processing strategies and the comprehension of sentence-level input by l2 learners of german

19
Processing strategies and the comprehension of sentence-level input by L2 learners of German Carrie N. Jackson * Department of German, University of Wisconsin-Madison, 818 Van Hise Hall, 1220 Linden Dr., Madison, WI 53706, USA Received 9 October 2007; received in revised form 7 February 2008; accepted 18 February 2008 Abstract This study investigates how L2 learners of German (English L1) process structural and semantic information when reading German sentences. In a timed comprehension task, intermediate and advanced L2 learners of German read sentences that varied according to word order (subject-first versus object-first) and subject animacy (inanimate subject versus animate subject). Both learner groups exhibited higher comprehension rates and faster reading times for sentences with an inani- mate subject, and higher comprehension rates for subject-first compared to object-first sentences. Only among the advanced L2 learners did this subject-first preference translate into significant reading time differences. Comprehension rates for advanced L2 learners were also higher than for intermediate L2 learners in sentences with an animate subject. The results suggest that while both learner groups adopted semantic-based processing strategies, the advanced L2 learners were better able to utilize structural-based strategies to interpret case markings and word order when completing the task. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: German; Syntactic processing; Second language learning; Case; Reading comprehension; Semantic processing 0346-251X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.system.2008.02.003 * Permanent address: Department of Germanic and Slavic Languages and Literatures, Pennsylvania State University, 311 Burrowes Building, University Park, PA 16802, USA. Tel.: +1 814 863 7488; fax: +1 814 863 8882. E-mail address: [email protected] Available online at www.sciencedirect.com System 36 (2008) 388–406 www.elsevier.com/locate/system SYSTEM

Upload: carrie-n-jackson

Post on 29-Oct-2016

214 views

Category:

Documents


0 download

TRANSCRIPT

Available online at www.sciencedirect.com

System 36 (2008) 388–406

www.elsevier.com/locate/system

SYSTEM

Processing strategies and the comprehensionof sentence-level input by L2 learners of German

Carrie N. Jackson *

Department of German, University of Wisconsin-Madison, 818 Van Hise Hall,

1220 Linden Dr., Madison, WI 53706, USA

Received 9 October 2007; received in revised form 7 February 2008; accepted 18 February 2008

Abstract

This study investigates how L2 learners of German (English L1) process structural and semanticinformation when reading German sentences. In a timed comprehension task, intermediate andadvanced L2 learners of German read sentences that varied according to word order (subject-firstversus object-first) and subject animacy (inanimate subject versus animate subject). Both learnergroups exhibited higher comprehension rates and faster reading times for sentences with an inani-mate subject, and higher comprehension rates for subject-first compared to object-first sentences.Only among the advanced L2 learners did this subject-first preference translate into significantreading time differences. Comprehension rates for advanced L2 learners were also higher than forintermediate L2 learners in sentences with an animate subject. The results suggest that while bothlearner groups adopted semantic-based processing strategies, the advanced L2 learners were betterable to utilize structural-based strategies to interpret case markings and word order when completingthe task.� 2008 Elsevier Ltd. All rights reserved.

Keywords: German; Syntactic processing; Second language learning; Case; Reading comprehension; Semanticprocessing

0346-251X/$ - see front matter � 2008 Elsevier Ltd. All rights reserved.

doi:10.1016/j.system.2008.02.003

* Permanent address: Department of Germanic and Slavic Languages and Literatures, Pennsylvania StateUniversity, 311 Burrowes Building, University Park, PA 16802, USA. Tel.: +1 814 863 7488; fax: +1 814 8638882.

E-mail address: [email protected]

C.N. Jackson / System 36 (2008) 388–406 389

1. Introduction

Correctly interpreting argument structure, specifically who does what to whom, is a cru-cial component of language comprehension. In second language (L2) comprehension,learners must develop strategies to effectively identify the grammatical subject and directobject in a sentence when processing L2 input. This task can be difficult when the mostsalient cues for discerning grammatical roles in the L2 differ from those in learners’ firstlanguage (L1) (Jackson, 2007; LoCoco, 1987). Especially at lower proficiency levels, learn-ers may also rely more on semantic-based strategies instead of processing key structuralinformation in a sentence (VanPatten, 1996). Such strategies may be effective when dealingwith natural L2 input, such as in conversation or when reading longer texts, where one hasaccess to a variety of information, including a larger discourse context. However, by over-looking certain L2 grammatical features, learners may hinder their ability to develop themost optimal strategies for processing L2 input and, in turn, incorporate such features intotheir developing L2 linguistic system (Sharwood Smith, 1993; VanPatten, 2004). Using atimed sentence comprehension task, the present study explores how L2 learners of German(English L1) process and comprehend German sentences. By examining both comprehen-sion accuracy and reading time data, this study goes beyond the assessment of L2 sentencecomprehension and offers a window into the strategies L2 learners at different proficiencylevels employ when processing sentences in their non-native language (Harrington, 2001;Papadopoulou, 2005).

The key questions for this study were whether or not L2 learners of German utilizedstructural information provided by a grammatical feature that distinguishes Germanand English, specifically the German case marking system, and whether or not semanticinformation, namely noun animacy, interacted with case marking information when learn-ers read and comprehended German sentences. Unlike English, which relies largely onword order to indicate the subject and direct object of a sentence, German word orderis more flexible and relies largely on case endings to mark these grammatical roles (Lenerz,1977; Zubin, 1977). Although previous research has shown that German native speakersstill have greater difficulty when confronted with object-first sentences compared to sub-ject-first sentences, they still process case marking information effectively (e.g., Bornkesselet al., 2002; Hemforth et al., 1993; Matzke et al., 2002). In contrast, many L2 learners ofGerman do not necessarily recognize the importance of case markings and continue to relypredominantly on semantic information and word order cues when comprehending Ger-man input (Jackson, 2007; Kempe and MacWhinney, 1998; LoCoco, 1987).

2. Processing strategies and L2 acquisition

In his theory of input processing and its relationship to learners’ developing L2 linguisticsystem, VanPatten (1996, 2004) has advanced several hypotheses to describe the strategiesL2 learners use to interpret L2 input. Relying on the idea that learners are limited-capacityprocessors (McLaughlin et al., 1983), VanPatten argues that especially less proficient L2learners will be selective in how they allocate their attentional resources, choosing to pro-cess L2 input ‘‘for meaning before they process it for form” (VanPatten, 2004, p. 7). In sodoing, learners prefer to process content words over grammatical items, such as inflectionalmorphology (VanPatten, 2004, p. 9). Similarly, learners may rely heavily on lexical seman-tics and the likelihood of a given event occurring in the real world (VanPatten, 2004,

390 C.N. Jackson / System 36 (2008) 388–406

pp. 17–19). Furthermore, when learners do utilize grammatical information, they are morelikely to pay attention to structures perceived as ‘‘more meaningful” than those perceived tobe of little communicative value (VanPatten, 2004, p. 10). Finally, learners tend to adopt asubject-first strategy, in which they ‘‘process the first noun or pronoun in a sentence as thesubject or agent” (VanPatten, 2004, p. 15), even when other morphological informationidentifies a different entity as the subject of an utterance. Only after they have successfullyincorporated other grammatical cues, such as verbal agreement or case markings, into theirdeveloping L2 linguistic system, are learners able to attend to such formal features duringcomprehension.

One premise underlying VanPatten’s (1996, 2004) model of input processing isSchmidt’s (1993, 2001) noticing hypothesis. Schmidt (2001) argues that ‘‘SLA is largely dri-ven by what learners pay attention to and notice in the target language input and whatthey understand the significance of noticed input to be” (2001, pp. 3–4). With regard tocase marking cues in German, the question is whether L2 learners of German recognizethe significance of this morphological cue and, in turn, will utilize this information to iden-tify grammatical roles, especially when such information does not play a significant role intheir L1.

In one recent study, Jackson (2007) found evidence among fifth and sixth semester L2learners of German in favor of processing strategies similar to those outlined by VanPat-ten (1996, 2004), in which learners relied predominantly on semantic-based strategies andless on structural-based cues, such as word order or case marking information. The L2learners read individual sentences, such as 1(a–d), which varied according to whetherthe subject of the sentence was an animate or inanimate noun, and word order (subject-first versus object-first).

1a. Peter kann sehen, dass das Spiel denTrainer

argert. (subject-first;inanimate subject)

Peter can see, that [thegame]NOM

[thecoach]ACC

angers

‘‘Peter can see that the game angers the coach.”1b. Peter kann sehen, dass den

Trainerdas Spiel argert. (object-first;

inanimate subject)Peter can see, that [the

coach]ACC

[thegame]NOM

angers

‘‘Peter can see that the game angers the coach.”

1c. Peter kann sehen, dass das Kind denTrainer

argert. (subject-first; animatesubject)

Peter can see, that [thechild]NOM

[thecoach]ACC

angers

‘‘Peter can see that the child angers the coach.”1d. Peter kann sehen, dass den

Trainerdas Kind argert. (object-first; animate

subject)Peter can see, that [the

coach]ACC

[thechild]NOM

angers

‘‘Peter can see that the child angers the coach.”

C.N. Jackson / System 36 (2008) 388–406 391

Learners’ comprehension of sentences containing only one animate noun, such as 1aand 1b, was significantly better than that of sentences containing two animate nouns. Theyhad the greatest difficulty comprehending object-first sentences containing two animatenouns, like 1d. Jackson (2007) suggested that learners’ high performance on sentences con-taining only one animate noun could be traced to a preference for semantic-based strate-gies, in that they could correctly interpret that the game angered the coach rather than thatthe coach angered the game without recourse to grammatical information, like word orderor case markings. When the meaning of individual target words and real world knowledgealone were not sufficient to interpret the target sentence—as was the case in sentences con-taining two animate nouns, in which it was just as likely that the child angered the coach asthe other way around—learners adopted a subject-first strategy. This led to higher com-prehension rates on subject-first sentences containing two animate nouns compared toobject-first sentences containing two animate nouns, in which case markings were the onlyreliable cue for determining the grammatical subject of the target sentence.

In another recent study, Ritterbusch et al. (2006) tested L2 learners of German at a sim-ilar proficiency level as Jackson (2007) on their ability to fill in the correct definite articleand, thus, their ability to correctly identify the grammatical role of each noun via casemarking information. While the L2 learners provided the correct article a majority ofthe time, their accuracy was lower on object-first compared to subject-first sentences. Ina second task, when asked to identify the case if given the definite article and the grammat-ical gender of the noun, the learners were accurate 77% of the time (p. 38). These findingssuggest that even when intermediate-level L2 learners of German possess explicit knowl-edge of the German case system, they may still have difficulty understanding the impor-tance of these case markings in light of German’s more flexible word order.

2.1. L2 sentence processing research

While the results reported by Jackson (2007) and Ritterbusch et al. (2006) support VanP-atten’s (1996, 2004) hypotheses regarding the preferred strategies L2 learners employ wheninterpreting L2 input, these results did not examine what the L2 learners did while process-ing the target sentences, focusing instead on the final outcome, namely their performance ona comprehension or grammar production task. In order to more precisely address the strat-egies L2 learners employ during comprehension, a growing body of literature has begun toadopt research tools common in the field of psycholinguistics to examine sentence process-ing among L2 learners. One research tool commonly used in this type of research, namelyself-paced reading, relies on the collection of reading time data alongside comprehensionrates or grammaticality judgments. Within this research paradigm, reading times areregarded as a measure of the relative processing difficulty of a particular sentence or gram-matical structure, with longer reading times indicating greater processing difficulties.

Recent studies utilizing this research paradigm have shown that even highly proficientL2 learners may rely more on lexical-semantic and pragmatic information when processingand comprehending L2 sentences compared to L1 speakers (see Clahsen and Felser, 2006,for a recent review). At the same time, there is evidence that at higher proficiency levels, L2learners are able to process crucial grammatical information in the L2, even when thegrammatical structure in question is not part of their L1 grammar (e.g., Hoover and Dwiv-edi, 1998; Hopp, 2006; Juffs and Harrington, 1995; Williams et al., 2001, but see Jiang,2004, 2007 for counter evidence).

392 C.N. Jackson / System 36 (2008) 388–406

To date, however, most of the research in the area of L2 processing literature hasfocused on highly proficient L2 speakers (see Dekydtspotter and Outcalt, 2005; Frenck-Mestre, 1997; Osterhout et al., 2004; Tokowicz and MacWhinney, 2005, for exceptionsto this trend). To explore the potential connections between L2 learners’ processing strat-egies and their developing L2 linguistic knowledge, researchers must also look at less pro-ficient L2 learners. To this end, using the same target sentences as Jackson (2007), thepresent study investigated the relationship between comprehension and reading timeswhen L2 learners of German read sentences like 1(a–d), and whether strategy use, as mea-sured by reading times, changed with increased proficiency.

2.2. Competition model research

One area of research that has directly addressed the relationship between L2 proficiencyand processing strategies, focusing in particular on how L2 learners assign the grammat-ical roles of agent and patient in an utterance, is research within the framework of theCompetition Model. Within this research paradigm, the cue or cues that are the most reli-able predictors of the agent in an utterance are presumed to have a high level of cue valid-

ity, and are therefore more important in that particular language (MacWhinney et al.,1984). In English, word order is the most reliable cue for determining the subject of thesentence, whereas in German, case markings are the most salient cue (Bates and Mac-Whinney, 1989; Kilborn, 1989). Research has shown that especially less proficient L2learners tend to rely on the most salient cues in their L1 when determining agency inL2 sentences. For example, LoCoco (1987) found that beginning L2 learners of German(English L1) transferred L1 English word order strategies and overlooked L2 German casemarkings when listening to sentences and determining the agent. Nevertheless, as profi-ciency increases, L2 learners appear to become more attuned to the relative strength ofcues in the L2, although this trend toward increasing native-likeness is rarely complete(Kilborn, 1989; McDonald, 1987; Sasaki, 1994).

Relevant for the current study, Kempe and MacWhinney (1998) showed that wheninterpreting sentences, such as examples 2(a–d), L2 learners of German tended to pickthe first noun as the agent of the utterance, regardless of the case markings.

2a. Die Mutter

sucht die Tochter. [The mother]NOM/ACC looks-for [the daughter]NOM/ACC

‘‘The mother looks for the daughter.”‘‘The daughter looks for the mother.”

2b. Den Loffel

sucht die Mutter. [The spoon]ACC looks-for [the mother]NOM

‘‘The mother looks for the spoon.”

2c. Der Vater

sucht den Teller. [The father]NOM looks-for [the plate]ACC

‘‘The father looks for the plate.”

2d. Der Loffel

sucht die Tochter. [The spoon]NOM looks-for [the daughter]ACC

‘‘The spoon looks for the daughter.”

At the same time, faster response latencies were recorded when the first noun was ani-

mate, and therefore a logical agent or grammatical subject of the utterance, as well aswhen the second noun was inanimate, and therefore a logical patient or direct object.These findings led Kempe and MacWhinney (1998) to conclude that ‘‘learners of Germanconsider semantic information immediately and regardless of whether an unambiguouscase marker is present or not” (1998, p. 567). Studies utilizing the Competition Model par-adigm with other language pairings have also reported that animacy is a salient cue for L2learners, regardless of the relative importance of this cue in the learners’ L1 (e.g., Gass,1987; Sasaki, 1994).

C.N. Jackson / System 36 (2008) 388–406 393

2.3. Present study

To explore the relationship between the use of semantic- and structural-based strategiesamong L2 learners of German, the present study manipulated target sentences accordingto two factors; whether they contained one animate noun or two animate nouns (inani-mate subject versus animate subject) and word order (subject-first versus object-first).All target sentences were unambiguous if one paid attention to case marking information.In this manner, the present study also measured the extent to which intermediate andadvanced L2 learners of German would utilize this structural cue when reading the targetsentences.

In line with previous research (Jackson, 2007; Ritterbusch et al., 2006), I hypothesizedthat the L2 learners of German in the present study would have difficulty utilizing casemarking information to assign grammatical roles when reading the target sentences, whichwould manifest itself through lower comprehension rates on object-first sentences. Withregard to the reading time data, if the L2 learners had not yet reached a proficiency levelwhere they could devote sufficient attentional resources to processing case marking infor-mation while reading the target sentences, then their reading times would exhibit few if anydifferences according to word order. In contrast, if they did process this information whilereading the target sentences, then they should exhibit greater processing difficulties—asindicated by longer reading times—on object-first sentences, similar to previous findingsin the L1 German processing literature (e.g., Hemforth et al., 1993). For both the compre-hension accuracy and reading time data, I expected that these difficulties would be exac-erbated in object-first sentences containing two animate entities since case markingswould be potentially the only means of correctly interpreting grammatical roles in this tar-get sentence condition.

At the same time, if learners relied on semantic-based strategies, then the difficultiesposed by object-first sentences should be attenuated in conditions containing only one ani-mate entity, as one could correctly interpret grammatical roles based on semantic informa-tion and real world knowledge, without necessarily processing structural cues in the input(VanPatten, 1996, 2004). This would be reflected in comparable comprehension rates andreading times for sentences containing only one animate entity, regardless of word order.Finally, I hypothesized that as proficiency increased, the L2 learners would become betterable to utilize structural information when processing the target sentences, as evidenced byhigher comprehension rates on sentences containing two animate entities and greater dif-ferences in reading times according to word order among the advanced L2 learners ascompared to the intermediate L2 learners.

394 C.N. Jackson / System 36 (2008) 388–406

3. Method

3.1. Participants

Seventeen intermediate L2 learners of German enrolled in intermediate German lan-guage courses at a large Midwestern university and 26 advanced L2 learners of Germanenrolled in advanced literature, linguistics, and conversation courses at the same universityparticipated in the study. All participants were native speakers of English who beganlearning German after age 11.1 Furthermore, all participants demonstrated knowledgeof the target sentence vocabulary, as measured by a vocabulary translation task in whichparticipants provided English translations of the German words used in the timed readingcomprehension task.2 This minimized the possibility that poor vocabulary knowledgewould unduly influence performance on the task.

The participants also completed a language background questionnaire to gather infor-mation about their language learning experiences. Participants were asked to self-rate theiroverall proficiency in German, along with their reading, writing, speaking, and oral com-prehension skills on a four-point scale: 1 = beginner, 2 = intermediate, 3 = advanced, and4 = near-native. Participants also listed how many years they had been learning German,and whether or not they had spent time in a German-speaking country. The results fromthis questionnaire are presented in Table 1. T-tests were performed to compare the twolearner groups to one another on each of these measures.

Compared to the intermediate L2 learners, the advanced L2 learners had been learningGerman longer, had spent more time on average in a German-speaking country, and ratedthemselves as significantly more proficient in German, both in overall proficiency and withregard to oral proficiency and their ability to comprehend spoken German (all p-values < .01).Although the advanced L2 learners also rated themselves as more proficient in reading andwriting in German than the intermediate L2 learners, these differences were not statisticallysignificant (both p-values > .1). Table 1 also presents the overall comprehension results,including filler items, from the timed reading comprehension task. Here too, the advancedL2 learner group outperformed the intermediate L2 learner group (t(1,41) = 2.32, p < .05).

Twenty-five German native speakers also completed the timed comprehension task.Their data provide a baseline measure of performance on the task. However, becausethe main questions in this study investigate how L2 learners process the target sentences,the native German data is provided for descriptive comparisons only. Their results are notincluded in any of the statistical analyses (Fender, 2003; Juffs, 2005).

4. Materials

All target sentences consisted of a main clause and a subordinate clause. The subordi-nate clause contained the crucial information in each target sentence. Because German

1 Four participants (3 intermediate; 1 advanced) indicated on their language background questionnaire thatthey had knowledge of Russian and two intermediate participants also indicated knowledge of Latin. All of theseparticipants, however, still considered German their dominant and most proficient L2. No other participantsreported knowledge of any L2 that uses case markings to mark grammatical roles. Furthermore, no participantsreported any exposure to a foreign language prior to age 11.

2 This task was administered after the completion of the timed reading comprehension task.

Table 1Biographical information for L2 learner groups

Intermediate L2 learners (n = 17) Advanced L2 learners (n = 26)

Self-ratings of L2 proficiencya

Reading 2.5 2.7Writing 2.4 2.7Speaking 2.3 2.9Comprehension 2.6 3.3Overall 2.4 2.8

Total years learning German 4.9 6.8Time abroad (in months) 0.2 6.5Overall comprehension task scoreb 83.1 88.1

a Self-proficiency ratings are on a scale of 1–4, with 1 = beginner and 4 = near-native.b Overall comprehension score is out of 100%.

C.N. Jackson / System 36 (2008) 388–406 395

requires that the finite verb appear in the final position in subordinate clauses, the thematicverb in the crucial portion of each target sentence appeared after the two nominal argu-ments. All introductory clauses were designed to be as neutral as possible so as not toinfluence the meaning of the targeted subordinate clause. Furthermore, the words usedin the target sentences were all found in a selection of first year German textbooks to maxi-mize the participants’ familiarity with the target sentence vocabulary.

Within the key portion of each sentence, the direct object was always an animate entity.Sentences varied according to whether the subject was animate or inanimate, and wordorder, namely subject-first order versus object-first order.3 The combination of thesetwo variables led to four variations for each target sentence, shown previously as sentences1(a–d).

German is a language that contains three grammatical genders, masculine, feminine,and neuter. Due to case syncretism, only masculine nouns are unambiguous in the nom-inative and accusative case, the two cases used to indicate the grammatical subject andmost direct objects in a sentence. In contrast, the definite articles die and das identify bothnominative and accusative forms for feminine and neuter nouns, respectively. Therefore,nouns from all three genders appeared as grammatical subjects in the target sentences, butall direct objects were masculine, to ensure that all target sentences were unambiguouswhen read in their entirety.

In total, 24 target sentences were created (see Appendix for complete list of target sen-tences). Each sentence was manipulated to create four different versions, as seen in thesample sentences 3(a–d). From these sentences, four different lists were created, each con-taining six different sentences per condition, such that no list contained more than one ver-sion of any given target sentence. Participants received one of two lists, thus ensuring that

3 Although German exhibits an overall subject-first preference, animacy can impact this preference. Specifically,object-first sentences containing an animate direct object and an inanimate subject are considered lessunacceptable and easier to process by German native speakers (Jackson, 2005; Scheepers et al., 2000; Zubin andKopcke, 1985). In contrast, sentences containing an animate subject exhibit a subject-first preference, regardlessof the animacy of the direct object. Therefore, subject animacy was varied in the target sentences, as opposed toobject animacy, so that object-first sentences containing an inanimate subject and an animate object would not beas pragmatically marked.

396 C.N. Jackson / System 36 (2008) 388–406

they did not read the same sentence more than once.4 So that participants would be lesslikely to discover the purpose of the experiment, the 24 target sentences were presentedin randomized order along with 52 filler items.

5. Procedure

All participants completed a timed reading comprehension task containing the targetand filler sentences. The experiment was designed and implemented using Micro Experi-mental Laboratory (MEL) version 2.01 (Schneider, 1997). Participants read each targetsentence individually, in its entirety, on a computer screen. They were instructed to readeach sentence as quickly and accurately as possible. Once they had read the sentence, theypressed the space bar. At this point, the target sentence disappeared and two statementsappeared. Participants were prompted to choose the statement that best captured themeaning of the original target sentence. These statements focused on aspects crucial tocomprehending the target sentences, including grammatical role assignment and the mean-ing of the thematic verb.

Current pedagogical trends favor testing L2 reading comprehension in the L1 (Lee andVanPatten, 2003). Therefore, the comprehension statements were presented to the L2learners in English.5 When tested via the L2, learners might have relied on memorizingthe target sentence and responding based on whether the comprehension statement repli-cated the form of the original target sentence. By presenting the comprehension statementsin English, L2 learners were forced to interpret the original German target sentence andprocess it for meaning. Presenting the comprehension statements in English also elimi-nated the possibility that the L2 learners would not understand the summary statements,which could have potentially confounded the comprehension results. In this manner, read-ing times for each target sentence were recorded, along with whether participants chose thecorrect comprehension statement.

6. Results

6.1. Comprehension accuracy

All participants correctly comprehended at least 65% of the target sentences correctly,indicating that they were paying attention while completing the task. The comprehensionresults for the German native speakers and both L2 learner groups are presented in Table2. Again, the native speaker results are presented for descriptive purposes only and werenot included in the statistical analysis. The data suggest that both L2 learner groups per-formed well on sentences containing only one animate noun, regardless of word order. Forsentences containing two animate nouns, however, it appears that the advanced L2 learn-ers generally outperformed the intermediate L2 learners, although both groups had diffi-culty comprehending object-first sentences.

4 Participants only received two of the possible four target sentence lists for the timed reading task. The othertwo lists were used in a separate acceptability judgment task not reported here (see Jackson, 2005 for details).

5 German translations of the comprehension statements were used in the version of the task completed by theGerman native speaker control group.

Table 2Mean percent accuracy for L2 learner groups and German native speakers (standard deviations in parentheses)

Intermediate L2 learners Advanced L2 learners German native speakers

Inanimate subject/subject-first 97.1 (8.8) 98.7 (4.5) 98.0 (5.5)Inanimate subject/object-first 95.1 (7.8) 94.2 (12.4) 94.0 (9.5)Animate subject/subject-first 77.5 (15.5) 87.8 (13.0) 90.7 (12.8)Animate subject/object-first 57.8 (20.5) 67.3 (23.8) 88.7 (15.0)

C.N. Jackson / System 36 (2008) 388–406 397

A 2 � 2 � 2 repeated-measures ANOVA, with animacy (inanimate subject versus ani-mate subject) and word order (subject-first versus object-first) as within-subjects factors,and proficiency as a between-subjects factor (intermediate versus advanced), was per-formed on the percentage of correct responses for each condition. The results revealed amain effect of animacy (F(1, 41) = 102.09, p < .0001), indicating that overall, performancewas better on sentences containing only one animate noun compared to those containingtwo animate nouns. At the same time there was also a significant animacy � proficiencyinteraction (F(1,41) = 4.13, p < .05). A follow-up ANOVA revealed that there was no sig-nificant difference in the percentage of correct responses between intermediate andadvanced L2 learners on sentences containing only one animate noun (F < 1.0). In con-trast, the percentage of correct responses on sentences containing two animate nounswas significantly higher among advanced L2 learners (M = 77.6%) than among intermedi-ate L2 learners (M = 67.6%; F(1, 42) = 4.22, p < .05).

There was also a main effect for word order (F(1, 41) = 32.25, p < .0001) but no signif-icant word order � proficiency interaction (F < 1.0), indicating that overall, both interme-diate and advanced L2 learners appeared to have greater difficulty comprehending object-first sentences compared to subject-first sentences. Finally, there was a significant anima-cy � word order interaction effect (F(1, 41) = 26.10, p < .0001). Simple effects tests showedthat for both groups, although performance on object-first sentences was significantlylower than performance on subject-first sentences regardless of animacy, this differencewas much greater on sentences containing two animate nouns than on sentences contain-ing only one animate noun (animate subject: F(1, 42) = 39.35, p < .0001; inanimate sub-ject: F(1, 42) = 4.62, p < .05). Although there was a significant trend for proficiency(F(1, 41) = 3.07, p < .1), there was no significant three-way interaction between animacy,word order, and proficiency (F < 1.0), indicating that in spite of the differences betweenthe two groups on sentences containing two animate nouns, the overall pattern of resultswas similar for both groups.

7. Reading times

As is common in sentence processing research, only reading times from correctly compre-hended sentences were included in the analysis (Fender, 2003). This criterion excluded18.1% of the data from the intermediate L2 learner group and 13.0% of the data from theadvanced L2 learner group. Reading times that were longer than 20000 ms and shorter than1000 ms were considered outliers and removed from further analysis. Furthermore, readingtimes that were greater than three SDs from the mean for a given condition across all par-ticipants within each L2 learner group were also considered outliers and removed from fur-ther analysis. These steps were taken to eliminate reading times that were artificially high or

Table 3Mean reading times (ms) for L2 learner groups and German native speakers (standard deviations in parentheses)

Intermediate L2 learners Advanced L2 learners German native speakers

Inanimate subject/subject-first 6982 (1742) 6578 (1456) 4500 (1692)Inanimate subject/object-first 6631 (1587) 6909 (1807) 4703 (1350)Animate subject/subject-first 7219 (2054) 6870 (1696) 4917 (1468)Animate subject/object-first 7103 (2247) 7797 (2658) 6493 (2700)

398 C.N. Jackson / System 36 (2008) 388–406

low as a result of factors, such as momentary loss of concentration on the part of an indi-vidual participant, that were not directly related to the experiment itself. This resulted in theexclusion of an additional 3.2% of the data for the intermediate L2 learner group and 1.6%of the data for the advanced L2 learner group.

Mean reading times for the German native speakers and each L2 learner group are pre-sented in Table 3. These data suggest that while reading times in both learner groups weresensitive to whether a sentence contained one animate noun or two animate nouns, onlyreading times within the advanced L2 learner group varied according to word order.

A repeated-measures ANOVA with animacy (inanimate subject versus animate subject)and word order (subject-first versus object-first) as within-subjects factors and proficiency(intermediate versus advanced) as a between-subjects factor was performed on the meanreading times for each condition. There was a main effect for animacy (F(1, 41) = 5.83,p < .05) and no significant animacy � word order or animacy � proficiency interaction(animacy � word order: F(1, 41) = 1.32, p > .1; animacy � proficiency: F < 1.0) indicatingthat overall, sentences containing only one animate noun were read more quickly thanthose containing two animate nouns.

There was no main effect for word order (F < 1.0), however there was a significant wordorder � proficiency interaction (F(1, 41) = 4.70, p < .05). A follow-up ANOVA indicatedthat while there was no significant difference in reading times on object-first sentences(M = 7100 ms) compared to subject-first sentences (M = 6867 ms) among the intermediateL2 learners (F < 1.0), advanced L2 learners took significantly longer to read object-firstsentences (M = 7353 ms) compared to subject-first sentences (M = 6724 ms; F(1, 25) =5.19, p < .05).

There was no three-way interaction effect between animacy � word order � proficiency(F < 1.0) nor was there an overall effect for proficiency (F < 1.0). This indicates that therewas no difference in overall reading times between the intermediate and advanced partic-ipants. When viewed in light of the significant word order � proficiency interaction, theseresults suggest that the main difference between the intermediate and advanced L2 learnergroup stemmed from the attention they paid to structural information when comprehend-ing the target sentences, not in their overall speed on the task, nor in their ability to utilizesemantic information during comprehension.

8. Discussion

By and large, the results from both L2 learner groups parallel the comprehensionresults from intermediate-level L2 learners of German reported by Jackson (2007). Boththe intermediate and advanced L2 learners appeared to have fewer difficulties on sentencescontaining an inanimate subject, with higher comprehension rates and faster reading times

C.N. Jackson / System 36 (2008) 388–406 399

for these sentences compared to sentences containing both an animate subject and an ani-mate direct object. Similarly, both L2 learner groups had higher comprehension rates forsubject-first sentences compared to object-first sentences, although this difference wasmuch greater for sentences containing two animate nouns. Only among the advancedL2 learners, however, did this subject-first preference lead to significant reading time dif-ferences between subject-first and object-first sentences. Finally, although there was no sig-nificant difference in comprehension rates for advanced versus intermediate L2 learners onsentences containing only one animate noun, advanced L2 learners performed better onsentences containing two animate nouns.

The first hypothesis that the L2 learners would have greater difficulty processing casemarking information was supported based on both groups’ lower comprehension rateson object-first sentences. This result diverges from the comprehension results of the Ger-man native speakers, presented in Table 2, who exhibited relatively high comprehensionrates regardless of word order. This indicates that the German native speakers in this studysuccessfully utilized case marking information to identify grammatical roles in a sentence(cf. Bates and MacWhinney, 1989; Lenerz, 1977; Zubin, 1977). In contrast, the compre-hension results from the intermediate and advanced L2 learners suggest that they adopteda subject-first strategy, in which all else being equal, they preferred to interpret the firstnoun they encountered as the grammatical subject. As such, these results corroborate pre-vious research showing that L2 learners of German have difficulty correctly identifyingcase marking information in object-first sentences (Ritterbusch et al., 2006).

Interestingly, while both learner groups had lower comprehension rates on object-firstsentences, only among the advanced L2 learners did this difficulty lead to significant read-ing time differences according to word order. Given that case marking information is whatsignaled whether a sentence was a subject-first or an object-first sentence, this reading timepreference for subject-first sentences among the advanced L2 learner group suggests thateven though their comprehension of object-first sentences was significantly lower, theywere sensitive to case marking information when reading and interpreting the target sen-tences. This subject-first preference in the advanced L2 learner group parallels findings inthe German psycholinguistic literature, in which German native speakers have greater dif-ficulties reading object-first sentences, even in situations where accompanying comprehen-sion tasks have shown relatively low error rates regardless of word order (e.g., Bornkesselet al., 2002; Hemforth et al., 1993). Thus, while the results from both L2 learner groupsindicate that word order had an impact on their ultimate comprehension of the target sen-tences, only among the advanced L2 learners did this preference appear to translate into aprocessing strategy favoring subject-first sentences while reading the target sentences.When viewed in combination with the advanced L2 learners’ higher comprehension of sen-tences containing two animate nouns overall, the reading time differences between the twoL2 learner groups suggest that the advanced learners had incorporated the German casemarking system into their L2 linguistic system and had begun to recognize the meaningfulrole this morphological cue plays in German (Schmidt, 1993, 2001; VanPatten, 2004).

While there was evidence that the advanced L2 learners had begun to recognize theimportant role case marking information plays in German, the fact remains that compre-hension rates for the object-first sentences containing two animate entities were still rela-tively low among both L2 learner groups. This finding highlights the need to develop anawareness among L2 learners of German at all proficiency levels that because word orderin German is more flexible compared to English, the German case marking system is not

400 C.N. Jackson / System 36 (2008) 388–406

simply an abstract set of rules, but rather a central and meaningful component of Germangrammar. By explicitly addressing this fact and discussing the pragmatic reasons that canfavor object-first sentences (cf. Lenerz, 1977), and including activities that require learnersto utilize case marking information when deducing the meaning of an utterance, teachersmight be able to encourage even less proficient L2 learners to develop more efficient strat-egies for processing German input (cf. Wong, 2004). Research that directly tests the effec-tiveness of such instructional methods would be a valuable path for future studies.

The second hypothesis, that the availability and reliability of semantic informationwould minimize the difficulty of object-first sentences was also supported. On the onehand, the higher comprehension rates and faster reading times on sentences containingonly one animate noun parallel the results reported by Kempe and MacWhinney(1998), who showed that response latencies among L2 learners of German were sensitiveto the manipulation of noun animacy. Based on their findings, Kempe and MacWhinney(1998) concluded that noun animacy aids comprehension because animate nouns are likelyagents or subjects in a sentence, whereas inanimate nouns are likely patients or directobjects. However, in the present study, the direct object was always animate and the ani-macy of the grammatical subject was manipulated. Therefore, it is possible that the anima-cy effects found here and in previous L2 research may not derive from the relative ease ofassigning agent and patient roles as such, but rather how easy it is to interpret a sentencebased on the meaning of individual words and how those individual words relate to oneanother based on real world event probabilities.

If one assumes that reading times provide an indication of the processing strategies theL2 learners used when completing this task (Harrington, 2001; Papadopoulou, 2005), thenthe results also offer evidence to support VanPatten’s (1996, 2004) hypothesis that whenpossible, L2 learners will adopt processing strategies that rely less on structural informa-tion and more on lexical semantics and event probabilities. Specifically, when L2 partici-pants at both proficiency levels could utilize a strategy that relied on the application ofsemantic information and the likelihood of a particular event occurring in the real world,they were able to process and interpret the target sentence more quickly than when theyalso had to pay attention to structural information, be it word order or case markings,to comprehend the target sentence. Furthermore, that the advanced L2 learner group alsoexhibited a reading time difference according to animacy, independent of word order, sug-gests that even when L2 learners’ knowledge of certain L2 grammatical structuresimproves, they may continue to rely heavily on semantic-based strategies when processingL2 input (Clahsen and Felser, 2006). This highlights the need for language teachers to rec-ognize that even highly successful and proficient L2 learners may not process informationin the same manner as native speakers of a language.

9. Conclusion

The current study set out to test the validity of VanPatten’s (1996, 2004) processingprinciples among intermediate and advanced L2 learners of German. By analyzing bothcomprehension and reading time data, this study attempted to measure not only the finaloutcome of L2 processing, but also examine the strategies L2 learners use when processingand comprehending L2 sentences. The results suggest that semantic information and realworld event probabilities played a crucial role in the strategies the L2 learners used to com-prehend the target sentences. At the same time, with increased proficiency learners became

C.N. Jackson / System 36 (2008) 388–406 401

better able to take advantage of structural information when completing the experimentaltask. The results from this study demonstrate how psycholinguistic research tools can beused to explore the development of L2 learners’ processing strategies as L2 proficiencyincreases. As such, this line of research augments classroom-based pedagogical researchand furthers not only our understanding of L2 learners’ developing L2 linguistic system,but points to ways foreign language teachers may foster L2 learning at all levels.

Acknowledgement

This project was funded in part by a Dissertation Research Grant from Language

Learning—A Journal of Research in Language Studies, and a Travel Grant from the Centerfor German and European Studies at the University of Wisconsin-Madison. I would liketo thank Mark Louden, Monika Chavez, Maryellen MacDonald and the anonymous Sys-tem reviewers for their comments on earlier versions of this article. All errors, of course,are my own.

Appendix

1b. Peter kann sehen, dass das Spiel den Trainer argert. (SO/-ANIM)Peter can see that the game angers the coach.

1a. Peter kann sehen, dass den Trainer das Spiel argert. (OS/-ANIM)Peter can see that the game angers the coach.

1d. Peter kann sehen, dass das Kind den Trainer argert. (SO/ + ANIM)Peter can see that the child angers the coach.

1c. Peter kann sehen, dass den Trainer das Kind argert. (OS/ + ANIM)Peter can see that the child angers the coach.

2a. Sie merken, dass das Bild den Grafiker argert.2b. Sie merken, dass den Grafiker das Bild argert.2c. Sie merken, dass die Kunstlerin den Grafiker argert.2d. Sie merken, dass den Grafiker die Kunstlerin argert.

They notice that the picture/the artist angers the graphic designer.

3a. Ich weib, dass die Frage den Arzt argert.3b. Ich weib, dass den Arzt die Frage argert.3c. Ich weib, dass der Junge den Arzt argert.3d. Ich weib, dass den Arzt der Junge argert.

I know that the question/the boy angers the doctor.

4a. Sie erzahlt uns, dass das Foto den Maler beeinflusst.4b. Sie erzahlt uns, dass den Maler das Foto beeinflusst.4c. Sie erzahlt uns, dass die Schulerin den Maler beeinflusst.4d. Sie erzahlt uns, dass den Maler die Schulerin beeinflusst.

She tells us that the photo/the pupil influences the painter.

5a. Es scheint, dass das Gesetz den Kaufmann beeinflusst.5b. Es scheint, dass den Kaufmann das Gesetz beeinflusst.

(continued on next page)

Appendix (continued)

5c. Es scheint, dass die Sekretarin den Kaufmann beeinflusst.5d. Es scheint, dass den Kaufmann die Sekretarin beeinflusst.

It appears that the law/the secretary influences the businessman.

6a. Es ist schade, dass der Preis den Verkaufer beeinflusst.6b. Es ist schade, dass den Verkaufer der Preis beeinflusst.6c. Es ist schade, dass der Mann den Verkaufer beeinflusst.6d. Es ist schade, dass den Verkaufer der Mann beeinflusst.

It is too bad that the price/the man influences the salesman.

7a. Es ist toll, dass der Kurs den Onkel interessiert.7b. Es ist toll, dass den Onkel der Kurs interessiert.7c. Es ist toll, dass die Frau den Onkel interessiert.7d. Es ist toll, dass den Onkel die Frau interessiert.

It is great that the course/the woman interests the uncle.

8a. Es ist schon, dass die Zeitschrift den Teenager interessiert.8b. Es ist schon, dass den Teenager die Zeitschrift interessiert.8c. Es ist schon, dass das Madchen den Teenager interessiert.8d. Es ist schon, dass den Teenager das Madchen interessiert.

It is nice that the magazine/the girl interests the teenager.

9a. Wir horen, dass die Antwort den Lehrer interessiert.9b. Wir horen, dass den Lehrer die Antwort interessiert.9c. Wir horen, dass die Studentin den Lehrer interessiert.9d. Wir horen, dass den Lehrer die Studentin interessiert.

We hear that the answer/the student interests the teacher.

10a. Es ist klar, dass der Job den Computerprogrammierer erschreckt.10b. Es ist klar, dass den Computerprogrammierer der Job erschreckt.10c. Es ist klar, dass der Techniker den Computerprogrammierer erschreckt.10d. Es ist klar, dass den Computerprogrammierer der Techniker erschreckt.

It is clear that the job/the technician scares the computer programmer.

11a. Es ist unerwartet, dass das Fax den Chef erschreckt.11b. Es ist unerwartet, dass den Chef das Fax erschreckt.11c. Es ist unerwartet, dass der Bankrauber den Chef erschreckt.11d. Es ist unerwartet, dass den Chef der Bankrauber erschreckt.

It is unexpected that the fax/the bank robber scares the boss.

12a. Es ist schlecht, dass der Film den Sohn erschreckt.12b. Es ist schlecht, dass den Sohn der Film erschreckt.12c. Es ist schlecht, dass die Krankenschwester den Sohn erschreckt.12d. Es ist schlecht, dass den Sohn die Krankenschwester erschreckt.

It is bad that the film/the nurse scares the son.

402 C.N. Jackson / System 36 (2008) 388–406

Appendix (continued)

13a. Es ist bekannt, dass das Projekt den Manager langweilt.13b. Es ist bekannt, dass den Manager das Projekt langweilt.13c. Es ist bekannt, dass die Kauffrau den Manager langweilt.13d. Es ist bekannt, dass den Manager die Kauffrau langweilt.

It is known that the project/the business woman bores the manager.

14a. Es ist nicht gut, dass das Gedicht den Schriftsteller langweilt.14b. Es ist nicht gut, dass den Schriftsteller das Gedicht langweilt.14c. Es ist nicht gut, dass der Leser den Schriftsteller langweilt.14d. Es ist nicht gut, dass den Schriftsteller der Leser langweilt.

It is bad that the poem/the reader bores the writer.

15a. Sie denkt, dass das Experiment den Wissenschaftler langweilt.15b. Sie denkt, dass den Wissenschaftler das Experiment langweilt.15c. Sie denkt, dass der Student den Wissenschaftler langweilt.15d. Sie denkt, dass den Wissenschaftler der Student langweilt.

She thinks that the experiment/the student bores the scientist.

16a. Ich glaube, dass der Brief den Enkel beschreibt.16b. Ich glaube, dass den Enkel der Brief beschreibt.16c. Ich glaube, dass die Oma den Enkel beschreibt.16d. Ich glaube, dass den Enkel die Oma beschreibt.

I believe that the letter/the grandmother describes the grandson.

17a. Es ist gut, dass die Kritik den Sanger beschreibt.17b. Es ist gut, dass den Sanger die Kritik beschreibt.17c. Es ist gut, dass der Kritiker den Sanger beschreibt.17d. Es ist gut, dass den Sanger der Kritiker beschreibt.

It is good that the review/the reviewer describes the singer.

18a. Es stimmt, dass die Geschichte den Clown beschreibt.18b. Es stimmt, dass den Clown die Geschichte beschreibt.18c. Es stimmt, dass der Schauspieler den Clown beschreibt.18d. Es stimmt, dass den Clown der Schauspieler beschreibt.

It is true that the story/the actor describes the clown.

19a. Es ist prima, dass die Musik den Vater beruhigt.19b. Es ist prima, dass den Vater die Musik beruhigt.19c. Es ist prima, dass die Tante den Vater beruhigt.19d. Es ist prima, dass den Vater die Tante beruhigt.

It is great that the music/the aunt calms the father.

20a. Es wundert uns, dass die Idee den Politiker beruhigt.20b. Es wundert uns, dass den Politiker die Idee beruhigt.20c. Es wundert uns, dass der Senator den Politiker beruhigt.20d. Es wundert uns, dass den Politiker der Senator beruhigt.

(continued on next page)

C.N. Jackson / System 36 (2008) 388–406 403

Appendix (continued)It amazes us that the idea/the senator calms the politician.

21a. Es freut uns, dass die Praline den Schuler beruhigt.21b. Es freut uns, dass den Schuler die Praline beruhigt.21c. Es freut uns, dass die Mutter den Schuler beruhigt.21d. Es freut uns, dass den Schuler die Mutter beruhigt.

It makes us happy that the chocolate/the mother calms the pupil.

22a. Er liest, dass der Wind den Gartner stort.22b. Er liest, dass den Gartner der Wind stort.22c. Er liest, dass die Fotografin den Gartner stort.22d. Er liest, dass den Gartner die Fotografin stort.

He reads that the wind/the photographer disturbs the gardener.

23a. Alle sagen, dass der Regen den Arbeiter stort.23b. Alle sagen, dass den Arbeiter der Regen stort.23c. Alle sagen, dass der Mechaniker den Arbeiter stort.23d. Alle sagen, dass den Arbeiter der Mechaniker stort.

Everyone says that the rain/the mechanic disturbs the worker.

24a. Es ist sicher, dass der Plan den Ingenieur stort.24b. Es ist sicher, dass den Ingenieur der Plan stort.24c. Es ist sicher, dass der Architekt den Ingenieur stort.24d. Es ist sicher, dass den Ingenieur der Architekt stort.

It is certain that the plan/the architect disturbs the engineer.

404 C.N. Jackson / System 36 (2008) 388–406

References

Bates, E., MacWhinney, B., 1989. Functionalism and the competition model. In: MacWhinney, B., Bates, E.

(Eds.), The Cross-Linguistic Study of Sentence Processing. Cambridge University Press, Cambridge, pp. 3–73.

Bornkessel, I., Schlesewsky, M., Friederici, A., 2002. Grammar overrides frequency: evidence from the online

processing of flexible word order. Cognition 85 (2), B21–B30.

Clahsen, H., Felser, C., 2006. Grammatical processing in language learners. Applied Psycholinguistics 27 (1), 3–

42.

Dekydtspotter, L., Outcalt, S., 2005. A syntactic bias in scope ambiguity resolution in the processing of English–

French cardinality interrogatives: evidence for informational encapsulation. Language Learning 55 (1), 1–36.

Fender, M., 2003. English word recognition and word integration skills of native Arabic- and Japanese-speaking

learners of English as a second language. Applied Psycholinguistics 24 (2), 289–315.

Frenck-Mestre, C., 1997. Examining second language reading: an on-line look. In: Sorace, A., Heycock, C.,

Shillock, R. (Eds.), Proceedings of the GALA 1997 Conference on Language Acquisition. Human

Communications Research Center, Edinburgh, pp. 474–478.

Gass, S., 1987. The resolution of conflicts among competing systems: a bidirectional perspective. Applied

Psycholinguistics 8 (4), 329–350.

Harrington, M., 2001. Sentence processing. In: Robinson, P. (Ed.), Cognition and Second Language Instruction.

Cambridge University Press, Cambridge, pp. 91–124.

Hemforth, B., Konieczny, L., Strube, G., 1993. Incremental syntax processing and parsing strategies. In:

Proceedings of the Annual Conference of the Cognitive Science Society, USA, vol. 15, pp. 539–544.

Hoover, M.L., Dwivedi, V.D., 1998. Syntactic processing in skilled bilinguals. Language Learning 48 (1), 1–29.

Hopp, H., 2006. Syntactic features and reanalysis in near-native processing. Second Language Research 22 (3),

369–397.

C.N. Jackson / System 36 (2008) 388–406 405

Jackson, C.N., 2005. The sentence-level processing of case markings and word order by native and non-native

speakers of German. Unpublished doctoral dissertation, University of Wisconsin, Madison.

Jackson, C.N., 2007. The use and non-use of semantic information, word order, and case markings during

comprehension by L2 learners of German. Modern Language Journal 91 (3), 418–432.

Jiang, N., 2004. Morphological insensitivity in second language processing. Applied Psycholinguistics 25 (4), 603–

634.

Jiang, N., 2007. Selective integration of linguistic knowledge in adult second language learning. Language

Learning 57 (1), 1–33.

Juffs, A., 2005. The influence of first language on the processing of wh-movement in English as a second language.

Second Language Research 21 (2), 121–151.

Juffs, A., Harrington, M., 1995. Parsing effects in second language sentence processing: subject and object

asymmetries in wh-extraction. Studies in Second Language Acquisition 17 (4), 482–516.

Kempe, V., MacWhinney, B., 1998. The acquisition of case markings by adult learners of Russian and German.

Studies in Second Language Acquisition 20 (4), 543–587.

Kilborn, K., 1989. Sentence processing in a second language: the timing of transfer. Language and Speech 32 (1),

1–23.

Lee, J.F., VanPatten, B., 2003. Making Communicative Language Teaching Happen, second ed. McGraw-Hill,

Boston.

Lenerz, J., 1977. Zur Abfolge nominaler Satzglieder im Deutschen (On the sequence of nominal phrases in

German). Gunther Narr Verlag, Tubingen, Germany.

LoCoco, V., 1987. Learner comprehension of oral and written sentences in German and Spanish: the importance

of word order. In: VanPatten, B., Dvorak, T.R., Lee, J. (Eds.), Foreign Language Learning: A Research

Perspective. Newbury House, Cambridge, MA, pp. 119–129.

MacWhinney, B., Bates, E., Kliegl, R., 1984. Cue validity and sentence interpretation in English, German and

Italian. Journal of Verbal Learning and Verbal Behavior 23 (2), 127–150.

Matzke, M., Mai, H., Nager, W., Russeler, J., Munte, T., 2002. The cost of freedom: an ERP-study of non-

canonical sentences. Clinical Neurophysiology 113 (6), 844–852.

McDonald, J., 1987. Sentence interpretation in bilingual speakers of English and Dutch. Applied Psycholin-

guistics 8 (4), 379–413.

McLaughlin, B., Tammi, R., McLeod, B., 1983. Second language learning: an information-processing

perspective. Language Learning 33 (2), 135–158.

Osterhout, L., McLaughlin, J., Kim, A., Greenwald, R., Inoue, K., 2004. Sentences in the brain: event-related

potentials as real-time reflections of sentence comprehension and language learning. In: Carreiras, M.,

Clifton, C. (Eds.), The On-Line Study of Sentence Comprehension. Psychology Press, New York, pp. 271–

308.

Papadopoulou, D., 2005. Reading-time studies of second language ambiguity resolution. Second Language

Research 21 (2), 98–120.

Ritterbusch, R., LaFond, L., Agustin, M., 2006. Learner difficulties with German case: implications from an

action research study. Die Unterrichtspraxis 39 (1–2), 30–45.

Sasaki, Y., 1994. Paths of processing strategy transfers in learning Japanese and English as foreign languages.

Studies in Second Language Acquisition 16 (1), 43–71.

Scheepers, C., Hemforth, B., Konieczny, L., 2000. Linking syntactic functions with thematic roles: psych-verbs

and the resolution of subject-object ambiguity. In: Hemforth, B., Konieczny, L. (Eds.), German Sentence

Processing. Kluwer Academic Publishers, Dordrecht, The Netherlands, pp. 95–137.

Schmidt, R., 1993. Awareness and second language acquisition. Annual Review of Applied Linguistics 13, 206–

226.

Schmidt, R., 2001. Attention. In: Robinson, P. (Ed.), Cognition and Second Language Instruction. Cambridge

University Press, Cambridge, pp. 3–32.

Schneider, W., 1997. Micro Experimental Laboratory v2.01. Psychology Software Tools Inc., Pittsburgh, PA.

Sharwood Smith, M., 1993. Input enhancement in instructed SLA: theoretical bases. Studies in Second Language

Acquisition 15 (2), 165–179.

Tokowicz, N., MacWhinney, B., 2005. Implicit and explicit measures of sensitivity to violations in second

language grammar. Studies in Second Language Acquisition 27 (2), 173–204.

Williams, J.N., Mobius, P., Kim, C., 2001. Native and non-native processing of English wh-questions: Parsing

strategies and plausibility constraints. Applied Psycholinguistics 22 (4), 509–540.

406 C.N. Jackson / System 36 (2008) 388–406

Wong, W., 2004. The nature of processing instruction. In: VanPatten, B. (Ed.), Processing Instruction: Theory,

Research, and Commentary. Lawrence Erlbaum Associates, Mahwah, NJ, pp. 33–64.

VanPatten, B., 1996. Input Processing and Grammar Instruction in Second Language Acquisition. Ablex

Publishing Corp., Norwood, NJ.

VanPatten, B., 2004. Input processing in SLA. In: VanPatten, B. (Ed.), Processing Instruction: Theory, Research,

and Commentary. Lawrence Erlbaum Associates, Mahwah, NJ, pp. 5–32.

Zubin, D., 1977. The semantic bases of case alternation in German. In: Fasold, R.W., Shuy, R.W. (Eds.), Studies

in Language Variation: Semantics, Syntax, Phonology, Pragmatics, Social Situations, Ethnographic

Approaches. Georgetown University Press, Washington, DC, pp. 88–99.

Zubin, D., Kopcke, K.M., 1985. Cognitive constraints on the order of subject and object in German. Studies in

Language 9 (1), 77–107.