Predicting Upcoming Department Of Experimental Psychology .
C Cambridge University Press 2016 doi:10.1017/S1366728916000547Bilingualism: Language and Cognition: page 1 of 14 Predicting upcominginformation in native-languageand non-native-languageauditory word recognition ASTER DIJKGRAAFDepartment of Experimental Psychology, Ghent UniversityR O B E RT J. H A RT S U I K E RDepartment of Experimental Psychology, Ghent UniversityWO U T E R D U Y C KDepartment of Experimental Psychology, Ghent University(Received: April 7, 2015; final revision received: April 1, 2016; accepted: April 1, 2016)Monolingual listeners continuously predict upcoming information. Here, we tested whether predictive language processingoccurs to the same extent when bilinguals listen to their native language vs. a non-native language. Additionally, we testedwhether bilinguals use prediction to the same extent as monolinguals. Dutch–English bilinguals and English monolingualslistened to constraining and neutral sentences in Dutch (bilinguals only) and in English, and viewed target and distractorpictures on a display while their eye movements were measured. There was a bias of fixations towards the target object in theconstraining condition, relative to the neutral condition, before information from the target word could affect fixations. Thisprediction effect occurred to the same extent in native processing by bilinguals and monolinguals, but also in non-nativeprocessing. This indicates that unbalanced, proficient bilinguals can quickly use semantic information during listening topredict upcoming referents to the same extent in both of their languages.Keywords: bilingualism, prediction, speech perception, visual world paradigmIn monolingual (native) language comprehension, peoplecontinuously generate predictions about upcoming input(e.g., Altmann & Kamide, 1999; Boland, 2005; DeLong,Urbach & Kutas, 2005). In a seminal paper, Altmannand Kamide (1999) studied prediction in auditorylanguage comprehension using a visual world paradigm.Participants listened to sentences such as The boy will eatthe cake or The boy will move the cake. Eye movementswere recorded while participants viewed a visual scenewith four objects that could all be moved, but in whichonly one object (the cake) was edible. When participantsheard the verb eat, participants initiated fixations to thepicture of the cake more often before the onset of theword cake than after hearing the verb move. Altmannand Kamide concluded that the sentence context preactivated the representation of the target word. Variousrecent models of monolingual sentence comprehensionhave now incorporated predictive processing (e.g., Levy,2008; MacDonald, 2013; Pickering & Garrod, 2013).Using context information to generate predictionsis fundamental in efficient language processing: it canspeed up processing, solve ambiguities, and help thelistener determine when to start an overt response in a This research was supported by a Concerted Research Action (GOA)from the Special Research Fund, Ghent University.We thank Denis Drieghe and the Centre for Vision and Cognition(Southampton University) for making it possible for us to run a controlstudy with monolinguals.dialogue (Kutas, DeLong & Smith, 2011; Van Berkum,2010). These facilitatory functions could be particularlyrelevant in L2 comprehension, which is often consideredto be slower, less accurate, and more resource-consumingthan L1 processing (Cook, 1997; Hahne, 2001; Weber& Broersma, 2012). On the other hand, L2 processingdifficulty may also impede efficient prediction duringlanguage comprehension. However, in spite of its possibleincreased importance, there is very little research aboutwhether bilinguals predict input in their L2 as nativespeakers do in L1 or whether L2 words and theirfeatures are just integrated incrementally when they areencountered in the input rather than before.In a recent review, Kaan (2014) suggested thatpredictive processing in L2 is not inherently different frompredictive processing in L1, but that it may be modulatedby factors associated with non-native comprehension. Forexample, it is often assumed that predictions are based onstatistical regularities extracted from the input throughouta person’s lifetime (e.g., Bar, 2007; MacDonald, 2013).However, information stored in memory about how oftena word tends to occur in a certain context (e.g., anedible object following the verb eat) may be differentin L2 and L1 because the L2 has usually been practicedless (Gollan, Montoya, Cera & Sandoval, 2008) and indifferent settings (e.g., native learning versus classroomlearning). Less or different input in L2 may affect thecontent and strength of predictions. Importantly, if L2 ispracticed less than L1, representations of lexical form,Address for correspondence:A. Dijkgraaf, Department of Experimental Psychology, Ghent University, Henri Dunantlaan 2, 9000 Ghent, Belgiumaster.dijkgraaf@ugent.beDownloaded from https://www.cambridge.org/core. Universiteit Gent, on 05 Sep 2017 at 12:04:10, subject to the Cambridge Core terms of use, available at rg/10.1017/S1366728916000547
2Aster Dijkgraaf, Robert J. Hartsuiker and Wouter Duyckmeaning and use as well as the links between themmay be less consistent and less accurate in L2 (Gollanet al., 2008). Weaker representations may lead to lessefficient retrieval. And less efficient retrieval of lexicalform or semantic associations may in turn lead to slower,less accurate or weaker predictions. Likewise, becausebilinguals divide language use between L1 and L2, andtherefore also have less L1 practice, L1 processing toomay be different for monolinguals and bilinguals. Ifinconsistency of lexical representations indeed affectsprediction skill during comprehension, then predictionskill is expected to increase with increased consistencyof representations. This implies that predictive processingin L2 should become more native-like as L2 proficiencyincreases.Furthermore, lexical competition is increased in L2processing because of simultaneous activation of L1words and because L2 speakers often misperceivephonemes, thereby increasing the number of wordsperceived as similar (Lagrou, Hartsuiker & Duyck, 2013a;Weber & Cutler, 2004). Increased competition can causea delay in the selection of a predicted word, as well asin processing the context information used to generate aprediction. Finally, a number of other factors are thought tomodulate prediction in monolingual language processing,such as resource limitations, emotional state and cognitivecontrol. Kaan (2014) suggests that the effect of each ofthese factors may in turn interact with processing language(native or non-native), so that L2 data is required toevaluate the generalizability of each demonstration ofprediction in monolingual language processing.Some studies reveal effects of semantic context ontarget word recognition (Chambers & Cooke, 2009;FitzPatrick & Indefrey, 2007; Lagrou, Hartsuiker &Duyck, 2013b) in L2 processing. However, effects foundat presentation of the target word do not allow usto distinguish facilitation of semantic integration fromsemantic prediction. A constraining sentence contextmay facilitate word integration upon presentation of theword in L2 processing, but whether or not bilingualsactively predict information, online and during sentenceprocessing, to the same extent in L1 and L2, remainsunclear.Prediction in L2 ReadingIn a study in the visual domain by Martin, Thierry,Kuipers, Boutonnet, Foucart and Costa (2013), nativespeakers of English and late Spanish–English bilingualsread sentences in English with predictable or lesspredictable sentence-final nouns. Event-related potentialswere measured at the article preceding the sentencefinal noun. The article was always congruent withthe final noun, but not with the expected noun (e.g.,Since it is raining, it is better to go out with anumbrella [EXPECTED]/ a raincoat [UNEXPECTED]).If participants indeed predicted umbrella, a semanticanomaly effect should be elicited by the article a relativeto an, because a is incongruent with umbrella. Thus,the target for prediction is the lexical form and thecongruent article. The target is predicted based onsemantic information from the sentence context. Martinet al. indeed found an N400-effect for the incongruentarticle for L1 readers, but not for L2 readers. The lack ofan effect on the article was taken to indicate that L2 readersdid not predict the target word (at least not as efficientlyas L1 readers). For the target noun, the authors did find asignificant N400-effect in central and parietal regions inboth L1 and L2 readers, but the effect was significantlylarger in L1 than in L2 readers. The N400-effect on thenoun showed that even though the participants reading inL2 did not predict upcoming input, integration of a targetword in the sentence was still easier if the sentence wasconstraining.The lack of a prediction effect on the article inL2 comprehension in Martin et al.’s study (2013) mayhave resulted from the particular manipulation used. Inparticular, the lexical prediction effect was measuredon the basis of the congruency of an article (a/an)with the predicted word. The particular phonologicalagreement rule manipulated does not exist in the bilingualparticipants’ L1. Martin et al. (2013) tested whethera group of intermediate L2 proficient participants, notparticipating in their experiment, knew the phonologicalarticle-noun agreement rule. Both an online and an offlinetest showed that intermediate L2 proficient participantswere sensitive to the agreement rule. However, theintermediate L2 proficient group actually participating inthe experiment may not have been able to apply the rulequickly enough for a prediction-incongruent determinerto modulate the N400 effect.Therefore, in a second study in the visual domain,Foucart, Martin, Moreno and Costa (2014), used a similarsentence reading paradigm but measured the predictioneffect by manipulating prediction congruency of thedeterminers’ gender in Spanish sentences (e.g., The piratehad the secret map, but he never found the [masc] treasure[EXPECTED]/ the [fem] cave [UNEXPECTED] he waslooking for.). As in Martin et al., the target for predictionwas the lexical form and the congruent article. Thetarget is predicted based on semantic information fromthe sentence context. However, in this study the genderagreement rule between the target article and noun existedboth in the late bilingual participants’ L1 (French) and L2(Spanish). Here, the authors found an effect of congruencyof the article and the predicted noun on the N400 elicitedby the article both in L1 reading (by Spanish monolingualsand early Spanish–Catalan bilinguals) and in L2 reading(by late French–Spanish bilinguals), although the effectlasted for a shorter time in the late bilingual group. TheDownloaded from https://www.cambridge.org/core. Universiteit Gent, on 05 Sep 2017 at 12:04:10, subject to the Cambridge Core terms of use, available at rg/10.1017/S1366728916000547
Prediction in non-native auditory language processingresults demonstrate that bilinguals reading in L2 can usesemantic information from the sentence context to predictupcoming words and their gender. Foucart et al. suggestedthat the similarity between the article-noun agreement rulein late bilingual participants’ L1 and L2 may have madeit easier for the participants to generate a prediction intime. In addition, half of the expected nouns includedin the experiment were cognates, possibly adding to thefacilitatory effect. The two studies described above showthat bilinguals can predict lexical information in sentencereading, but that whether or not prediction occurs maydepend on L1 and L2 language similarity.Prediction in L2 listeningBoth studies described above were conducted in the visualdomain, but predictive language processing may well bemore challenging in the auditory modality. For instance,the fact that auditory input unfolds over time, unlikewritten input, may make prediction more relevant becausethe listener cannot return to prior input or influenceinput rate, unlike in reading1 . Predictive processing mayalso be more difficult in the auditory modality than inthe visual modality for bilinguals because of increasedcross-language co-activation due to misperceptions andmisrepresentation in listening (Weber & Cutler, 2004).Foucart, Ruiz-Tada and Costa (2015) tested predictionin the auditory modality using an EEG paradigm similar toFoucart et al. (2014). Again, the target for prediction wasthe lexical form with the congruent article, and predictionswere based on semantic information from the sentencecontext. The authors found that bilinguals listening in L2are able to predict upcoming words based on sentencecontext. The participants in this study were all bilingualand they were only tested in their L2. Therefore, no directcomparison could be made between the size of the effectin L1 and L2 in bilinguals, or between the size of the effectin monolinguals (L1) and bilinguals (L1 or L2).Visual world paradigm studies on prediction in L2auditory processing have mainly focused on predictionbased on morphosyntactic information. In a visual worldexperiment, Hopp (2013) investigated whether Germannative and English–German bilingual listeners wouldshow predictive looks to target objects whose genderagreed with an article in the auditory signal. Like nativelisteners, English–German bilinguals listening in L2 weremore likely to look at the target objects whose genderagreed with an afore-mentioned article before the onsetof the target object in the auditory signal, but only inthe bilinguals who had native-like mastery of genderassignment.1In Martin et al. (2013), the first half of each stimulus sentence waspresented on the screen as a whole. After pressing spacebar, one wordwas presented every 700 ms.3Hopp (2015) used a visual world paradigm toinvestigate whether English–German bilinguals integratemorphosyntactic information and verb semantics togenerate predictions about upcoming semantic inputduring L2 auditory comprehension. In this experiment,picture displays including three possible actors and acontrol object were paired with an SVO (e.g., TheNOMwolf kills soon theACC deer) or an OVS (e.g., TheACC wolfkills soon theNOM hunter) sentence in German. Nativelisteners were more likely to look at expected patients(e.g., the deer) before the onset of the second NP inSVO sentences and at expected agents (e.g., the hunter)in OVS sentences. English–German bilinguals, on theother hand, were more likely to fixate patients beforethe onset of the second NP, independently of the casemarking (nominative or accusative) of the first NP. Hoppconcluded that there was an effect of semantic predictionin L2 based on information extracted at the verb, but thatcase information did not modulate predictions like in L1listeners. Bilingual participants seemed unable to applyan L2 agreement rule not present in their L1 on the fly, orat least not quickly enough to support prediction. Hopp’sfindings are in line with recent findings of Mitsugi andMacwhinney (2016), who demonstrated that L1 Englishlearners of Japanese with good offline knowledge of theJapanese case-marking system were unable to employ thisknowledge online in order to generate predictions in avisual world eye-tracking experiment.Dussias, Valdés Kroff, Guzzardo Tamargo andGerfen (2013) also focused on prediction based onmorphosyntactic information, specifically, predictionbased on article-noun gender agreement. A group ofEnglish–Spanish bilinguals (high and low proficiency),Italian–Spanish bilinguals and Spanish monolingualssaw a display with two pictures of items with thesame or different grammatical gender. While lookingat the display, they heard a sentence with an articlethat either agreed with the gender of one of the twoitems in the display, or with both. Spanish monolingualslooked at the target picture sooner in the differentgender condition (when the article was a cue) than inthe same gender condition. Highly proficient English–Spanish bilinguals, but not low proficient English–Spanish bilinguals, also looked at the target pictureearlier in the different gender condition. Unlike thelow proficient English–Spanish bilinguals, low proficientItalian–Spanish bilinguals looked at the target picturesignificantly earlier in the different gender condition,but only when the target item was feminine. Dussias etal.’s results suggest that highly proficient bilinguals usegender cues to anticipate information as monolingualsdo, whereas low proficient bilinguals do not, unlesstheir native language has a similar article-noun genderagreement system. Even though the effects Dussias et al.found for monolinguals and highly proficient bilingualsDownloaded from https://www.cambridge.org/core. Universiteit Gent, on 05 Sep 2017 at 12:04:10, subject to the Cambridge Core terms of use, available at rg/10.1017/S1366728916000547
4Aster Dijkgraaf, Robert J. Hartsuiker and Wouter Duyckare likely to be anticipatory in nature, given their timecourse, the authors do not explicitly distinguish betweeneffects anticipation and facilitation of integration.These recent visual world studies on prediction in L2listening reveal that it is especially difficult for bilingualsto process morphosyntactic features quickly enough touse them as a cue to generate predictions in L2. However,it remained unclear whether bilinguals also have difficultyanticipating semantic information in L2 processing,which would always lead to weaker L2 predictioneffects, or, whether they selectively have difficultyapplying language-specific and, difficult, grammaticalrules quickly enough during predictive processing.Hopp (2015) explicitly distinguishes prediction based onverb semantics and prediction based on case-marking.However, as Hopp proposes, the significant effect ofprediction based on verb semantics (predictive looks tothe patient object in both SVO and OVS sentences) inL2 listening can be interpreted in two ways: either theL2 listeners used semantic information extracted at theverb to guide predictive looks towards the most plausiblesentence object in the picture display (the patient), or,on the basis of the first NP, fixations were directed toa plausible patient object, regardless of verb semantics.Therefore, it remains unclear whether bilinguals are ableto use verb semantics to guide their predictions duringnon-native sentence comprehension as they do in L1.Koehne and Crocker (2015) provided evidence thatlanguage learners are able to use semantic restrictionsat the verb to predict upcoming referents. Participantslearned novel, artificial verb, subject (man and woman)and object names by exposure to verbs with visualcontext, followed by exposure to nouns in SVO sentencecontext, in a visual world paradigm. Anticipatory eyemovements to the sentence target objects were foundduring presentation of the constraining verb. As eachverb type was combined with each subject type, theanticipatory eye-movements to the target object could nothave been based on information extracted at the sentencesubject alone. Koehne and Crocker show that people canuse verb semantics to predict upcoming information inearly language learning. However, instruction specificallystressed semantic processing of the sentences. Also, alimited number of artificial verbs (six at most) and objects(18 at most) were used in the study. These two factorsmay have greatly inflated predictive processing whencompared to natural L2 language processing.Present StudyAll previous studies on anticipating information in L2listening have either focused on L2 listening alone, or theyhave compared a group of L2 listeners to a group of L1listeners in a between-participants design. In the presentexperiment, Dutch–English bilinguals were tested in thenative and non-nat
auditory word recognition ASTER DIJKGRAAF Department of Experimental Psychology, Ghent University ROBERT J. HARTSUIKER Department of Experimental Psychology, Ghent University WOUTER DUYCK Department of Experimental Psychology, Ghent University (Received: April 7, 2015; final revision received: April 1, 2016; accepted: April 1, 2016)
5.3.2 Predictors of academic performance.211 5.3.2.1 Academic overload predicting academic performance .211 5.3.2.2 Amotivation predicting academic performance .213 5.3.2.3 Perceived stress directly and indirectly predicting
Keywords: Power analysis, minimum detectable effect size, multilevel experimental, quasi-experimental designs Experimental and quasi-experimental designs are widely applied to evaluate the effects of policy and programs. It is important that such studies be designed to have adequate statistical power to detect meaningful size impacts, if they .
Experimental and quasi - experimental desi gns for generalized causal inference: Wadsworth Cengage learning. Chapter 1 & 14 Campbell, D. T., & Stanley, J. C. (1966). Experimental and quasi -experimental designs for research. Boston: Hougton mifflin Company. Chapter 5. 3 & 4 Experimental Design Basics READINGS
Predicting U.S. Business Cycle Turning Points Using Real-Time Diffusion Indexes Based on a Large Data Set Herman Steklera and Yongchen Zhaob1 a Department of Economics, George Washington University b Department of Economics, Towson University Abstract This paper considers the issue of predicting cyclical turning points using real-time diffusion
Quasi experimental designs are similar to true experimental designs but in quasi experiments, the experimenter lacks the degree of control over the conditions that is possible in a true experiment Some research studies may necessitate the use of quasi designs rather than true experimental designs Faulty experimental design on the .
experimental or quasi-experimental designs. The eval . Principles in Experimental Designs (New York, McGraw Hill, 1962). the details of experimental design, attentionisfocused on
6. Stages in experimental design and scientific methodology 7.Development of experimental design disciplines and their implementation 4. Can provide examples of experimental designs in real cases. 5. Can explain several concepts of treatment design types, environment, and measurement (response). 6. Can explain the stages of experimental design and
Pick up the Baldrige Excellence framework for education today or download free content. Become a Baldrige examiner or attend examiner training for a better understanding of the criteria. Attend a national or regional Baldrige conference. Complete a Baldrige self-assessment. The Baldrige Criteria are built on these interrelated core values and concepts. They represent beliefs and .
