed recallMemory

Focus sensitive particles highlight the relevance of contextual alternatives for the interpre-tation of a sentence. Two experiments tested whether this leads to better encoding andtherefore, ultimately, better recall of focus alternatives. Participants were presented withauditory stimuli that introduced a set of elements (‘‘context sentence’’) and continued inthree different versions: the critical sentences either contained the exclusive particle nur(‘‘only’’), the inclusive particle sogar (‘‘even’’), or no particle (control condition). After beingexposed to blocks of ten trials, participants were asked to recall the elements in the contextsentence. The results show that both particles enhanced memory performance for thealternatives to the focused element, relative to the control condition. The results supportthe assumption that information-structural alternatives are better encoded in memory inthe presence of a focus sensitive particle.

� 2013 Elsevier Inc. All rights reserved.

Introduction

The ultimate aim of language comprehension is the suc-cessful construction of a mental model (e.g., Johnson-Laird,1983; van Dijk & Kintsch, 1983), which represents themeaning of an utterance that has been perceived. Further-more, this meaning representation should be long-lasting,such that an interlocutor will be able to remember the gistof a conversation at some later point in time. In the presentpaper we investigate how recall of a dialogue is modulatedby the presence of focus sensitive particles like only andeven.

In prevalent linguistic theories on focus, the function offocus is to indicate ‘‘the presence of alternatives that arerelevant for the interpretation of linguistic expressions’’(Krifka, 2007: p. 18; see also Rooth, 1992). The additional

function of a focus sensitive particle is that it establishesa specific relation between a focused element and its setof alternatives (cf. König, 1991; Rooth, 1992). Consider asentence like the one in (1):

(1) John only saw Sue at the dinner party

In (1), the particle only associates with the element infocus John and expresses that, among the set of possiblealternatives {Sue, Mary, Paul,. . .}, John saw no-one elsebut Sue. Thereby, the focus operator excludes the alterna-tives to the focused element. From a theoretical perspec-tive, focus operators can be divided into subclasses ofinclusive, exclusive, and scalar particles (see e.g., Krifka,1999; König, 1991). While exclusives like nur (‘‘only’’) ex-press that the focused element and its alternatives do notshare the property expressed by the predicate, inclusiveslike auch (‘‘also’’) and sogar (‘‘even’’) have an additivemeaning component, indicating that the assertion alsoholds for at least one of the alternatives. Standard theoriesof focus particles further identify a presupposition of like-lihood in the scalar particle sogar. That is, sogar induces a

K. Spalek et al. / Journal of Memory and Language 70 (2014) 68–84 69

scale of likelihood on members of the alternative set andassigns the focused element the lowest value.

An alternative set is not only induced by focus sensitiveparticles but also by a contrastive focus accent (L + H� inthe ToBI system, cf. Silverman et al., 1992). Using contras-tive prosody, Husband and Ferreira (2012) investigatedhow a contrast set is generated during language compre-hension. Using a cross-modal priming paradigm, they pre-sented participants with auditory sentences like ‘‘Themuseum thrilled the sculptor when they called . . .’’. Theword sculptor (the prime) was either pronounced with con-trastive or with neutral prosody. Only the cases in whichsculptor was pronounced with contrastive prosody are rel-evant here. In their first experiment, a visual word was pre-sented right after sculptor. The word could be a contrastivesemantic associate (painter), a non-contrastive semanticassociate (statue), or unrelated. Reaction times to both con-trastive and non-contrastive semantic associates were fas-ter than reaction times to an unrelated control word.However, in a second experiment, a delay of 750 ms wasintroduced between prime and target. In this experiment,only contrastive associates were replied to faster. Husbandand Ferreira assume that, initially, activation spreads to allsemantically related words, but that a later suppressionprocess generates the correct alternative set. They reasonthat later access should be inhibited for elements that donot belong to the alternative set (see also a study by Braunand Tagliapietra (2010), demonstrating a cross-modalpriming effect for contrast alternatives after primes pro-duced with contrastive focus accent but not after primesproduced with neutral prosody).

The study described above shows evidence for the on-line activation of an alternative set during language com-prehension. However, a related question would bewhether these alternatives also feature prominently in arepresentation of the processed discourse, that is, whetherthey will be remembered better.

Most research on the effect of information structure onmemory has investigated memory for the focused element.Several studies suggest that focused elements are privilegedin memory representation compared to non-focused con-stituents, i.e., they are remembered with a higher accuracyand are represented at a more fine-grained semantic level(e.g., Birch & Garnsey, 1995; Sturt, Sanford, Stewart, &Dawydiak, 2004). Hence, information structure seems toplay a role in the mental representation of the meaningof an utterance. If the theoretical assumptions of Rooth’sAlternative Semantics (Rooth, 1992) hold, focus indicatesthe presence of alternatives. Therefore, focus should notonly improve memory for the focused element but alsofor information-structural alternatives.

However, this question has been investigated lessintensively. In fact, we only know of one study about thememory representation of information-structural alterna-tives. Fraundorf, Watson, and Benjamin (2010) comparedthree theoretical accounts about the function of contras-tive accenting, two of which were concerned with its effecton alternatives to a focused expression. The granularity ac-count (based on a study by Sanford, Sanford, Molle, & Em-mott, 2006) predicts that a focused item is representedmore specifically, leading to less activation of related items

(i.e., the alternatives). The contrast representation accountdeveloped by Fraundorf et al. (2010) assumes that contras-tive pitch accenting might also strengthen the representa-tion of the contrast items (the alternatives), leading notonly to better encoding for what did happen but also tobetter encoding for what did not happen.

Fraundorf et al. presented participants with short re-corded discourses containing contrast sets consisting oftwo items. One of these items was specified, using eithernon-contrastive (H�) or contrastive (L + H�) pitch accent.In a later recognition memory test, the L + H� accent in-creased both the number of hits to correct statements,and the number of correct rejections of the contrast item.The latter finding suggests that contrastive pitch accenton a focused element also enhances memory for its alter-natives, facilitating participants’ decision that somethingdid not happen. The findings support the contrast repre-sentation account: contrastive pitch accent is used by lis-teners to encode information about all elements in thecontrast set.

To sum up, previous findings suggest that focused infor-mation and information-structural alternatives have aprivileged memory representation. In the present study,we investigate the impact of focus sensitive particles onthe representation of the mental model that listeners con-struct from an utterance. In contrast to focus accenting,which indicates the presence of alternatives, a focus parti-cle necessarily instantiates a contextually-salient set ofalternatives, because this alternative set is part of itsmeaning (cf. Beaver & Clark, 2008, who distinguish weakfocus induced by contrastive pitch accent and strong focusinduced by focus particles, and König, 1991). Therefore,alternatives might be even more salient when a focus par-ticle is used, compared to cases where focus is indicated bymeans of pitch accent alone. In the following two experi-ments, we investigated if alternatives to a focused elementare recalled even better if the focused element is not onlyaccented but also preceded by a focus particle. In addition,we investigated whether exclusive and inclusive particlesdifferentially affected the representation and hence the re-call of focus alternatives.

Experiment 1

We used a delayed recall experiment to measure the ef-fects of the presence or absence of inclusive and exclusivefocus sensitive particles on memory for information-struc-tural alternatives. Participants were presented auditorilywith short dialogues and were later asked to recall the ele-ments mentioned in the story. The critical manipulationwas whether the utterances contained an inclusive (sogar),an exclusive (nur), or no (control condition) focus sensitiveparticle. The measure of interest was the number ofremembered alternatives.

Based on the lexical meaning of the focus particles, thefollowing hypothesis can be formulated: memory for thealternative set should be worst for exclusives, intermediatein the control condition and best for inclusives. This pre-diction was based on the assumption that focus sensitiveparticles alter the representation of elements in the alter-

70 K. Spalek et al. / Journal of Memory and Language 70 (2014) 68–84

native set. While inclusive particles mark the alternativesto be part of the predication, exclusives express that thealternatives to the focused constituent do not hold. There-fore, memory for information-structural alternativesshould be enhanced when the alternatives are highlightedby an inclusive particle like sogar, while memory should bereduced in the case of the particle nur. We will refer to thishypothesis as the lexical meaning hypothesis.

Alternatively, however, both conditions with focusoperators might lead to better memory for the alternativeset, compared to the control condition without a focus sen-sitive particle, since both particles make reference to a setof alternatives and highlight its relevance for interpreta-tion (Krifka, 2007). We will refer to this hypothesis as thecontrast hypothesis.

Method

Participants26 native speakers of German (21 female, mean age

27.1 years, SD 3.8, age range 22–31) were recruited fromthe subject pool at the Institute of Psychology of HumboldtUniversity and paid 7 Euros in compensation. None ofthem reported any vision or hearing difficulties. The dataof two subjects were excluded from the analysis due totechnical problems. The remaining participants were 19women and 5 men with a mean age of 27.4.

MaterialsParticipants listened to 50 prerecorded dialogues con-

taining two context sentences and a critical sentence. Thecontext sentences introduced a set of three elements, aperson, and an assumption about a particular event refer-ring to the previously mentioned elements (cf. (2)). Thecritical sentence was spoken by a different speaker and re-vised the assumption of the first speaker. The context sen-tences were recorded by a male speaker and the criticalsentences were spoken by a female speaker (the secondauthor of this paper) who was trained on focus accentua-tion. Both speakers had a middle German accent close tothe standard variety of German and recording took placein a sound proof room. For the critical sentence, three ver-sions were recorded: it either contained (a) the exclusiveparticle nur (‘‘only’’), (b) the inclusive particle sogar(‘‘even’’), or (c) no focus operator as a control condition.In the experiment, one of these three versions was ran-domly assigned to a given experimental list (i.e., neithercontext sentences nor critical sentences were repeatedwithin participants).

(2)Context sentences (speaker 1):In der Obstschüssel liegen Pfirsiche, Kirschen und Bananen.In the fruit bowl, there are peaches, cherries, and bananas.Ich wette, Carsten hat Kirschen und Bananen gegessen.I bet Carsten ate cherries and bananas.

Critical sentences (speaker 2):(a) Nein, er hat nur Pfirsiche gegessen.(b) Nein, er hat sogar Pfirsiche gegessen.(c) Nein, er hat Pfirsiche gegessen.No, he (a) only/(b) even/(c) _ ate peaches.

The three elements given in the context sentences werepresented in random order. The second sentence was spo-ken with a pitch accent on the two elements (in example(2): cherries, bananas). In the critical sentences, the thirdelement which had not been mentioned in the second con-text sentence appeared as the focused element (in example(2): peaches), thereby ensuring that each element wasmentioned twice in the dialogue. Whether the first, second,or third element presented in the first context sentencewas focused in the critical sentence, was counterbalancedacross items. The critical sentences with particles werepronounced with a hat contour (cf. Féry, 1993), having apitch accent on the focus particle as well as the focusedelement and the condition without particle had a fallingpitch accent on the focused constituent. The pitch contourof the critical sentence from example (2) is shown inAppendix A.

Acoustic analyses were conducted on the focused ele-ment to compare the acoustic parameters in the threeconditions. Table 1 presents means and standard errorsfor duration, maximum pitch, minimum pitch, differencebetween maximum and minimum pitch, mean pitch, andthe relative points of maximum and minimum pitch forboth the stressed syllable and the entire word. The fo-cused elements had similar prosodic properties acrossthe three conditions except for the mean pitch whichwas higher in the condition without a particle. However,since the pitch difference and maximum and minimumpitch did not differ significantly across conditions, the fo-cused element is unlikely to have been more prominentin the condition without a particle. Fig. 1 displays themean pitch contours on the focused element for thethree experimental conditions. Importantly, the accenttype of the focused element was the same acrossconditions.

For the construction of each of the 30 experimentalitems,1 a unique setting was chosen, that is, if the ele-ments appeared in a fruit bowl in one item (as in ourexample (2)), no fruit bowl was mentioned in any of theother items. This was necessary to avoid confusion duringthe recall phase (see below), when participants would beasked ‘‘What was in the fruit bowl?’’. 90 German high fre-quency nouns (more than 1 occurrence per million accord-ing to the DLEX database, Heister et al., 2011) fromtaxonomic categories served as the list of elements (cf.Schröder, Gemballa, Ruppin, & Wartenburger, 2012, forthe category norms). In each item, there were three nounsand the number of syllables was kept as constant as pos-sible within an item (see Appendix B for all experimentalitems). For the second context sentence, 23 common Ger-man verbs (normalized type frequency range: 1–68) wereused that described simple actions (some verbs like ein-kaufen (to buy) were used twice) and 30 German names(male and female).

A set of 20 filler items was constructed in order to dis-courage participants from concentrating only on the list ofelements. The fillers had exactly the same structure as theexperimental items and the elements chosen were also

1 We refer to an entire dialogue as an experimental item.

Table 1Mean acoustic parameters of the focused element in the critical sentences (e.g., Pfirsiche in (2a,b,c)).

Measure only no particle even F(2,29) p

Mean SE Mean SE Mean SE

Stressed syllableDuration (s) 0.24 .01 0.23 .01 0.23 0.01 <1 .77Maximum pitch (Hz) 213 2.53 225 8.00 207 7.63 <1 .58Minimum pitch (Hz) 194 2.29 197 4.38 185 3.85 <1 .37Pitch difference 20 1.70 28 8.49 22 5.59 <1 .58Mean pitch (Hz) 203 2.18 210 3.93 195 4.53 11.58 .0001Relative point of pitch maximum (s) 0.15 0.01 0.13 0.01 0.12 0.01 <1 .67Relative point of pitch minimum (s) 0.16 0.02 0.16 0.02 0.18 0.02 1.52 .23

Entire wordDuration (s) 0.49 0.02 0.49 0.02 0.49 0.01 <1 .87Maximum pitch (Hz) 225 8.00 222 8.12 212 6.66 2.58 .85Minimum pitch (Hz) 148 2.46 148 2.94 145 3.03 7.43 .1Pitch difference 78 8.40 74 7.80 67 6.74 <1 .56Mean pitch (Hz) 189 1.65 186 1.74 179 2.37 7.87 .0001Relative point of pitch maximum (s) 0.15 0.02 0.17 0.02 0.16 0.02 2.26 .01Relative point of pitch minimum (s) 0.46 0.02 0.46 0.01 0.43 0.02 <1 .38

1 2 3 4 5

160

180

200

220

Intervall Part

F0 m

ean

Particle Conditionsonly no particle even

Fig. 1. Mean pitch contour of focused element in the critical sentencesacross particle condition.

K. Spalek et al. / Journal of Memory and Language 70 (2014) 68–84 71

from various categories (either taxonomic or non-taxo-nomic). The filler items were also combined with one ofthe particle conditions (nur, sogar, no particle).

After having listened to a block of ten items, partici-pants were presented with questions about these items(see Procedure). For the experimental items, the partici-pants were asked to recall the list of elements. The criticalquestion always mentioned the category and the setting,e.g., Which fruits were in the fruit bowl? for example (2).For the filler items, comprehension questions were pre-sented to make sure that participants listened carefullyto the entire dialogue, not just to the nouns mentioned inthe first context sentence. The filler questions either askedabout the setting, the name of the protagonist, the action,or the focus particle. A question about the focus particlewas always posed in such a way that the answer couldbe determined easily, e.g., the question Did Carsten eat ba-nanas and cherries? requires a no-response in the context ofexample (2a) (i.e., No, he only ate peaches.).2 In total, weconstructed a set of 60 filler questions: 21 about the action,19 about the setting, 5 about the person, and 15 about thefocus particle. A participant received 20 of those questionschosen randomly from three different lists.

Each participant was exposed to all 30 experimentalitems and 20 filler items in five blocks of 10 items contain-ing different particle conditions. The conditions were ro-tated across items according to a Latin Square Design, sothat one subject heard an item only once in a particularcondition. This resulted in a total of three lists that werepseudo-randomized for each participant. For randomiza-tion, the program Mix (van Casteren & Davis, 2006) wasused with the following constraints: no more than threefiller or experimental items were presented in a row, a

2 Note that in the case of a sentence without focus particle this questionis possibly indeterminate, since this is not explicitly stated by the sentence.We avoided these kinds of questions in order not to confuse participants.An example question about the particle sogar was Did Carsten eat severalkinds of fruits?, requiring a yes-response.

particular particle condition (nur, sogar, no particle) ap-peared only twice in a row and, within a block, a category(e.g., fruits) appeared only once to control for interferenceeffects.

ApparatusParticipants were seated in a darkened room in front of

an Acer TFT monitor (type Asus 1923d) with a resolution of1280 � 1024 and a refresh rate of 75 Hz (13.3 ms). Stimu-lus presentation was controlled by Neurobehavioral Sys-tems Presentation software (Version 15.1). Subjects woreSennheiser headphones with an integrated microphone.The answers were recorded as wav-Files.

ProcedureThe experiment started with an on-screen-instruction

informing participants about the structure of the experi-ment and the task they would have to perform. Subjects

Table 2Experimental procedure in Experiments 1 and 2.

BLOCK 1Auditory Presentation PhasePresentation: 1 Experimental Item There are shirts, trousers, and jackets in the catalogue.

I bet Matthias has bought shirts and trousers.No, he only bought jackets.

Presentation: 2 Filler They have the flavours chocolate, mint, and vanilla in the ice cream parlor.I bet Nicole has eaten chocolate and vanilla.No, she has even eaten mint.

Presentation: 3 Filler In the botanic garden grow palm trees, cacti, and bonsai.I bet Johannes has made a sketch of palm trees and cacti.No, he has sketched bonsai.

[. . .]Presentation: 10 Experimental Item The storybook deals with witches, princes, and dragons.

I bet Simon has dreamed of princes and dragons.No, he dreamed of witches.

Test Phase (Recall)Cue/ Question: 1 Experimental Item What was in the catalogue?

[Expected response: shirts, trousers, jackets]Cue/ Question: 2 Filler Where was Nicole?

[Expected response: in the ice cream parlor]Cue/ Question: 3 Filler What did Johannes do?

[Expected response: sketched Bonsai][. . .]Cue/ Question: 10 Experimental Item What did the storybook deal with?

[Expected response: witches, princes, dragons]n-back counting taskBLOCK 2Auditory Presentation Phase (10 Items)Test Phase (10 Items)n-back counting task[. . .]

only no particle even

% re

calle

d50

6070

8090

100

Fig. 2. Percentage correct recall for the alternatives.

72 K. Spalek et al. / Journal of Memory and Language 70 (2014) 68–84

were told that, during the test phase, they would only havea limited time frame to respond and that they were sup-posed to respond aloud. After the instructions were dis-played, subjects performed 4 practice trials (a block offour dialogues followed by four questions) and were al-lowed to adjust the sound volume of the headphones.Phases of auditory discourse presentation alternated withtest phases in which participants were cued for recall.The overall structure of the experiment is exemplified inTable 2.

In the auditory presentation phase, each trial beganwith a central fixation cross displayed for 500 ms. Then, athree sentence discourse was presented over headphones.The next trial was immediately initiated by a fixation cross.After a block of 10 trials (lasting approximately 4 min), thetest phase was initiated by a self-paced button press.

In the recall phase, a trial was again started with a fix-ation cross for 500 ms, then the question appeared on thescreen for 3 s, followed by an asterisk that was displayedfor 650 ms. Subjects were supposed to respond orally assoon as the asterisk appeared on the screen. Responseswere recorded, and participants had 14 s to respond(in order to not loose information if a participantresponded too early, i.e., while the question was still onthe screen, recordings actually started from the onset ofthe question and lasted 17 s). Immediately after record-ing, the next trial was initiated by a fixation cross andthe next question was displayed on the screen. All itemswere tested in the same order as presented during thepresentation phase. Thus, the amount of delay between

presentation and test was kept constant and subjects couldeasily keep track of the sequence.

At the end of a block (i.e., auditory presentation phaseand recall phase, cf. Table 2), subjects were asked to per-form a simple n-backward counting task progressing from2 to 6 increments and to take a small break. This was doneto reduce interference effects between blocks, becausesome categories were used more than once (but only oncewithin a given block).

Table 3Fixed effect estimates (top) and variance estimates (bottom) for LMER ofcorrectly recalled alternatives (% recalled � particle condition + trial(centered) + (1|subject) + (0 + trial|subject) + (1|item), N = 713, log-likeli-hood: �3506).

Fixed effect Coefficient SE t p (MCMC)

Intercept 70.57 4.1 17.2Nur 6.29 2.9 2.2 .02Sogar 6.60 2.9 2.3 .03Trial (centered) 0.35 0.1 3.2 .01Random effect s2

Participant 230.37Item 94.53Random Slope s2

Trial 0.13

3 In this analysis, we only included a random factor for item and subject,since the additional effect for trial did not increase explanatory power ofthe model as revealed by model comparison.

4 We carried out separate analyses for recall of focus alternatives andrecall of the focused element, because our theoretical interest lies in theassociation of focus particles with an alternative set and the effect thiswould have on a listener’s ability to recall the alternatives. However, if wewant to claim that the effect was uniquely present for alternatives, wewould have to statistically bolster this claim by showing that the effect offocus condition (nur, sogar, control) interacts with the type of recalledelement (focused vs. alternative). We investigated this by an omnibusgeneralized mixed model that contained all data coded binomially. Thefinal model included particle condition and type of element (focused oralternative) as well as an interaction of these two factors and it furthercontained trial, item and subjects as factors as well as random slopes fortrial. The condition without a particle in the alternatives was chosen asreference level. The model showed a main effect of type of recalled element(p < .0001, showing that a focused element was recalled better than analternative). There was also a significant effect of the focus particles(control vs. nur: p < .05, control vs. sogar: p < .01). However, the interactionwas not significant, suggesting that the effect was equally present forfocused elements and alternatives. Yet, when setting the focused elementas reference level, the effect of the particles disappears, suggesting thatfocus particles affect foremost the alternatives.

K. Spalek et al. / Journal of Memory and Language 70 (2014) 68–84 73

After the experiment, a questionnaire was administeredasking the participant for basic demographic information,what he or she thought the experiment was about andwhether he or she employed any specific strategies. An en-tire testing session lasted about 40 min.

Results

The recorded answers were transcribed and the numberof correct responses was calculated. If a subject mentioneda variant of the presented nouns (e.g., handbag instead ofthe presented noun bag), we coded it as a correct response.

Since we were interested specifically in memory for thealternative set, we split the data into recall of the alterna-tives (e.g., bananas and cherries in (2)) and recall of the fo-cused element (e.g., peaches in (2)). First, we analysed theeffect of the focus particle on the amount of correctly re-called alternatives. The data are presented in Fig. 2.

Without a focus sensitive particle, 70.7% of the alterna-tive items were correctly recalled, with a focus sensitiveparticle, recall performance improved to 76.4% and 77.4%correctly recalled items for nur and sogar, respectively.We fitted a series of linear mixed effects models followingthe procedure described in Baayen (2008). We started outwith a model that had a single random factor (subjects)and subsequently added additional random factors andrandom slopes. Model comparisons by means of F testswere performed on log-likelihood values to single out themodel with the best fit. Only factors that increased themodel’s prediction were kept in the final model.

The model with the best fit included condition and trialnumber (centered) as fixed factors, subjects and items as ran-dom factors, and random slopes for trial number. The controlcondition was chosen as reference level, and p-values wereextracted with Markov chain Monte Carlo (MCMC) samplingwith 10,000 runs. A post hoc Tukey test on the final model(with the R-package multcomp, cf. Bretz, Hothorn, & West-fall, 2010) was carried out for the comparison of nur and so-gar. A summary of the model is given in Table 3.

The analysis revealed a significant difference betweenthe control condition and sogar (t = 2.3, SE = 2.9, p < .05)and a significant effect for nur (t = 2.2, SE = 2.9, p < .05).The post hoc Tukey test showed that the recall perfor-mance for nur and sogar did not differ significantly, p = .90.

Second, we analysed the effect of the focus sensitiveparticle on the amount of correctly recalled focused ele-

ments. The focused element was correctly recalled 81% ofthe time without a focus particle, 82% with nur and 83%with sogar. Because the dependent variable was binomial(the focused element was either recalled or it was notrecalled), we carried out a logit mixed model with particlecondition as fixed factor and subjects and items as randomfactors (cf. Jäger, 2008). There were no effects of the twoparticles (control vs. sogar: p = .86; control vs. nur: p = .62;nur vs. sogar: p = .79).3 Hence, the difference in memoryperformance across particle conditions was uniquely drivenby an effect of the particles on the alternative set.4

Discussion

Experiment 1 used dialogue structures in three ver-sions: a condition with the exclusive particle nur (‘‘only’’),the inclusive particle sogar (‘‘even’’) and a control condi-tion without a particle that also carried a pitch accent onthe focused element. The separate analyses suggested thatboth particles improved participants‘ memory for thealternatives to the focused element, but not memory forthe focused element itself. However, the differential effect(i.e., no effect of focus particles on the focused element)was not borne out by a combined analysis (cf. Footnote4). Hence, it might be the case that a beneficial effect ofthe presence of a focus particle is also observed for recallof the focused element itself (although it did not reach sig-nificance in the separate analysis, see above). Importantly,though, this does not dispute the observation that thepresence of focus particles improves memory for focusalternatives. We will discuss the latter finding first, be-cause we were mainly interested in the effect of focus par-ticles on memory of the alternative set.

Perhaps surprisingly, it was the presence or absence of afocus sensitive particle that affected the recall performancefor the alternatives, whereas the additional meaning com-ponent (inclusive vs. exclusive) of the particle had no ef-fect. That is, memory performance for the alternativeswas not differentially affected by the fact that alternativeswere included or excluded, but only by the fact that focus

74 K. Spalek et al. / Journal of Memory and Language 70 (2014) 68–84

sensitive particles highlight the relevance of an alternativeset, thereby enhancing memory for these alternatives. Thisfinding is in line with the contrast hypothesis outlinedabove.

Note that we are not saying that focus alone did nothave an effect on memory. However, in our experiment,we did not manipulate the presence or absence of (contras-tive) focus. Rather, we investigated the effect of expres-sions that associate with a focused element. With regardto memory for the focused elements, the percentage of cor-rectly recalled items was quite high – on average, 82% ofthe focused items were correctly recalled, correspondingroughly with the findings of Birch and Garnsey (1995).

Does it hurt our conclusion that we cannot rule out thatthe presence of focus particles had a similar effect on the re-call of the focused element itself? We think not, but we be-lieve that the potential beneficial effect on recall for thefocused element stems from a different source. Previousstudies on improved memory for focus (e.g., Birch & Garnsey,1995) have emphasized the greater prominence of the fo-cused element and have argued that this makes it more sali-ent and therefore better remembered later on. In our study,we had either one (pitch accent) or two (pitch accent and fo-cus particle) cues for focus. Therefore, the emphasis on thefocused element might have been even greater in the twoparticle conditions than in the control condition.

We had tried to find a context that allowed the use ofminimal pairs, that is, critical sentences which differedonly in the presence/ absence of a focus particle and thespecific particle used. In particular, the pitch accent onthe focused element also had to be identical in all threecontexts. Given that the two particles we used tend to leadto different contextual assumptions, it was not easy to finda type of context that supported both of them and the ab-sence of a particle equally well. In the materials used, thecritical sentences were a correction of previously intro-duced expectations. In hindsight, these items might nothave been ideal for two reasons, one theoretical, one basedon the exact materials used. First, because a correction wasused, the focus in all critical sentences was contrastive in anarrow sense (cf. e.g., Krifka, 2007) and there was a strongexhaustiveness implicature in the condition without a fo-cus operator. That is, if we consider the sentence ‘‘No, Car-sten ate pears’’, the natural interpretation is that he did noteat the other fruits mentioned before. Hence, it could be ar-gued that the condition without a focus operator did notdiffer a lot conceptually from the condition with only. Note,however that the condition without particle and the condi-tion with only still differ theoretically, since only modifiesthe assertion of the sentence, while exhaustivity comesas an implicature in the case of the condition without par-ticle. This theoretical difference is also reflected in the im-proved memory performance with only compared to thecontrol condition.

Second, in the items used, the use of even appearedawkward to some, but not all listeners.5 We carried out arating study in which we presented the target sentences to

5 As pointed out to us by an anonymous reviewer and verified by therating study described next.

a group of naïve listeners and asked them to rate the contex-tual appropriateness of the critical sentence, given the twocontext sentences (a scale from 1 to 7 was used, with 1‘‘completely inappropriate’’ and 7 ‘‘very appropriate’’; forfurther details of the rating study, see Appendix C). The aver-age rating for critical sentences containing only was 6.51, forsentences without focus particle it was 6.44, and for sen-tences containing even it was 5.06. While the ratings for onlyand no particle did not differ significantly, the ratings forsentences containing even were significantly lower than forsentences containing only, t1(29) = �5.49, p < .001;t2(29) = �12.7, p < .001. However, the ratings for sentencescontaining even were still a lot higher than the ratings forfillers containing semantic or pragmatic violations like‘‘Speaker 1: The adult education centre offers classes indrawing, photography, and sewing. I bet, Marius has at-tended courses in photography and sewing. Speaker 2: Yes,he has painted pictures.’’, or ‘‘Speaker 1: On the playground,there are monkey bars, swings, and slides. I bet, Sophie hasused slides and swings. Speaker 2: Yes, she has only usedmonkey bars.’’ The results of a subsequent correlationalanalysis are detailed below (discussion of Experiment 2).

In Experiment 2, we wanted to test whether the effectsgeneralize to contexts where the exhaustiveness implicat-ure is less prominent, where only and even were equallyappropriate contextually, and with stimuli that have a nar-rative structure without a correction.

Experiment 2

Method

ParticipantsA total of 33 native speakers of German (21 female and

12 male, mean age 25.7 years, SD 2.65, age range 21–32)were recruited from the same subject pool and paid 7Euros in compensation. None of them reported any visionor hearing difficulties.

MaterialsWe created a set of 45 experimental items (30 items were

based on the material in Experiment 1 and 15 additionalexperimental items) and 35 filler items (20 items based onExperiment 1 and 15 additional filler items). The narrativestructure differed from Experiment 1 and followed the sche-ma in (3). The first context sentence mentioned a set of threeelements and connected the person with the setting. The sec-ond context sentence continued the story and served to easeaccommodation. Finally, the critical sentence referred backto the list and focused one of the elements. We used thesame particle conditions as in Experiment 1: the exclusiveparticle nur, the inclusive particle sogar and the control con-dition without a focus sensitive particle. All sentences wererecorded by a trained female speaker (NG).

(3)Context sentencesMatthias erhält ein Paket mit Hemden, Hosen und JackenMatthias receives a package with shirts, trousers and jackets.Er guckte, was ihm gefiel.

6

de

K. Spalek et al. / Journal of Memory and Language 70 (2014) 68–84 75

He considered what he liked.

Critical sentence(a) Er hat nur die Hemden behalten(b) Er hat sogar die Hemden behalten(c) Er hat die Hemden behaltenHe kept (a) only/ (b) even/ (c) _ the shirts.

7 Again, we carried out an omnibus analysis which showed a main effect

Care was taken that, across items, the focused elementin the critical sentence was equally often the first, second,or third element from the first context sentence. Again, thecritical sentences with particles were pronounced with ahat contour (Féry, 1993) and the condition without particlehad a falling pitch accent on the focused constituent. Thepitch contour of the critical sentence from example (3) isshown in Appendix D.

As in Experiment 1, we conducted acoustic analyses onthe focused element to compare the pitch contour acrossconditions (see Table 4). For the stressed syllable, none ofthe parameters differed across conditions. On the entireword, only the minimum pitch was significantly higherin the condition with sogar. Fig. 3 displays the pitchcontours on the focused word for the three experimentalconditions. As the Figure shows, the accent type was againthe same across conditions.

The items were constructed according to the same cri-teria as in Experiment 1 (a list of the Material of Experi-ment 2 can be found in Appendix E). Half of theexperimental items contained the definite determiner be-fore the noun and half of them did not, depending on whatsounded natural in the given contexts.6 This time, the filleritems were only recorded in the version without particle andconsequently, while there were still filler questions aboutthe setting, the name of the protagonist, or the action, nomore filler questions were asked concerning the particles(cf. Experiment 1). We omitted filler questions about parti-cles in order to avoid drawing attention to their existencein the stimuli. Hence, in the second experiment, the focusmanipulation was more subtle than in the first. Each partic-ipant received the same filler items with the same questions.The experimental questions mentioned the category and thesetting (e.g., Which pieces of clothes were in the package? forexample (3)), requiring participants to recall the focused ele-ment and the alternatives.

Each participant was exposed to all 45 experimentalitems and 35 filler items in eight blocks of 10 stimuli.The conditions were rotated across the experimental stim-uli according to a Latin Square Design. This resulted in a to-tal of three lists that were pseudo-randomized for eachparticipant. The following constraints were set for random-ization: no more than three filler or experimental trialswere presented in a row, the particular focus condition ap-peared at most three times in a row and, within a block, acategory (e.g., fruit) appeared only once.

ApparatusThe apparatus was the same as in Experiment 1.

Judgments concerning the determiner were checked by three indepen-nt research assistants.

ProcedureThe procedure was the same as in Experiment 1, with

one exception: Instead of 5 blocks in Experiment 1, 8blocks were presented in Experiment 2, and the experi-ment lasted 50 min.

Results

The recorded answers were transcribed and coded inthe same way as in Experiment 1. Again, we first lookedat the correctly recalled focus alternatives. Fig. 4 depictsparticipants’ performance for the alternatives. 59.7% ofthe items were recalled in the control condition, 64.3%with nur, and 64.2% with sogar.

The same mixed effects model as in Experiment 1 wasfitted to the data. A summary of the model is given in Table5. The model yielded a significant effect for sogar (t = 2.1,SE = 2.2, p < .05) and nur (t = 2.1, SE = 2.2, p < .05). A posthoc test revealed no difference between sogar and nur(p = 1.0).

The focused element was correctly recalled 75% of thetime without a focus particle, 77% with nur and 76% withsogar. The recall of the focused element was not affectedby the particle condition (control vs. sogar: p = .59, controlvs. nur: p = .28, nur vs. sogar: p = .85).7

Discussion

Experiment 2 replicates the main results from Experi-ment 1: The presence of a focus particle improved memoryfor focus alternatives. Recall was affected by the presenceof a focus particle, but not by the particular focus particle,supporting the contrast hypothesis and refuting the lexicalmeaning hypothesis.

As with Experiment 1, we compared the contextualappropriateness of the critical sentences in the three focusparticle conditions in a subsequent rating study (see Dis-cussion of Experiment 1 for details). The average ratingswere 6.04 for sentences without a focus particle, 5.64for sentences containing only, and 5.10 for sentencescontaining even. While the discrepancy between sentencescontaining only and even was less strong than for the mate-rials of Experiment 1 (as evidenced in a significant interac-tion of particle condition and experiment, cf. Appendix C),it was still present (t1(29) = �3.45, p < .01; t2(44) = �3.29,p < .01) (but note again that ratings for sogar were againbetter than semantically or pragmatically violated sen-tences, see discussion of Experiment 1 above). Hence, itmight be possible that recall performance was not only af-fected by the presence/absence of a focus particle, but alsoby the contextual appropriateness of the target sentences.In particular, it might be the case that, in line with the lex-ical meaning hypothesis outlined in the introduction ofExperiment 1, alternatives are even more salient (andtherefore recalled better) in the presence of even than in

of focus (p < .0001, focused elements were again recalled better thanalternatives), better recall in the conditions with nur and sogar compared tothe control condition (p < .05 respectively) but no interaction between typeof element and particle condition.

Table 4Mean acoustic parameters of the focused element in the critical sentences (e.g., Hemden in (3a,b,c)).

Measure only no particle even F(2,44) p

Mean SE Mean SE Mean SE

Stressed syllableDuration (s) 0.24 0.01 0.23 0.01 0.23 0.01 1.0 .37Maximum pitch (Hz) 224 12.5 231 15.3 230 15.8 <1 .78Minimum pitch (Hz) 180 3.48 175 3.74 171 3.11 1.83 .17Pitch difference 45 12.7 56 15.1 59 15.9 1.17 .31Mean pitch (Hz) 195 3.87 196 5.05 194 5.07 <1 .69Relative point of pitch maximum (s) 0.10 0.01 0.11 0.01 0.10 0.01 <1 .98Relative point of pitch minimum (s) 0.17 0.02 0.16 0.02 0.15 0.02 <1 .50

Entire wordDuration (s) 0.43 0.01 0.43 0.01 0.42 0.01 1.85 .16Maximum pitch (Hz) 233 14.3 240 16.5 239 17.1 <1 .77Minimum pitch (Hz) 148 3.24 148 3.28 143 3.87 12.17 .0001Pitch difference 85 14.4 92 16.9 95 17.1 1.0 .365Mean pitch (Hz) 183 2.31 184 3.21 181 3.01 <1 .84Relative point of pitch maximum (s) 0.13 0.02 0.12 0.02 0.12 0.02 1.0 .36Relative point of pitch minimum (s) 0.40 0.02 0.38 0.02 0.38 0.02 1.0 .36

1 2 3 4 5

160

180

200

220

Intervall Part

F0 m

ean

Particle Conditionsonly no particle even

Fig. 3. Mean pitch contour of focused element in the critical sentencesacross particle condition.

only no particle even

% re

calle

d50

6070

8090

100

Fig. 4. Percentage correct recall for the alternatives.

Table 5Fixed effect estimates (top) and variance estimates (bottom) for LMER ofcorrectly recalled alternatives (N = 1476, log-likelihood: �7373).

Fixed effect Coefficient SE t p (MCMC)

Intercept 59.70 3.9 15.4Nur 4.63 2.2 2.1 .04Sogar 4.51 2.2 2.1 .05Trial (centered) 0.14 0.06 2.4 .03Random effect s2

Participant 314.02Item 141.00Random Slope s2

Trial 0.06

76 K. Spalek et al. / Journal of Memory and Language 70 (2014) 68–84

the presence of only, because even includes alternativeswhile only excludes them. At the same time, memoryencoding is hampered in the presence of even, reducingthe size of the recall effect, because target sentences witheven are less contextually appropriate.8 If this is the case,we can make a straightforward prediction: If recall of thealternatives is worse for items in which the contextualappropriateness of the target sentence is lower (i.e., foritems that sound ‘‘more awkward’’), we should observe acorrelation between the average appropriateness rating ofan item and the recall performance such that, all else beingequal, recall is better for more appropriate items. We carriedout separate correlations for the materials of Experiment 1and Experiment 2 and separate correlations for the two fo-cus particles, paying particular attention to the conditionwith even, because this is where an appropriateness effectshould emerge. In Experiment 1, the correlation between

8 We thank an anonymous reviewer for suggesting this explanation to us.

the ratings and the recall performance was r = �.09, p = .66for items with even and r = �.43, p < .05 for items with only.In Experiment 2, the correlation was r = �.0004, p = .99 foritems with even and r = .08, p = .62 for items with only.

K. Spalek et al. / Journal of Memory and Language 70 (2014) 68–84 77

Hence, the only significant correlation showed that recallwas worse for items which were contextually more appro-priate, and this was only the case for one set of items andonly in the presence of only. These observations speakagainst a trade-off of the effect of the focus particles andcontextual appropriateness which might have masked anadditional effect of lexical meaning.

Another difference between Experiments 1 and 2 is thedecrease of performance in the two experiments. While75% of the alternatives were recalled on average in Exper-iment 1, this number dropped to 63% in Experiment 2. Per-formance also decreased for the recall of the focusedelement, but less steeply, namely from 82% to 76%. We be-lieve that this is due to two factors. First, Experiment 2 waslonger than Experiment 1 (8 blocks with 10 items each in-stead of 5 blocks with 10 items each), which might havedecreased performance due to both fatigue and also moreinterference from previously recalled information inExperiment 2. Second, while both the focused elementand the alternatives were named twice in Experiment 1,the alternatives were only named once in Experiment 2(whereas the focused element was still named twice).The first factor affects recall of alternatives and the focusedelement equally, while the second factor only affects recallof the alternatives which might explain why performancefor alternatives dropped more than performance for the fo-cused element.

General discussion

In two delayed recall experiments, we presented partic-ipants with auditory discourses in three different versionsand assessed their memory for focus alternatives as a func-tion of the focus particle condition (only, even, no particle).We found that the exclusive particle only as well as theinclusive particle even facilitated recall for the informationstructural alternatives. The prosody of the sentences waskept constant across conditions, so that there was alwaysa falling accent on the focused element. In the following,we will discuss the findings with regard to memory repre-sentations of discourse and with regard to the differencebetween intonation focus and a combination of a focusparticle with intonation focus.

Memory representations

As discussed in the introduction, empirical evidencefor improved memory for focused constituents is rela-tively wide-spread, whereas memory for focus alterna-tives has only been investigated once before. Fraundorfand colleagues (2010) had explicitly pitted three differ-ent accounts against each other, but only two of thesehave anything to say about memory for discourse9:According to the granularity account, based on Sanfordet al. (2006), focus improves the encoding of and hence

9 The third account which is quickly discarded by Fraundorf andcolleagues (2010) assumes that focus (in the form of pitch accent) onlyserves to speed the initial identification of the discourse status.

memory for the focused element. The focused element isencoded more specifically which leads to less activationof related words. According to the contrast account, focusnot only serves to encode the focused information butalso the members of a contrast set at a deeper level.Fraundorf and colleagues’ findings support the latterview.

Unlike Fraundorf et al. (2010), we did not compare (theencoding of) an alternative to a focused element to analternative to an unfocused element. In our case, the alter-natives were always alternatives to the focused element,and the critical manipulation was whether or not a focusparticle was present in addition to the pitch accent onthe focused element. Still, we can say that our data alsosupport the contrast account: The additional manipulationimproved memory for the alternatives, suggesting that it isnot only the focused element itself but the entire contrastset that has become more salient and therefore betterremembered. Given that Fraundorf et al. had been the onlyones so far who reported a beneficial effect for focus alter-natives, it is important to corroborate this finding. Ourstudy has used a more difficult memory task, namely de-layed recall instead of delayed recognition memory. Also,whereas Fraundorf and colleagues had a set of two items(the focused element and the contrast element) that weredirectly contrasted, we had a set of three items (the fo-cused element and two alternatives), and the contrastwas expressed less strongly (in particular in Experiment2). Finding very similar results in spite of these ratherstrong differences suggests that the underlying mechanismis reliable.

Focus particles

Our main theoretical interest was whether focus sen-sitive particles would strengthen the mental representa-tion of focus alternatives above and beyond intonationalfocus marking. Our results support the contrast hypothe-sis formulated in the introduction to Experiment 1: Bothfocus particles improved memory for alternatives relativeto the control condition. Why should this be the case? Ifwe follow Rooth (1992) and Krifka (2007), the primaryfunction of focus is to indicate the presence of interpreta-tion-relevant alternatives. Beaver and Clark (2008) distin-guish types of focus sensitivity that result frompragmatic forces from those that are lexically encoded.They argue that the focus sensitivity of focus particlesis conventionalized (‘‘strong focus’’), that is, lexically en-coded. We can explain the improved memory perfor-mance in the presence of a focus particle in twodifferent ways: Either, the conventionalized focus sensi-tivity of focus particles makes the alternative set moresalient. Or the alternative set is signaled by only onecue without focus particles, namely by the intonation fo-cus, and by two cues in the presence of a focus particle,namely by the intonation focus and the particle. Both ofthese solutions could be implemented in an activation-spreading framework where either a strong activationburst from the focus particle or a combined activationfrom two sources (intonation focus and focus particle)lead to higher activation of the alternatives, ultimately

Fig. A1. Pitch Contour of the critical sentences in Experiment 1, example(2): nur.

Fig. A2. Pitch Contour of the critical sentences in Experiment 1, example(2): sogar.

78 K. Spalek et al. / Journal of Memory and Language 70 (2014) 68–84

improving their encoding in long-term memory. We can-not decide which of these solutions is true, but somedata from our lab on immediate recognition memory(Gotzner, Spalek, & Wartenburger, 2013) show that theeffects of intonation focus and the presence of a focusparticle are not always additive.

Why didn’t we observe a difference between the inclu-sive particle even and the exclusive particle only? Intui-tively, the fact that only expresses that none of thealternatives can take the place of the value in its scopeshould weaken their representation. In contrast, the factthat even licenses the inference that the utterance appliesto at least one alternative and, in addition, that the alterna-tive is more likely than the focused value, should strength-en the representation of the alternatives. A domain inwhich a different impact of only and even can be observedis sentence comprehension: In a reading study, Filik, Pater-son, and Liversedge (2009) investigated sentences whichdescribed a likely event (e.g., students being taught bythe best teacher passing an exam) or an unlikely event(e.g., students being taught by the worst teacher passingan exam). The sentential subjects were preceded by eitheronly or even. The results demonstrated that readers expecta likely event after only (e.g., Only the students taught by thebest teacher passed the exam) and an unlikely event aftereven (e.g., Even the students taught by the worst teacherpassed the exam), and they showed processing difficultiesif this assignment was reversed (i.e., in sentences like Onlythe students taught by the worst teacher passed the exam andEven the students taught by the best teacher passed theexam).

Kim (2012, Chapter 4) compared only and also ratherthan only and even. Like even, also is an inclusive focus par-ticle, but unlike even, it does not impose an order of likeli-hood on the alternative set. Participants were presentedwith little narratives consisting of a context sentence (4)and a target sentence (5):

(4) Mark has some pears and some apples.(5) Jane only/ also has some apples.

Participants were presented with a visual display withfour quadrants on which the set of mentioned items (i.e.,pears and apples), a superset (i.e., pear, apples and or-anges), a subset (i.e., apples) and a novel set (i.e., or-anges) were presented. Participants were instructed toclick on the things Jane has and their eye movementswere monitored while they listened to the sentences.The results show that target sentences containing onlytriggered early fixations to a subset, whereas target sen-tences containing also triggered early fixations to a sup-erset. Kim concludes that only facilitated identificationof a recently mentioned item because listeners expecteda restriction on an alternative set whereas also causedlisteners to expect a discourse-new item from the sameset.

These two studies show that the inclusive or exclusivenature of a focus particle affects immediate expectationsfor the upcoming discourse. By contrast, the long-termmemory representation of an utterance appears to besensitive to the fact that a focus particle requires a set of

alternatives for its interpretation, and this set is subse-quently committed to memory.

Acknowledgments

This work was funded by the Deutsche Forschungs-gemeinschaft (DFG) as part of the Collaborative ResearchCentre SFB 632 ‘‘Information Structure’’. IW was supportedby the Stifterverband für die Deutsche Wissenschaft(Claussen–Simon–Stiftung).

We thank Sophie Repp and Berry Claus for helpfuldiscussions. We are also grateful to Janine Klose andVinzent Müller for recording the experimental stimuliand running the experiments.

Appendix A

Pitch Contour of the critical sentences in Experiment 1,example (2): (a) nur, (b) sogar, (c) no particle

(a) Fig. A1(b) Fig. A2(c) Fig. A3

Fig. A3. Pitch Contour of the critical sentences in Experiment 1, example(2): no particle.

K. Spalek et al. / Journal of Memory and Language 70 (2014) 68–84 79

Appendix B

List of items in Experiment 1.

1. Im Katalog sind Hemden, Hosen und Jacken. Ich wette, Matthias haJacken gekauft.

There are shirts, trousers, and jackets in the catalogue. I bet Matthias h2. In der Obstschüssel liegen Pfirsiche, Kirschen und Bananen. Ich wette

sogar Pfirsiche gegessen.There are peaches, cherries, and bananas in the fruit bowl. I bet Carsten3. Im Getränkemarkt gibt es Wasser, Cola und Saft. Ich wette, Angelika

gekauft.There is water, coke, and juice available at the drinks cash-and-carry.

bought coke.4. Im Zoo leben Zebras, Löwen und Affen. Ich wette, Peter hat Zebras unZebras, lions, and monkeys live in the zoo. I bet Peter has taken pictures5. Im Baumarkt gibt es Pinsel, Sägen und Feilen. Ich wette, Jens hat Pin

nachbestellt.There are brushes, saws, and files in the DIY. I bet Jens reordered brus6. Im Karton liegen Bleistifte, Lineale und Scheren. Ich wette, Sarah hat

Bleistifte weggeschmissen.There are pencils, rulers, and scissors in the box. I bet Sarah has throw

pencils.7. Im Musikzimmer stehen Geigen, Gitarren und Harfen. Ich wette, An

Gitarren gestimmt.There are violins, guitars, and harps in the music room. I bet Anja has8. In der Schatulle befinden sich Ketten, Ringe und Broschen. Ich wette

sogar Ringe angelegt.There are necklaces, rings, and brooches in the casket. I bet Karoline has9. Im Geräteraum liegen Reifen, Matten und Seile. Ich wette, Martin haThere are hoops, mats, and ropes in the gym. I bet Martin has got rop10. Im Schuppen stehen Spaten, Besen und Harken. Ich wette, Doris ha

gesäubert.There are spades, brooms, and rakes in the shed. I bet Doris has cleane11. Im Waffenmuseum befinden sich Dolche, Pistolen und Speere. Ich

nur/sogar Pistolen fotografiert.There are daggers, pistols, and spears in the arms museum. I bet Stefan

pictures of pistols.12. Im Kulturbeutel befinden sich Seife, Shampoo und Duschgel. Ich we

nur/sogar Seife benutzt.There is soap, shampoo, and shower gel in the toilet bag. I bet Michael h13. Im Möbelgeschäft gibt es Tische, Regale und Betten. Ich wette, Anna

sogar Regale angeschaut.

Appendix C

For the rating study, we asked participants to indicatethe contextual appriateness of the third sentence of an itemin the context of the first two sentences. We did a combinedrating for the items of both experiments, adding filler itemswhich contained (mild) semantic or pragmatic violations.

C.1. Method

C.1.1. ParticipantsTwenty-four participants (6 male) from the same popu-

lation as in the experiments reported in the body of the pa-per participated in the rating study. Their mean age was24.5, with a range from 20 to 30. None of them had takenpart in the recall experiments.

C.1.2. ItemsThe experimental items from Experiment 1 (30) and

Experiment 2 (45) were used. Three lists were created suchthat an item appeared only in one condition (with nur,with sogar, control without a particle) within a given listbut that the three conditions appeared equally often on

t sich Hemden und Hosen gekauft. Nein, er hat sich _/nur/sogar

as bought shirts and trousers. No, he _/only/even bought jackets., Carsten hat Kirschen und Bananen gegessen. Nein, er hat _/nur/

has eaten cherries and bananas. No, he _/only/even ate peaches.hat Saft und Wasser gekauft. Nein, sie hat _/nur/sogar Cola

I bet Angelika has bought juice and water. No, she _/only/even

d Löwen fotografiert. Nein, er hat _/nur/sogar Affen fotografiert.of zebras and lions. No, he _/only/even took pictures of monkeys.sel und Feilen nachbestellt. Nein, er hat _/nur/sogar Sägen

hes and files. No, he _/only/even reordered saws.Lineale und Scheren weggeschmissen. Nein, sie hat _/nur/sogar

n away rulers and scissors. No, she _/only/even threw away

ja hat Harfen und Geigen gestimmt. Nein, sie hat _/nur/sogar

tuned harps and violins. No, she _/only/even tuned guitars., Karoline hat Ketten und Broschen angelegt. Nein, sie hat _/nur/

put on necklaces and brooches. No, she _/only/even put on rings.t Seile und Matten geholt. Nein, er hat _/nur/sogar Reifen geholt.es and mats. No, he _/only/even got hoops.t Spaten und Besen gesäubert. Nein, sie hat _/nur/sogar Harken

d spades and brooms. No, she _/only/even cleaned rakes.wette, Stefan hat Dolche und Speere fotografiert. Nein, er hat _/

has taken pictures of daggers and spears. No, he _/only/even took

tte, Michael hat Shampoo und Duschgel benutzt. Nein, er hat _/

as used shampoo and shower gel. No, he _/only/even used soap.hat sich Betten und Tische angeschaut. Nein, sie hat sich _/nur/

(continued on next page)

There are tables, shelves, and beds in the furniture shop. I bet Anna has looked at beds and tables. No, she _/only/even looked at shelves.14. Im Spülbecken sind Schüsseln, Töpfe und Pfannen. Ich wette, Maria hat Töpfe und Schüsseln abgewaschen. Nein, sie hat _/nur/sogar

Pfannen abgewaschen.There are bowls, pots, and pans in the sink. I bet Maria has washed pots and bowls. No, she _/only/even washed pans.15. Im Kinderzimmer befinden sich Murmeln, Kreisel und Bälle. Ich wette, Max hat mit Bällen und Kreiseln gespielt. Nein, er hat _/nur/

sogar mit Murmeln gespielt.There are marbles, spinning tops, and balls in the nursery. I bet Max has played with balls and spinning tops. No, he _/only/even played

with marbles.16. Im Kunstmuseum sind Statuen, Gemälde und Fotografien. Ich wette, Janine hat Fotografien und Statuen betrachtet. Nein, sie hat _/

nur/sogar Gemälde betrachtet.There are statues, paintings, and photographs in the art museum. I bet Janine has looked at photographs and statues. No, she _/only/even

looked at paintings.17. Im Elektrogeschäft gibt es Mikrowellen, Friteusen und Toaster. Ich wette, Florian hat Mikrowellen und Friteusen gekauft. Nein, er hat

_/nur/sogar Toaster gekauft.There are microwaves, chip pans, and toasters in the electric shop. I bet Florian has bought microwaves and chip pans. No, he _/only/

even bought toasters.18. Im Gemüseregal gibt es Paprikas, Gurken und Karotten. Ich wette, Katharina hat Karotten und Paprikas mitgenommen. Nein, sie hat

_/nur/sogar Gurken mitgenommen.There are bell peppers, cucumbers, and carrots at the vegetables section. I bet Katharina has taken carrots and bell peppers. No, she _/

only/even took cucumbers.19. Im Garten wachsen Erbsen, Bohnen und Zwiebeln. Ich wette, Felix hat Bohnen und Zwiebeln geerntet. Nein, er hat _/nur/sogar Erbsen

geerntet.Peas, beans, and onions grow in the garden. I bet Felix has picked beans and onions. No, he _/only/even picked peas.20. In der Dose sind Bonbons, Kekse und Lutscher. Ich wette, Mark hat Lutscher und Bonbons gegessen. Nein, er hat _/nur/sogar Kekse

gegessen.There are candies, cookies, and lollipops in the jar. I bet Mark has eaten lollipops and candies. No, he _/only/even ate cookies.21. Auf dem Blumenbeet wachsen Rosen, Nelken und Lilien. Ich wette, Susanne hat Rosen und Lilien gegossen. Nein, sie hat _/nur/sogar

Nelken gegossen.Roses, carnations, and lilies grow on the bed. I bet Susanne has watered roses and lilies. No, she _/only/even watered carnations.22. Auf der Wiese sind Bienen, Fliegen und Mücken. Ich wette, Karl hat Mücken und Bienen gefangen. Nein, er hat _/nur/sogar Fliegen

gefangen.There are bees, flies, and mosquitos in the meadow. I bet Karl has caught mosquitos and bees. No, he _/only/even caught flies.23. Auf der Einkaufsliste stehen Käse, Eier und Milch. Ich wette, Isabell hat Eier und Milch mitgebracht. Nein, sie hat _/nur/sogar Käse

mitgebracht.There is cheese, eggs, and milk on the shopping list. I bet Isabell has brought eggs and milk. No, she _/only/even brought cheese.24. Auf dem Bauernhof leben Hühner, Ziegen und Kühe. Ich wette, Torsten hat Hühner und Kühe gefüttert. Nein, er hat _/nur/sogar

Ziegen gefüttert.Chicken, goats and cows live at the farm. I bet Torsten has fed chicken and cows. No, he _/only/even fed goats.25. Im Wald leben Füchse, Rehe und Igel. Ich wette, Lisa hat Füchse und Rehe gesehen. Nein, sie hat _/nur/sogar Igel gesehen.Foxes, deer, and hedgehogs live in the woods. I bet Lisa has seen foxes and deer. No, she _/only/even saw hedgehogs.26. Im Märchenbuch geht es um Hexen, Prinzen und Drachen. Ich wette, Simon hat von Prinzen und Drachen geträumt. Nein, er hat _/

nur/sogar von Hexen geträumt.The storybook deals with witches, princes, and dragons. I bet Simon has dreamed of princes and dragons. No, he _/only/even dreamed of

witches.27. Im Wäschekorb liegen Socken, Pullover und Kleider. Ich wette, Sebastian hat Kleider und Socken gewaschen. Nein, er hat _/nur/sogar

Pullover gewaschen.There are socks, sweaters, and dresses in the laundry basket. I bet Sebastian has washed dresses and socks. No, he _/only/even washed

sweaters.28. Im Schuhgeschäft gibt es Stiefel, Sandalen und Turnschuhe. Ich wette, Paula hat Stiefel und Sandalen gekauft. Nein, sie hat _/nur/

sogar Turnschuhe gekauft.There are boots, sandals, and sneakers at the shoe shop. I bet Paula has bought boots and sandals. No, she _/only/even bought sneakers.29. In der Schublade befinden sich Taschen, Schals und Hüte. Ich wette, Julia hat Schals und Hüte herausgenommen. Nein, sie hat _/nur/

sogar Taschen herausgenommen.There are bags, scarves, and hats in the drawer. I bet Julia has taken out scarves and hats. No, she _/only/even took out bags.30. Im Korb liegen Äpfel, Birnen und Pflaumen. Ich wette, Daniel hat Pflaumen und Äpfel herausgenommen. Nein, er hat _/nur/sogar

Birnen herausgenommen.There are apples, pears, and plums in the basket. I bet Daniel has taken out plums and apples. No, he /only/even took out pears.

80 K. Spalek et al. / Journal of Memory and Language 70 (2014) 68–84

each list. Thirty-eight filler items were created that con-tained violations of semantic or pragmatic appropriate-ness, like, for example: Speaker 1: ‘‘The adult educationcentre offers classes in drawing, photography, and sewing.I bet, Marius has attended classes in photography and sew-ing.’’ Speaker 2: ‘‘Yes, he has painted pictures.’’, where thecontinuation is contextually incongruous, or: Speaker 1:‘‘On the playground, there are monkey bars, swings, andslides. I bet, Sophie has used slides and swings.’’ Speaker

2: ‘‘Yes, she has only used monkey bars.’’, where the useof only in combination with the affirmative ‘‘yes’’ is infelic-itous. Each list contained the same filler items, such that agiven list contained 113 items.

C.1.3. ProcedureParticipants listened to the sentences via headphones and

were instructed to indicate the contextual appropriateness ofthe third sentence, given the preceding two sentences, by

Fig. D2. Pitch Contour of the critical sentences in Experiment 2, example(3): sogar.

K. Spalek et al. / Journal of Memory and Language 70 (2014) 68–84 81

pressing a number from 1 (‘‘not at all appropriate’’) to 7 (‘‘veryappropriate’’) on the keyboard. Each participant was pre-sented with one experimental list only.

C.2. Results

A linear mixed effects model with the fixed factorsExperiment (1 vs. 2) and particle condition (nur, sogar, noparticle; nur was chosen as the reference level) and therandom factors participants and items showed a significanteffect of Experiment (p < .001), a significant difference be-tween nur and sogar (p < .001) and a significant interactionbetween Experiment and condition, illustrating that thedifference between nur and sogar was larger in Experiment1 than in Experiment 2. There was no difference betweennur and the condition without a particle (p = .5).

Appendix D

Pitch Contour of the critical sentences in Experiment 2,example (3): (a) nur, (b) sogar, (c) no particle.

(a) Fig. D1(b) Fig. D2(c) Fig. D3

Fig. D1. Pitch Contour of the critical sentences in Experiment 2, example(3): nur.

Fig. D3. Pitch Contour of the critical sentences in Experiment 2, example(3): no particle.

Appendix E

List of items in Experiment 2.

1. Mathias erhält ein Paket mit Jacken, Hosen und Hemden. Er überlegte sich, was ihm am besten stand. Er hat _/nur/sogar die Hemdenbehalten.

Mathias receives a parcel with jackets, trousers and shirts. He wondered what suited him best. He _/only/even kept the shirts.2. Carsten greift in einen Korb mit Pfirsichen, Kirschen und Bananen. Er überlegte sich, auf was er Appetit hatte. Er hat_/nur/sogar die

Pfirsiche herausgeholt.Carsten reaches for a basket full of peaches, cherries, and bananas. He wondered what he would like to eat. He _/only/even took out the peaches.3. Angelika holt aus dem Supermarkt Wasser, Cola und Saft. Sie wollte ihren Durst stillen. Sie hat_/nur/sogar das Wasser kaltgestellt.Angelika bought water, coke, and juice at the supermarket. She wanted to quench her thirst. She _/only/even cooled the water.4. Peter sieht im Zoo Zebras, Löwen und Affen. Er wollte sich später daran erinnern. Er hat_/nur/sogar Affen fotografiert.Peter watches zebras, lions, and monkeys in the zoo. He wanted to remember that. He _/only/even took pictures of monkeys.5. Jens zählt in seinem Baumarkt Pinsel, Sägen und Feilen. Er stellte fest, dass manches fehlt. Er hat _/nur/sogar Sägen nachbestellt.

(continued on next page)

Jens counts brushes, saws and files in his DIY store. He noticed that some tools are missing. He _/only/even reordered saws.6. Sarah ordnet eine Umzugskiste mit Bleistiften, Linealen und Scheren. Sie überprüfte, was nicht mehr zu gebrauchen ist. Sie hat_/nur/

sogar die Bleistifte weggeschmissen.Sarah arranges a packing case full of pencils, rulers, and scissors. She checked what she did not need anymore. She _/only/even threw

away the pencils.7. Anja kommt in ein Musikzimmer mit Geigen, Gitarren und Harfen. Sie möchte ihren Musikunterricht vorbereiten. Sie hat _/nur/sogar

die Gitarren gestimmt.Anja enters a music room with violins, guitars and harps. She wants to prepare her music lesson. She _/only/even tuned the guitars.8. Karoline betrachtet in ihrer Schatulle Ketten, Ringe und Broschen. Sie überlegte, was zu ihrem Outift passt. Sie hat _/nur/sogar die

Ringe herausgenommen.Karoline looks at the necklaces, rings, and brooches in her casket. She wondered what would go well with her outfit. She _/only/even

took out the rings.9. Martin entdeckt im Geräteraum Reifen, Matten und Seile. Er überlegte, welche Übungen er machen wollte. Er hat _/nur/sogar Reifen

herausgeholt.Martin discovers hoops, mats, and ropes in the gym. He decided which exercises he would like to do. He _/only/even got out hoops.10. Doris sieht im Schuppen Spaten, Besen und Harken. Sie überlegte, was sie gebrauchen kann. Sie hat _/nur/sogar Harken

mitgenommen.Doris detects spades, brooms, and rakes in the shed. She wondered what she would need later on. She _/only/even picked up rakes.11. Stefan sieht im Waffenmuseum Dolche, Pistolen und Speere. Er war sehr interessiert. Er hat _/nur/sogar Pistolen fotografiert.Stefan discovers daggers, pistols, and spears in the arms museum. He was very fascinated. He _/only/even took pictures of pistols.12. Michael hat in seinem Kulturbeutel Seife, Shampoo und Duschgel. Er wollte sich waschen. Er hat _/nur/sogar die Seife ausgepackt.Michael has soap, shampoo, and shower gel in his toilet bag. He wanted to get clean. He _/only/even took out the soap.13. Anna betrachtet im Möbelgeschäft Tische, Regale und Betten. Sie überlegte, was in ihre Wohnung passen könnte. Sie hat _/nur/sogar

Regale ausgesucht.Anna looks at tables, shelves, and beds in the furniture shop. She considered what would look nice in her apartment. She _/only/even

chose shelves.14. Maria findet im Spülbecken Schüsseln, Töpfe und Pfannen. Sie überlegte, was sie zum Kochen brauchte. Sie hat _/nur/sogar die

Pfannen abgewaschen.Maria spots bowls, pots, and pans in the sink. She wondered what she would need for cooking. She _/only/even washed the pans.15. Max sucht in seinem Kinderzimmer nach Murmeln, Kreiseln und Bälle. Er konnte nicht alles finden. Er hat _/nur/sogar die Murmeln

verbummelt.Max looks for marbles, spinning tops, and balls in his nursery. He wasn’t able to find everything. He had _/only/even lost the marbles.16. Janine betrachtet in der Ausstellung Statuen, Gemälde und Fotografien. Sie sollte eine Rezension schreiben. Sie hat _/nur/sogar die

Gemälde erwähnt.Janine looks at statues, paintings, and photographs at the exhibiton. She had to write a review. She _/only/even mentioned the paintings.17. Florian testet im Elektrogeschäft Mikrowellen, Friteusen und Toaster. Er überlegte, was er noch gebrauchen kann. Er hat _/nur/sogar

Toaster gekauft.Florian tries out microwaves, chip pans, and toasters in the electric shop. He wondered what he would need. He _/only/even bought

toasters.18. Katharina steht vor einem Gemüseregal mit Paprikas, Gurken und Karotten. Sie überlegte, was sie noch zu Hause hat. Sie hat _/nur/

sogar Gurken mitgenommen.Katharina looks at bell peppers, cucumbers, and carrots in the vegetables section. She considered what she still had at home. She _/only/

even bought cucumbers.19. Felix begutachtet in seinem Garten Erbsen, Bohnen und Zwiebeln. Er pflegte den Garten regelmäßig. Er hat _/nur/sogar die Erbsen

gegossen.Felix examines peas, beans, and onions in his garden. He took care of the garden regularly. He _/only/even watered the peas.20. Mark öffnet eine Dose mit Bonbons, Keksen und Lutschern. Er verspürte Lust auf Süßes. Er hat _/nur/sogar Kekse gegessen.Mark has candies, cookies, and lollipops in the jar. He wanted to eat something sweet. He _/only/even ate cookies.21. Susanne hat auf ihrem Blumenbeet Rosen, Lilien und Nelken. Sie wollte einen Strauß verschenken. Sie hat _/nur/sogar die Nelken

geschnitten.Susannes grows roses, lilies, and carnations on her flower bed. She wanted to give someone a bouquet. She _/only/even used the

carnations.22. Karl jagt auf der Wiese Bienen, Fliegen und Mücken. Er hatte Spaß dabei. Er hat _/nur/sogar Fliegen gefangen.Karl chases bees, flies, and mosquitos in the meadow. He had a lot of fun. He _/only/even caught flies.23. Isabell notiert auf ihrer Einkaufsliste Käse, Eier und Milch. Sie hatte nicht viel Zeit. Sie hat _/nur/sogar den Käse vergessen.Isabell notes cheese, eggs, and milk on her shopping list. She was in a hurry. She _/only/even forgot to buy cheese.24. Torsten züchtet auf seinem Bauernhof Hühner, Ziegen und Kühe. Er überlegte, was er bereits erledigt hat. Er hat _/nur/sogar die

Ziegen gefüttert.Torsten breeds hens, goats, and cows on his farm. He considered what he had already taken care of. He had _/only/even fed the goats.25. Lisa sucht im Wald Füchse, Rehe und Igel. Es war eine lange Wanderung. Sie hat _/nur/sogar Igel gesehen.Lisa looks for foxes, deer, and hedgehogs in the woods. She had a long walk. She _/only/even saw hedgehogs.26. Simon liest im Märchenbuch von Hexen, Prinzen und Drachen. Er las gerne vor dem Einschlafen. Er hat _/nur/sogar von Hexen

geträumt.Simon reads about witches, princes, and dragons in the storybook. He liked reading before going to bed. He _/only/even dreamed of

witches.27. Sebastian holt aus dem Wäschekorb Socken, Pullover und Kleider. Er schaute nach, was besonders dreckig war. Er hat _/nur/sogar die

Pullover eingeweicht.

82 K. Spalek et al. / Journal of Memory and Language 70 (2014) 68–84

Sebastian takes out socks, sweaters, and dresses from the laundry basket. He checked what was notably dirty. He _/only/even soaked thesweaters.

28. Paula betrachtet im Schuhgeschäft Stiefel, Sandalen und Turnschuhe. Sie überprüfte, was sie sich leisten kann. Sie hat _/nur/sogarTurnschuhe anprobiert.

Paula looks at boots, sandals, and sneakers at the shoe shop. She considered what she could afford. She _/only/even tried on sneakers.29. Julia durchsucht eine Schublade nach Taschen, Schals und Hüten. Sie wollte aufräumen. Sie hat _/nur/sogar Taschen aussortiert.Julia browes for bags, scarves, and hats in her drawer. She wanted to tidy up. She _/only/even sorted out bags.30. Daniela nimmt aus dem Kühlschrank Äpfel, Birnen und Pflaumen. Sie möchte backen. Sie hat _/nur/sogar die Birnen abgewaschen.Daniela takes out apples, pears, and plums from the fridge. She wanted to bake a cake. She _only/even rinsed pears.31. Leoni pflanzt auf ihrem Balkon Krokusse, Dahlien und Veilchen. Sie überlegte, was sie noch tun muss. Sie hat _/nur/sogar die Veilchen

gedüngt.Leoni grows crocuses, dahlias, and violets on her balcony. She wondered what she still had to take care of. She _/only/even gave fertilizer

to the violets.32. Falk findet in seinem Modellbaukasten Züge, Boote und Schiffe. Er überlegte, was er am liebsten machen möchte. Er hat _/nur/sogar

Züge zusammengebaut.Falk discovers trains, boats, and ships in his model kit. He wondered what he would like to do. He _/only/even assembled trains.33. Cornelia findet auf dem Sperrmüll Sofas, Stühle und Truhen. Sie wollte ihre Wohnung umgestalten. Sie hat _/nur/sogar Stühle

mitgenommen.Cornelia discovers couches, chairs, and chests in the bulk rubbish. She wanted to rearrange her apartment. She _/only/even took chairs

with her.34. Erik betrachtet im Musikgeschäft Pauken, Flöten und Cellos. Er war auf der Suche nach einem neuen Hobby. Er hat _/nur/sogar Cellos

ausprobiert.Erik looks at kettledrums, flutes, and cellos in the music store. He was looking for a new hobby. He _/only/even tried out cellos.35. Petra legt auf ihren Schreibtisch Füller, Blöcke und Locher. Sie musste etwas vorbereiten. Sie hat _/nur/sogar Füller benutzt.Petra puts pens, blocks, and hole punches on her desk. She had to prepare something. She _/only/even used pens.36. Robert sucht in seiner Werkstatt Zangen, Hämmer und Schrauben. Er suchte eine Weile. Er hat _/nur/sogar die Zangen gefunden.Robert searches for pliers, hammers, and screws in his garage. He searched for a while. He _/only/even found the pliers.37. Tamara lagert in ihrem Tresor Rubine, Perlen und Saphire. Sie benötigte Geld. Sie hat _/nur/sogar die Perlen verkauft.Tamara has rubies, pearls, and sapphires in her vault. She needed some money. She _/only/even sold the pearls.38. Klaus trifft auf der Baustelle Maurer, Maler und Schlosser. Er wollte die Arbeit begutachten. Er hat sich _/nur/sogar mit Schlossern

unterhalten.Klaus meets bricklayer, painters, and locksmiths at the construction site. He wanted to examine the work. He _/only/even talked to

locksmiths.39. Franziska sucht im Badezimmer nach Bürsten, Schwämmen und Lappen. Sie wollte putzen. Sie hat _/nur/sogar die Bürsten gefunden.Franziska looks for brushes, sponges, and rags in the bathroom. She wanted to clean up. She _/only/even found brushes.40. Norman sieht im Biologiebuch Herzen, Mägen und Nieren. Er sollte Zeichnungen anfertigen. Er hat _/nur/sogar Nieren abgezeichnet.Norman sees hearts, stomachs, and kidneys in the biology book. He had to make drawings. He _/only/even copied the kidneys.41. Saskia trifft auf dem Wochenmarkt Bäcker, Gärtner und Bauern. Sie wollte selbst einen Stand aufmachen. Sie hat sich _/nur/sogar mit

den Bauern abgesprochenSaskia meets bakers, gardeners, and farm on the market. She wanted to have her own booth. She _/only/even talked to the farmers.42. Ole arbeitet im Theater mit Tänzern, Sängern und Künstlern. Er plante eine neue Aufführung. Er hat _/nur/sogar Tänzer engagiert.Ole works together with dancers, singers, and artists at the theatre. He was planning a new show. He _/only/even hired dancers.43. Dominik trifft bei der Weltmeisterschaft Ringer, Läufer und Schwimmer. Er wollte eine Reportage drehen. Er hat _/nur/sogar die

Schwimmer interviewt.Dominik meets wrestlers, runners, and swimmers at the Olympic games. He wanted to do a report. He _/only/even interviewed the

swimmers.44. Susanne benötigt für ihr Auto Reifen, Bremsen und Felgen. Sie muss durch den TÜV kommen. Sie hat _/nur/sogar die Bremsen

erneuert.Susanne needs tires, brakes, and wheels for her car. She had to pass the inspection. She _/only/even renewed the brakes.45. Maik sucht auf der Landkarte China, Peru und Indien. Er wollte sehen, wo er schon gewesen ist. Er hat sich _/nur/sogar an Indien

erinnert.Maik looks for China, Peru, and India on the map. He wanted to see where he had already been. He _/only/even remembered India.

K. Spalek et al. / Journal of Memory and Language 70 (2014) 68–84 83

References

Baayen, R. H. (2008). Analyzing linguistic data: A practical introduction usingR. Cambridge: Cambridge University Press.

Beaver, D., & Clark, B. (2008). Sense and sensitivity: How focus determinesmeaning. Oxford: Blackwell.

Birch, S. L., & Garnsey, S. (1995). The effect of focus on memory for wordsin sentences. Journal of Memory & Language, 34, 232–267.

Braun, B., & Tagliapietra, L. (2010). The role of contrastive intonationcontours in the retrieval of contextual alternatives. Language andCognitive Processes, 25, 1024–1043.

Bretz, F., Hothorn, T. & Westfall, P. (2010). Multiple comparisons using R.Boca Raton.

Féry, C. ( (1993). German intonational patterns. Tübingen: Niemeyer.

Filik, R., Paterson, K. B., & Liversedge, S. P. (2009). The influence of only andeven on online semantic interpretation. Psychonomic Bulletin &Review, 16, 678–683.

Fraundorf, S. H., Watson, D. G., & Benjamin, A. S. (2010). RecognitionMemory reveals just how CONTRASTIVE contrastive accenting reallyis. Journal of Memory and Language, 63, 367–386.

Gotzner, N., Spalek, K., & Wartenburger, I. (2013). How pitch accents andfocus particles affect the recognition of contextual alternatives. In M.Knauff, M. Pauen, N. Sebanz, & I. Wachsmuth (Eds.), Proceedings of the35th Annual Conference of the Cognitive Science Society(pp. 2434–2439). Austin, TX: Cognitive Science Society.

Heister, J., Würzner, K.-M., Bubenzer, J., Pohl, E., Hanneforth, T., Geyken,A., et al. (2011). DlexDB – eine lexikalische Datenbank für diepsychologische und linguistische Forschung. PsychologischeRundschau, 62, 10–20.

84 K. Spalek et al. / Journal of Memory and Language 70 (2014) 68–84

Husband, E. M., & Ferreira, F. (2012). Generating contrastive alternatives:Activation and suppression mechanisms. Talk given at the 25th CUNYhuman sentence processing conference, New York, NY.

Jäger, T. F. (2008). Categorical data analysis: Away from ANOVAs(transformation or not) and towards logit mixed models. Journal ofMemory and Language, 59, 434–446.

Johnson-Laird, P. N. (1983). Mental models. Cambridge, MA: HarvardUniversity Press.

Kim, C. (2012). Generating alternatives: Interpreting focus in discourse.Doctoral dissertation, University of Rochester.

Krifka, M. (1999). Additive particles under stress. Proceedings of SALT, 8,111–128.

Krifka, M. (2007). Basic notions of Information Structure. In C. Féry, G.Fanselow, & M. Krifka (Eds.), The notions of information structure.Interdisciplinary Studies on Information Structure (ISIS), 6. Potsdam:Universitätsverlag Potsdam.

König, E. (1991). The meaning of focus particles: A comparative perspective.London: Routledge.

Rooth (1992). A theory of focus interpretation. Natural LanguageSemantics, 1, 75–116.

Silverman, K., Beckman, M., Pitrelli, P., Ostendorf, M., Wightman, C., Price,P., et al. (1992). TOBI: A standard for labeling English prosody. InInternational Conference on Speech and Language Processing (ICSLP), 2, 867–870.

Sanford, A. J. S., Sanford, A. J., Molle, J., & Emmott, C. (2006). Shallowprocessing and attention capture in written and spoken discourse.Discourse Processes, 42, 109–130.

Sturt, P., Sanford, A. J., Stewart, A., & Dawydiak, E. (2004). Linguistic focusand good-enough representations: An application of the change-detection paradigm. Psychonomic Bulletin & Review, 11, 882–888.

Schröder, A., Gemballa, T., Ruppin, S., & Wartenburger, I. (2012). Germannorms for semantic typicality, age of acquisition, and conceptfamiliarity. Behaviour Research Methods, 44, 380–394.

Van Casteren, M., & Davis, M. H. (2006). Mix, a program forpseudorandomization. Behavior Research Methods, 38, 584–594.

van Dijk, T. A., & Kintsch, W. (1983). Strategies of discourse comprehension.New York: Academic Press.