Morphological knowledge affects processing of L2 derivational morphology: An event-related potential study

Journal of Neurolinguistics 37 (2016) 47e57

Contents lists available at ScienceDirect

Journal of Neurolinguistics journal homepage: www.elsevier.com/locate/jneuroling

Research paper

Morphological knowledge affects processing of L2 derivational morphology: An event-related potential study Taiping Deng a, b, Jiawei Shi a, Susan Dunlap c, Hongyan Bi d, Baoguo Chen a, * a

Beijing Key Laboratory of Applied Experimental Psychology, School of Psychology, Beijing Normal University, Beijing, 100875, China Zhejiang University of Finance & Economics, Hangzhou, 310018, China c Children's Learning Institute, University of Texas Health Science Center at Houston, United States d Key Laboratory of Behavioral Science, Institute of Psychology, Chinese Academy of Sciences, Beijing, 100101, China b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 5 February 2015 Received in revised form 1 September 2015 Accepted 2 September 2015 Available online 15 September 2015

This study used event-related potentials (ERPs) to investigate the effect of morphological knowledge on L2 derivational processing in a sentence reading task with Chinese-English bilinguals. Correctly derived words and pseudo-derived words were embedded in a semantically plausible sentence. Compared with the correctly derived words, a significant P600 to pseudo-derived words was elicited in the group with high morphological knowledge, indicating their sensitivity to rule violations and application of rule-based decomposition. For the group with low morphological knowledge, a significant N400 was observed, suggesting that participants in this group depend more on a whole-word processing mechanism. These results suggest that morphological knowledge plays an important role in L2 processing of derivational morphology. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Morphological knowledge Derivational morphology Second language

1. Introduction Recently, many researchers have shown great interest in second language (L2) morphological processing (Diependaele, ~ abeitia, Morris, & Keuleers, 2011; Gor & Cook, 2010; Lehtonen, Niska, Wande, Niemi, & Laine, 2006; NeubKauer & Dun Clahsen, 2009; Pliatsikas & Marinis, 2013; Silva & Clahsen, 2008). Much of this work has investigated whether second language (L2) learners, especially late learners who learned the L2 after puberty, can apply rule-based decompositions to process morphologically complex words. Following a dual-mechanism model, researchers have investigated whether morphologically complex words are stored in memory as whole-word units to be directly accessed from the separate lexical entries during processing, or whether these words are processed via some rule-governed decomposition (Pinker, 1999; Pinker & Ullman, 2002). However, there have been a limited number of studies, especially with online processing measures, and findings have been inconsistent (Clahsen, Felser, Neubauer, Sato, & Silva, 2010). Thus, the details of how late L2 learners process morphologically complex words remain largely unknown. Much research concerning L2 complex words has mainly focused on inflectional morphology, particularly for the past n-Galle s, Diaz, & Rodriguez-Fornells, 2005; Gor & Cook, 2010; Lehtonen et al., 2006; tense (De Diego Balaguer, Sebastia NeubKauer & Clahsen, 2009; Pliatsikas & Marinis, 2013; Silva & Clahsen, 2008). For instance, Silva and Clahsen (2008) adopted a masked priming paradigm to investigate regularly inflected words by groups of Chinese and German L2

* Corresponding author. School of Psychology, Beijing Normal University, Beijing, 100875, China. E-mail address: [email protected] (B. Chen). http://dx.doi.org/10.1016/j.jneuroling.2015.09.001 0911-6044/© 2015 Elsevier Ltd. All rights reserved.

48

T. Deng et al. / Journal of Neurolinguistics 37 (2016) 47e57

learners of English. Three types of primeetarget pairs, including identity (e.g. prayepray), test (e.g. prayed-pray), and unrelated (e.g. bake-pray), were presented to the participants. The native speakers showed full priming effects as evidenced by the amount of priming in the test condition being equivalent to the amount of priming in the identity condition. The results of the L2 learners showed no priming effect: The regular past-tense primes yielded the same target reaction times as unrelated primes, indicating that the L2 learners relied on the whole-word processing. In the masked visual priming experiment of NeubKauer and Clahsen (2009), the German regular past participles that are suffixed with -t were processed by highly proficient adult Polish (L1)-German (L2) learners. The three types of primeetarget pairs as in Silva and Clahsen (2008) were also included. The L1 group showed a full priming effect for the regular participles, suggesting that they rely on morphological parsing. The L2 learners, showed no prime effect, indicating that they rely more on lexical storage than morphological parsing. However, Pliatsikas and Marinis (2013) tested highly proficient Greek-English learners with naturalistic or classroom L2 exposure in a self-paced reading task focusing on past tenses forms. The inflected and pseudo-inflected verbs were embedded in the semantically plausible and syntactically simple sentences. The time that the participants spent on the segment that includes the verb was analyzed as the critical indicator. Reading times for the regularly inflected forms were significantly longer than those for the pseudo-inflected verbs, indicating that L2 learners rely on rule-based decomposition for regular forms. In brief, as concerns L2 inflectional processing, whether L2 learners can apply rule-based decompositions remains equivocal. As for L2 derivational morphology, another type of morphologically complex words, the number of studies is rather scarce, and the results are also inconsistent (Diependaele et al., 2011; Silva & Clahsen, 2008). For example, Silva and Clahsen (2008) observed a reduced priming effect for L2 learners when processing derived words, suggesting that the L2 learners applied the rule-based decompositions in a less automatic way compared with native speakers. However, Diependaele et al. (2011) investigated the performance of Spanish-English and Dutch-English bilinguals on a masked morphological priming lexical decision task. Three types of derivational words were selected, including transparent suffixed primes, opaque suffixed primes, and form control primes. The results indicated a graded facilitation pattern in both groups of L2 learners, with the largest priming effect in the transparent condition, intermediate in the opaque condition, and smallest in the form condition. The data suggested that L2 learners, similar to native speakers, can engage in the decomposing process of the derivational words. Based on the limited number of derivational studies and inconsistent results, it is difficult to draw any conclusions of whether L2 learners engage in decomposition of derived words. The declarative/procedural model posed by Ullman posited that late L2 learners may lack automated procedural memory for L2, leaving them to depend largely on declarative memory. Therefore, their abilities to compute aspects of morphologically complex words may be reduced. To be more specific, these morphologically complex words would be stored as whole-word units and could be directly accessed. However, with an increase in proficiency or exposure, L2 learners can begin to apply rulegoverned decomposition during online processing. This theory is also open in principle to the possibility that other factors besides proficiency might affect the attainment of rule-based decomposition (Ullman, 2005, 2012). Thus, the present study aims to explore other possible factors. Moreover, the studies reviewed above were all conducted using behavioral methods. As claimed by Morris and Holcomb (2005), it is difficult to test the representation and processing of morphologically complex words by behavioral methods based only on the reaction time data. The event related potentials (ERPs) technique, owing to its high temporal resolution, has been widely used to investigate the neural mechanisms of online language processing. In this study, we made use of this technique to further investigate how later L2 learners process derivational words. ERP studies on L1 derivational morphology have been conducted at the single word level and at the sentence level. At the single-word level, the priming paradigm is mostly employed (Kielar & Joanisse, 2011; McKinnon, Allen, & Osterhout, 2003). The N400, a negative going waveform that peaks around 400 ms after stimulus onset, is consistently elicited, and the reduction of this component is taken as an indication of semantic integration, possibly aided by morphemic segmentation (Kielar & Joanisse, 2011; Lavric, Clapp, & Rastle, 2007). At the sentence level, the violation paradigm is often used, and two ERP components are often focused on: N400 and P600. In sentence processing, the N400 has been related to lexical processing and semantic integration processes (Friederici & Weissenborn, 2007; Kutas & Federmeier, 2011). In derivational violations, the N400 has been interpreted to reflect the possibility that the pseudo-derived words are processed as whole units (Lück, Hahne, & Clahsen, 2006) or as pseudo-words (Janssen, Wiese, & Schlesewsky, 2006). Also in sentence contexts, the N400 has been €rvenpa €a €c, & thought to reflect a failed attempt to semantically integrate incorrectly derived words (Leinonen, Bratticoa, Ja Krausea, 2008). The P600, a late positive component for morphological violations, has been observed at sentence level processing involving affixes (Leinonen, Brattico, et al., 2008). This late positivity may reflect the repair process where the incorrectly combined stem and suffix is placed with a compatible suffix (Leinonen, Brattico, et al., 2008). Also, this positive component can be taken as indicative of combinatorial linguistic processes, specifically with respect to morphological violations (Havas, Rodrí guez-Fornells, & Clahsen, 2012). In summary, studies on L1 ERP derivational processing produce a relatively consistent pattern. Especially at the sentence level, the N400 component can be interpreted as the word-level lexical-semantic processing, and the P600 can be taken as an indication of combinatorial processes. In contrast to the relatively extensive ERP studies on L1 morphologically complex words, the ERP research on these words in L2 is less prevalent. In L2 inflectional processing, for instance, Hahne, Mueller, and Clahsen (2006) have investigated the processing of German inflected words by German learners with Russian as their native language. Two different types of German inflection were studied: participial inflection and noun plurals inflection. The inflected words were embedded in sentence contexts. ERP results showed that in L2 learners, processing regular violations elicited an anterior negativity and/or

T. Deng et al. / Journal of Neurolinguistics 37 (2016) 47e57

49

P600, indicating morphological decomposition. For irregular violations, the N400 component was found, suggesting that these words were stored as whole-word forms. Chen, Shu, Liu, Zhao, and Li (2007) investigated inflectional morpheme processing of subject-verb agreement in the obligatory context. Highly proficient Chinese (L1) - English (L2) learners showed a pattern of negativity in the 500e700 ms time window to the violations of subject-verb agreement, indicating ERP mechanisms distinct from native speakers, who showed a biphasic LAN-P600 profile of morpho-syntactic processing. In brief, where L2 inflectional processing is concerned, the N400 component can be taken as an indicator of whole-word processing, and the P600 or the pattern of LAN-P600 can be taken as an indicator of rule-based decomposition. However, inflection and derivation are deemed to represent functionally distinct processes (Anderson, 1982; Leinonen et al., 2008; Stump, 1998). Inflections do not create new words with new meanings, and are sensitive to the grammatical environment in which they occur. Inflectional morphemes specify the grammatical functions of the words in phrases or sentences (Bickel & Nichols, 2007; Bozic, Tyler, Su, Wingfield, & Marslen-Wilson, 2013). Therefore, the processing of inflectional morphology may not reflect the complete lexical processing of the inflected word itself. However, derivations create new words that may have new albeit related meanings. Processing of derivational morphology is a lexical process much more independent of grammatical context and is constrained by more selectional restrictions (Bickel & Nichols, 2007; Bozic et al., 2013). Therefore, the processing of derivational morphology may reflect a much purer processing of decomposition of derived words. Moreover, compared with inflectional processing, online processing of derived words in L2 remains a relatively unstudied area. Therefore the present study aims to address this gap by investigating derived processing with an ERP violation paradigm. To be specific, we focused on whether adult L2 learners employ decomposition mechanisms when processing derived words, and what factors may affect their processing patterns. A number of factors have been suggested to affect the processing of L2 morphologically complex words. Besides the linguistic properties of the affix that make the words semantically transparent or opaque, other factors such as age of acquisition (Johnson & Newport, 1989), L1 transfer, cognitive resources limitations (Clahsen et al., 2010), language proficiency, or exposure (Liang & Chen, 2014; Ullman, 2001, 2004) are all believed to influence the application of rule-based decomposition in online processing. However, at the present time, there has been no study to explore the relationship between L2 morphological knowledge and online processing of morphologically complex words. Therefore, the present study aimed to explore the effect of morphological knowledge on L2 derivational processing. Morphological knowledge, a more general or superordinate term (Bowers, Kirby, & Deacon, 2010; Nagy, Carlisle, & Goodwin, 2014), is a metacognitive skill (Clark, Gilbert, & Anderson, 2011) which includes morphological awareness (Bowers et al., 2010; Nagy et al., 2014). It refers to the understanding of the morphemic structure of morphologically complex words and the ability of the learner to reflect on and manipulate the structures of the words. In L1, many studies have demonstrated the bidirectional relationship between morphological knowledge and reading (Carlisle, 2000; Feldman & Bentin, 1994; McBride-Chang, Wagner, Muse, Chow, & Shu, 2005; Nagy, Berninger, & Abbott, 2006). On one hand, morphological knowledge, especially derivational morphological awareness, is closely associated with reading (Nagy et al., 2006; Walter, Wood, & D’zatko, 2009). Studies have consistently demonstrated that morphological knowledge made a unique contribution to word decoding ability (Deacon, Campbell, Tamminga, & Kirby, 2010; Deacon, Parrila, & Kirby, 2006; McBride-Chang, Tardif, Cho, Shu, Fletcher, Stokes, 2008; Nagy, Anderson, Schommer, Scott, & Stallman, 1989). More specifically, some studies have observed a relationship between morphological knowledge and reading morphologically complex words (Carlisle, 2000; Nagy et al., 2006). For instance, Nagy et al. (2006) conducted a study of English-speaking children from grade 4 to grade 8, and found that morphological knowledge was more related to the decoding of morphologically complex words than to the decoding of simple words. On the other hand, Deacon, Benere, and Pasquarella (2013) found a reciprocal relationship between reading and morphological awareness by evidence that reading contributes to the improvement of the morphological awareness (Deacon et al., 2013). This is consistent with the predication of Kuo and Anderson (2006), who claimed that extensive exposure to print could lead to better morphological awareness. The researchers hypothesized that, when encountering morphologically complex words, children do tend to analyze the subcomponent of the words (e.g. darkness ¼ dark þ ness). Thus their morphological awareness may be improved through their reading. In L1, many studies have demonstrated a robust bidirectional relationship between morphological knowledge and reading. Particularly considering that reading is closely associated with morphological knowledge in English, there is the possibility that L2 learners can also improve their understanding of word decomposition rules through reading and the continuous accumulation of morphological knowledge. Even as adults, L2 learners can be exposed to other input sources besides classroom reading, including reading articles, original novels, and English magazines. Understanding of word formation rules may be enhanced through accumulated reading. Therefore it is rational for us to predict that the processing mechanism of L2 derivational words may be different between groups with high versus low morphological knowledge. Compared to English, derivations and inflections in Chinese-both Mandarin and Cantonese-are very minimal (Li, Jin, & Tan, 2004; Zhang & Koda, 2014). As more than 75% of Chinese words are compounds (Institute of Language Teaching and Research [in China], 1986), lexical compounding is the predominant structure in this language. Many studies have been conducted to explore the relationship between morphological knowledge and reading in Chinese (Ku & Anderson, 2003; Li & Wu, 2015), with differing results. For example, some studies have indicated the important role of derivational knowledge (Kirby et al., 2012), and other studies have shown that Chinese derivational knowledge did not seem to make a unique contribution to Chinese reading and Chinese vocabulary (Ku & Anderson, 2003). Possibly, for Chinese-English bilinguals, this feature specific to their native language may have an influence on their learning of L2 English derivations (Hernandez & Li, 2007; Hernandez,

50

T. Deng et al. / Journal of Neurolinguistics 37 (2016) 47e57

Li, & MacWhinney, 2005), possibly causing greater difficulty for these Chinese learners. However, exposure may result in growth of the morphological knowledge and deepen the understanding of the rules about derivational words (Deacon et al., 2013). Thus, it may be possible that there will be processing differences between the participants with low L2 morphological knowledge and those with high L2 morphological knowledge. For those with low morphological knowledge, as they have a worse understanding of the rules of derivational knowledge (Deacon et al., 2013), they might tend to employ a different processing strategy in real time processing. The present study adopted an ERP technique to investigate the effect of morphological knowledge of Chinese-English bilinguals on L2 derivational processing in sentence contexts. Because the LAN is difficult to be elicited in adult L2 learners (Chen et al., 2007; Hahne, 2001; Hahne & Friederici, 2001; Morgan-Short, Sanz, Steinhauer, & Ullman, 2010; Morgan-Short, Steinhauer, Sanz, & Ullman, 2012; Ojima, Nakata, & Kakigi, 2005), we focused on the N400 and P600 components which respectively signify the neural mechanism of whole-word processing and rule-governed decomposition (Hahne et al., 2006). If morphological knowledge modulates the processing mechanism of derived words, a P600 should be elicited in the group with high morphological knowledge and a significant N400 component should be observed in the group with low morphological knowledge. 2. Method 2.1. Participants Participants were 36 Chinese (L1)-English (L2) bilinguals, enrolled in Beijing Normal University, who have passed the CET 4 with scores higher than 560 (CET, College English Test, is administered by the Ministry of Education of China for non-English major college students. The highest possible score is 710; the low cut-off score is 425.). According to their median score on the morphological knowledge test, participants were divided into two groups. The group with high morphological knowledge (HG) consisted of 18 participants (11 female), and the group with low morphological knowledge (LG) consisted of 18 participants (13 female). All participants reported being right-handed and having normal or corrected-to-normal vision and no history of neurological impairment. All the participants provided informed consent and received a monetary compensation for their participation. Proficiency scores were obtained through a self-rating questionnaire and the Oxford Placement Test (OPT). The self-rating questionnaire was a 6-point scale (1 ¼ quite poor, 6 ¼ highly proficient) to measure the participants' abilities in listening, speaking, reading, and writing. The OPT includes 25 multiple-choice questions and a cloze test, and the highest possible score is 50. The mean age, duration of L2 learning, morphological knowledge test score, second language proficiency ratings and OPT score of both participant groups are presented in Table 1. 2.2. Morphological knowledge test and materials 2.2.1. Morphological knowledge test A morphological test was administered to test the participants' derivational knowledge, that is, their knowledge about the morphological structure of derived words in English. It was designed specifically to assess the L2 learners' awareness of the relations of base and derived forms. The items in this morphological knowledge test were revised or reconstructed according to the rules adopted by previous studies (Carlisle, 2000; Wang, Cheng, & Chen, 2006). The whole morphological knowledge test included two parts. The first part required the production of a derived word in order to finish a sentence. For example, provided with the prompt word “bold” and the sentence “Linda's __ always gets her into trouble,” the correct response is “boldness.” This task was used to assess the L2 learners' awareness of the morphological structure of words and placed emphasis on their ability to produce the complex forms, given the base word and a sentence context. The second part of the test was a true/false test that required participants to judge whether the given word was the correct derivational form of the base word, for example “deaf”-“deafment” (incorrect) or “deaf” -“deafness” (correct). In this test, the correct base words were Table 1 The mean age, duration of L2 learning, morphological knowledge test score, English proficiency ratings score and OPT score (standard deviations) for both participant groups. Group

LG

Age Years of L2 learning Morphological knowledge Listening Speaking Reading Writing OPT

23.33 11.30 57.06 4.10 3.98 5.82 5.87 40.22

HG (2.11) (0.89) (6.08) (0.72) (0.54) (0.87) (0.36) (2.46)

24.09 10.70 70.50 4.00 4.12 5.68 5.60 39.50

(2.37) (1.92) (4.17) (1.03) (0.78) (0.56) (0.34) (1.79)

Note: LG ¼ group with low morphological knowledge; HG ¼ group with high morphological knowledge.

T. Deng et al. / Journal of Neurolinguistics 37 (2016) 47e57

51

matched with either correct or incorrect suffixes. The items assessed the correct understanding of the word relations between the base word and its derived word. Both parts of the morphological knowledge test contained transparent words and shift words. After a relatively large-scale testing with a separate group of 137 college students who have the similar background as the participants, any items with a passing rate higher than 0.8 or lower than 0.2 were removed, thus resulting in 100 items in total in the version administered to the study participants. The Cronbach's alpha coefficient of the test was 0.92. Each part included 50 items, and each correct answer was awarded one point so that scores on the complete test range from 0 to 100. For the grouping criterion, we adopted the median split method to classify the participants into two groups (Carver, Vafaei, Guerra, Freire, & Phillips, 2013; Gale-Ross, Baird, & Towson, 2009). First, the median of all the participants was established. Then, each individual score was compared to the median. If score was below the median, that individual was classified as having low morphological knowledge (n ¼ 18, range ¼ 44e65). If the individual's score on the morphological knowledge test was equal to or above the median, that person was classified as having high morphological knowledge (n ¼ 18, range ¼ 66e79, see Table 1). Words used in the morphological test were not included in the EEG stimuli. The two groups were statistically equivalent on mean age, duration of L2 learning, self-rated English proficiency, and OPT score (all ps > 0.05). As intended, the scores on the morphological knowledge test were significantly different (t ¼ 4.24, p < 0.001). 2.2.2. Materials Eighty derived English nouns with the suffix -ness, -nity, or -ment were selected because of their high-frequency and rich morphological productivity (Kuo & Anderson, 2006). The frequency of the derived words was calculated according to the frequency dictionary of Brysbaert and New (2009). A separate group of 28 college students from the same background as the participants rated the familiarity of the stimuli on a five-point scale (1 for quite unfamiliar, 5 for quite familiar). The mean familiarity rating of the words was 4.63 (SD ¼ 0.21); the mean length of words was 8.9 letters (SD ¼ 1.24). Eighty pseudoderived words corresponding the correctly derived words were created based on combining of the same stem with an incorrect suffix (e.g. stateness), according to the same procedure used by Havas, Todriguez-Fornells, and Clahsen (2012). The suffixes of the incorrectly derived words were constrained to the same three options: -ness, -nity, -ment, to ensure that the incorrectly derived words were similar to the correct words in form, length, and part of speech. Thus, 80 correctly derived words (e.g. statement) and 80 pseudo-derived words (stateness) were embedded in sentence contexts with SVO structure, resulting in 160 experimental sentences. Each experimental sentence contained 8 words, with the target word as the object and fifth word of each sentence. Two counterbalanced lists were constructed to ensure that participants never saw different versions of the same root words. Eighty filler sentences of the same length (8 words) were included, for a total of 160 sentences in each list. Example sentences are provided as follows: (1) The president gave a statement to the public. (sentence with a correctly derived word) (2) The president gave a stateness to the public. (sentence with a pseudo-derived word) (3) Jane loves the small restaurant in the hotel. (filler sentence) The same group of 28 raters was asked to assess cloze probability of the derived words and semantic plausibility for the correct sentences. The proportion of the raters who filled in the blank of the sentences with the derived target words was taken as the indicator of cloze probability of the derived words. The semantic plausibility of the sentences was rated on a fivepoint scale (1 for more implausible, 5 for more plausible). The mean sentence constraint was 0.29 (SD ¼ 0.13); the mean semantic plausibility rating was 4.73 (SD ¼ 0.23). The ratings results show that the sentence context constraints were low and the sentences were semantically plausible. 2.3. Procedure We employed the ERP violation paradigm similar to that of previous ERP studies on derived words (Havas et al., 2012). Stimuli were presented using E-Prime1.1. Each trial began with the presentation of a fixation cross (500 ms) in the center of the screen followed by sentences that were presented word-by-word in the center of the screen. Each word was presented for 500 ms, and the inter-stimulus interval was 500 ms (Havas et al., 2012). Two asterisks following the last word of each sentence were presented to indicate the end of the sentence. Thirty percent of the fillers were followed by a simple follow-up question to ensure the participants carefully read the sentences for comprehension. Participants answered by pressing “J” to respond “true” or “F” to respond “false” to the comprehension questions. Half of the comprehension questions were true, and half were false. Participants were tested individually in a quiet experimental room and were asked to minimize their blinks and body movements during EEG recording. Participants were allowed breaks after finishing every 40 trials. The experiment began with the presentation of 6 practice sentences. 2.4. ERP recordings and data analysis EEG signals were recorded from a 64-channel Quik-cap with Ag/AgCl electrodes placed according to the extended 10e20 system. All electrodes were referenced to the left mastoid during recording and were off-line referenced to the right mastoid.

52

T. Deng et al. / Journal of Neurolinguistics 37 (2016) 47e57

Impedances were kept below 5 KU. The vertical EOG was recorded with two electrodes placed above and below the left eyes, and the horizontal EOG was recorded with two electrodes placed on the right and left canthi. The EEG was amplified by Neuroscan SynAmps amplifiers and filtered on-line with a band-pass from 0.1 to 100 Hz with a sampling rate of 1000 Hz and later low-pass (40 Hz) filtered off-line. For ERP analysis, EEG data time-locked to the onset of the correctly derived or pseudoderived target words were averaged for each participant for all the electrodes, for the duration from 200 ms to 1000 ms, using a 200 ms pre-stimulus baseline. Scan 4.3 software was used for off-line data analysis. Epochs with EEG exceeding ±75 mV at any channel were automatically rejected off-line. The electrodes selected for analysis were from the occipital-parietal region according to the distribution of N400 and P600 which covered a three-level distribution: left (CP3, P7, PO5, O1), middle (Cz, Pz, POz, Oz), and right (CP4, P8, PO6, O2). In accordance with previous literature (Tanner, McLaughlin, Herschensohn, & Osterhout, 2013) and visual inspection of the waveforms, we chose three time windows for analysis: 100e200 ms for N1, 300e500 ms for N400, and 600e900 ms for P600. Mean amplitudes for each time window were analyzed using a global ANOVA with the between-subjects factor Group (HG, LG); the within-subject factors condition (correctly derived word, pseudo-derived word); and distributional factor laterality (left, middle, right). Significance levels of the F ratios were adjusted with the Greenhouse-Geisser correction.

3. Results The average comprehension accuracy of the 36 participants was 89%, indicating that the participants carefully read the sentences for comprehension. Fig. 1 and Fig. 2 represent the waveforms from the correctly derived and pseudo-derived conditions at the selected electrodes for each group (LG and HG, respectively).

3.1. 100e200 ms Time window In 100e200 ms time window (N1), an ANOVA showed that, the main effect of condition was not significant, F(1, 34) ¼ 0.20, p > 0.05, and the interaction of condition, group, and laterality was not significant, F(2, 68) ¼ 1.12, p > 0.05. The interaction between condition and laterality was significant, F(2, 68) ¼ 4.81, p ¼ 0.01. The interaction between group and condition was marginally significant, F(1, 34) ¼ 4.06, p ¼ 0.052. In order to clarify the marginally significant interaction between group and condition, paired t-tests were conducted for the group with high morphological knowledge and the group with low morphological knowledge. The results showed that there was no significant difference in amplitude between the correctly derived condition and pseudo-derived condition in both groups, HG: t(17) ¼ 1.71, p ¼ 0.11; LG: t(17) ¼ 1.12, p ¼ 0.28, HG: correct ¼ 0.75 mV vs. pseudo ¼ 1.25 mV; LG: correct ¼ 0.88 mV vs. pseudo ¼ 0.56 mV. To clarify the nature of the significant

Fig. 1. Derivational processing in participants with low morphological knowledge. Average waveforms for correctly derived words and pseudo-derived words, and scalp topographic maps representing difference waveforms (pseudo-derived minus correctly derived) across the relevant time-windows.

T. Deng et al. / Journal of Neurolinguistics 37 (2016) 47e57

53

Fig. 2. Derivational processing in participants with high morphological knowledge. Average waveforms for correctly derived words and pseudo-derived words, and scalp topographic maps representing difference waveforms (pseudo-derived minus correctly derived) across the relevant time-windows.

interaction between condition and laterality, follow-up analyses showed that this significant interaction was driven by the significant main effect of laterality in the pseudo-derived conditions, F(2, 70) ¼ 3.43, p ¼ 0.04, h2 ¼ 0.08. In sum, as far as the N1 component in this time window is concerned, there was no significant morphologically-related N1 component elicited in both groups. 3.2. 300e500 ms Time window In the 300e500 ms time window (N400), an ANOVA showed that, the main effect of condition was not significant, F(1, 34) ¼ 2.59, p > 0.05, the main effect of group was not significant, F(1, 34) ¼ 0.44, p > 0.05, the interaction between group and laterality was not significant, F(2, 68) ¼ 0.24, p > 0.05, the interaction between condition and laterality was not significant, F(2, 68) ¼ 0.53 p > 0.05, and the three-way interaction of group, condition, and laterality was not significant, F(2, 68) ¼ 1.33, p > 0.05. However, the main effect of laterality was significant, F(2, 68) ¼ 5.86, p < 0.01, h2 ¼ 0.15, and the interaction between condition and group was significant, F(1, 34) ¼ 5.56, p < 0.05, h2 ¼ 0.14. In order to clarify the nature of the interaction between condition and group, paired-sample t-tests were applied to each group to compare the amplitude differences between the correctly derived condition and the pseudo-derived condition. For the group with low morphological knowledge, the paired-sample t-test showed that the pseudo-derived condition elicited a significantly more negative component than the correctly derived condition, t (17) ¼ 3.02, p ¼ 0.008, d ¼ 0.31, (correct ¼ 2.42 mV vs. pseudo ¼ 1.66 mV). For the group with high morphological knowledge, the amplitude difference between the correctly derived condition and the pseudo-derived condition was not significant, t (17) ¼ 0.49, p ¼ 0.63, (correct ¼ 2.49 mV vs. pseudo ¼ 2.63 mV). 3.3. 600e900 ms Time window In the 600e900 ms time window (P600), an ANOVA showed that the main effect of condition was not significant, F(1, 34) ¼ 0.26, p > 0.05, the main effect of group was not significant, F(1, 34) ¼ 0.08, p > 0.05, and the interaction between group and laterality was not significant, F(2, 68) ¼ 1.14, p > 0.05. However, the main effect of laterality was significant, F(2, 68) ¼ 8.26, p < 0.01, h2 ¼ 0.20, the interaction between condition and group was significant, F(1, 34) ¼ 19.23, p < 0.01, h2 ¼ 0.36, and the interaction between condition and laterality was significant, F(2, 68) ¼ 3.89, p < 0.05, h2 ¼ 0.11, which was qualified by the higher three-way interaction of condition, group, and laterality, F(2, 68) ¼ 6.95, p < 0.01, h2 ¼ 0.17. To clarify the nature of the three-way interaction, an ANOVA including condition and laterality was conducted separately for each group. The results showed that for the low morphological knowledge group (LG), the main effect of condition was

54

T. Deng et al. / Journal of Neurolinguistics 37 (2016) 47e57

significant, F(1, 17) ¼ 6.00, p < 0.05, h2 ¼ 0.26, and was driven by the more negative component to the pseudo derived condition (1.04 mV) than that of the correctly derived condition (1.80 mV, see Fig. 1). According to its distribution, it can be termed an N400 because it seems to be the continuation of the N400 in the 300e500 ms time window. The main effect of laterality was not significant, F(2, 34) ¼ 1.68, p > 0.05, nor was the interaction between laterality and condition, F(2, 34) ¼ 0.40, p > 0.05. For the group with high morphological knowledge (HG), the main effect of condition was significant, F(1, 17) ¼ 16.00, p < 0.01, h2 ¼ 0.49, and was driven by a more positive component to the pseudo derived condition (1.56 mV) than that of the correctly derived condition (0.59 mV). According to its distribution, it was marked as a P600 (see Fig. 2). The main effect of laterality was significant, F(2, 34) ¼ 7.64, p < 0.01, h2 ¼ 0.31, indicating the higher amplitude in middle and right region than that in the left region (left ¼ 0.44 mV; middle ¼ 1.61 mV; right ¼ 1.18 mV). There was also a significant interaction between condition and laterality, F(2, 34) ¼ 7.81, p < 0.05, h2 ¼ 0.32. Further analysis on this interaction revealed that this positivity was shown to be medially distributed and right lateralized (middle location: t(17) ¼ 4.28, p < 0.01, d ¼ 0.57; right location: t(17) ¼ 3.76, p < 0.01, d ¼ 0.54). In summary, for the participants with higher morphological knowledge, a significant P600 was found in the time window of 600e900 ms, whereas for the participants with lower morphological knowledge, a reliable N400 was found in the time window of 300e500 ms and also continued into the following time window (600e900 ms). 4. Discussion The present study aimed to investigate the effect of morphological knowledge on L2 derivational processing using eventrelated potentials (ERP) methodology with Chinese-English bilinguals in a sentence reading task. Significant differences were found between the participants with higher (HG) versus lower (LG) morphological knowledge. For participants with lower morphological knowledge, an N400 to pseudo derived target words was reliably found 300e500 ms post-stimulus onset and persisted into the following time window (600e900 ms), indicating that these items were processed as wholes word rather than decomposed into morphemes (e.g., Janssen et al., 2006 Lück et al., 2006). For participants with higher morphological knowledge, a typical P600 was observed in 600e900 ms time window, indicating that these items were decomposed and detected as morphological violations. Together, these results suggest that morphological knowledge is an important variable modulating the neural mechanism of L2 derived words processing. Items in the pseudo-derived condition were correct stems affixed with an incorrect noun suffix. These items elicited an N400-like negativity in the group with lower morphological knowledge. This type of N400 effect has also found for the € rvenpa €a €, & Krause, 2008; Havas et al., derivational processing at the sentence level by native speakers (Leinonen, Brattico, Ja 2012). At the sentence level, the N400 to violations of suffixes may be taken as indicating the difficulty of semantic integration (Havas et al., 2012). For instance, in the study of Havas, Rodrí guez-Fornells, and Clahsen (2012), an N400 to derivational violations was regarded as an index of word-level lexical-semantic processing, which resulted from a failed attempt to semantically integrate the incorrectly derived words. In the present study, we posit that the N400 found in the group with lower morphological knowledge may indicate that these participants processed the pseudo-derived words as non-words rather than decomposing them into a root and affix (Janssen et al., 2006; Lück et al., 2006). Thus, participants with lower morphological knowledge seem to rely much more on a whole-word processing strategy. In some language-related ERP studies, with respect to morphological and syntactic violations, the P600 has been regarded as an indicator of combinatorial operations (Friederici, 2002; Havas et al., 2012). The P600 component elicited in our participants with higher morphological knowledge may be due to the online decomposition processing of the pseudo-derived words. This account is consistent with at least one study of native language processing. According to Havas et al. (2012), the familiarity and high productivity of Spanish words with the suffix “era” can result in a P600 component indexing combinatorial processing (Havas et al., 2012). In our study, the stems and suffixes were highly familiar and productive (Kuo & Anderson, 2006). Therefore, it should be possible for L2 learners to directly integrate the correct meaning of the pseudoderived words with the sentence context. More precisely, for the pseudo-derived word “stateness,” the L2 learners with high morphological knowledge may directly integrate the meaning of “state” with the correct noun suffix “-ment” and mentally correct it to “statement,” successfully integrating into the sentence contexts. If so, “stateness” would be an example of a rule-violated version of “statement” rather than a completely unknown word or pseudoword. Therefore, in this L2 group, only a P600 was elicited by these stimuli. Compared with the L1 study of Leinonen et al. (2008), we did not find an N400 component for this group of L2 learners. This may be due to the experimental stimuli. In Leinonen et al. (2008), the N400 elicited was taken as an index reflecting a morphological parsing processing. However, with respect to their construction of the pseudo-derived words, it is impossible to exclude the possibility of the difficulty in lexico-semantic integration. The pseudo-derived words were created by combining adjectival suffixes with a verb stem, which were not only grammatically illformed, but also semantically hard to interpret. What's more important is that the verbal roots used in their experiment were low-frequency words. Therefore, it is possible that their native speaker participants treated the pseudo-derived words as nonwords and had difficulty in online semantic integration. Thus, there is a possibility that the enhanced N400 component reflected semantic integration at the word-meaning level. In brief, in the present study, the P600 found for the group with higher morphological knowledge seems to suggest that these participants rely much more on the rule-governed decomposition mechanism to process L2 derived words. The present findings raise the question of why the same derived words are processed differently for learners with low versus high morphological knowledge. The derived words selected for the present study were highly productive (Kuo &

T. Deng et al. / Journal of Neurolinguistics 37 (2016) 47e57

55

Anderson, 2006). Previous masked priming studies consistently show the significant priming effects of these types of words (Diependaele, Sandra, & Grainger, 2005, 2009; Feldman, O'Connor, & Moscoso del Prado Martín, 2009; Morris, Franck, Grainger, & Holcomb, 2007). Therefore it may not be the linguistic properties of these words that decide whether they are processed as whole words or via the constituent morphemes. The participants in the present study were well-matched on their proficiency but differed in morphological knowledge. Hence, it may not be the proficiency that determines the different ERP patterns between these two groups but rather morphological knowledge that affects processing of L2 derived words. Learners may strengthen their understanding of the L2 word formation rules through their reading or exposure to the print. Each time they get access to morphologically complex words such as darkness, they have the opportunity to see the subcomponents as dark and ness. This process may help them learn about the word formation rules and also strengthen their understanding of the relationship between morphemes. Thus, the more they are exposed to derived words, the greater their morphological knowledge becomes and the more likely they will be to employ decomposition mechanisms. For our participants with lower morphological knowledge, their rule-based decomposition has not reached the level as high as the comparison group, so they have to rely on a whole-word processing mechanism. We acknowledge that this is our indirect inference on the performance of the group with high morphological knowledge. Future study should directly investigate the relationship between reading exposure and morphological knowledge. According to the declarative/procedural model proposed by Ullman, processing aspects of L1 morphologically complex words depend largely upon procedural memory, and native speakers can easily decompose morphologically complex words into their respective stem and suffixes (Ullman, 2004; Ullman et al., 2005). However, in a later-acquired second language, it is argued that later exposure to language may impair the ability of the procedural memory to learn or compute aspects of morphologically complex words. L2 learners may have to depend largely on declarative memory in L2. According to this model, the shift of dependence from declarative to procedural memory is expected to be a function of language proficiency, and rule-based decomposition similar to native speakers can be gradually realized. The results from our study suggest that, in addition to proficiency, the morphological knowledge that L2 learners have acquired is also an important variable modulating how L2 derived words are processed. For our participants with lower morphological knowledge, the significant negativity in the 600e900 ms time window was broadly distributed, similar to their N400 in the earlier time window. It may have been a continuation of the N400. Also, it may be reflecting the distinct neural mechanisms of derivational processing between the two groups. In a previous ERP study, Chinese learners of English showed a later negativity-named an N600-to the violations of subject-verb agreement (Chen et al., 2007). Compared with the P600 elicited in native speakers, the N600 in Chinese learners of English was taken as evidence that processing of subject-verb agreement by L2 learners was sufficiently different from native speakers. Chinese is quite different from Indo-European languages (Li, Bates, & MacWhinney, 1993; Li et al., 2004). For example, the Chinese language does not have any inflectional morphology to mark number or gender (Chen et al., 207). This lack of grammatical morphology may have consequences for native Chinese speakers acquiring L2 grammar, forcing reliance on different processing strategies for inflectional processing, even with increased proficiency in L2. Similarly, compared with Indo-European languages, Chinese has limited derivational morphology (Li et al., 2004; Zhang & Koda, 2014). In some cases, derivational morphology is not necessary to turn a noun into a verb or a verb into a noun (Li et al., 2004). Such language-specific characteristics may pose an obstacle for native Chinese learners of English trying to acquire derivational morphology, especially when they have relatively low morphological knowledge about the language. Therefore, our ERP results may be evidence that different neural mechanisms underlie different processing strategies across these two morphological knowledge levels. Although the findings of the present study indicate the important role of morphological knowledge in morphological processing, including a comparison with native speakers could provide a clearer picture. Therefore, in the future, we suggest investigating whether L2 learners with high morphological knowledge use native-like rule-based processing by directly comparing their performance to that of native speakers. In conclusion, to our knowledge, this is the first study to investigate the effect of morphological knowledge on the processing L2 derivational words. The results indicated that L2 learners with high morphological knowledge manifest rule-based decomposition, whereas L2 learners with low morphological knowledge rely more on lexical storage. These results provide strong evidence that morphological knowledge is an important variable modulating the neural mechanism of L2 derived word processing. Acknowledgments This research was supported by the Open Fund of State Key Lab of Cognitive Neuroscience and Learning (CNLYB1309) to Baoguo Chen. References Anderson, S. R. (1982). Where's morphology? Linguistic Inquiry, 13, 571e612. http://www.jstor.org/stable/4178297. Bickel, B., & Nichols, J. (2007). Inflectional morphology. In T. Shopen (Ed.), Grammatical categories and the lexicon: Vol. 3. Language typology and syntactic description (pp. 169e240). Cambridge: Cambridge University Press. Bowers, P. N., Kirby, J. R., & Deacon, S. H. (2010). The effects of morphological instruction on literacy skills: a systematic review of the literature. Review of Educational Research, 80(2), 144e179. http://dx.doi.org/10.3102/0034654309359353.

56

T. Deng et al. / Journal of Neurolinguistics 37 (2016) 47e57

Bozic, M., Tyler, L. K., Su, L., Wingfield, C., & Marslen-Wilson, W. D. (2013). Neurobiological systems for lexical representation and analysis in English. Journal of Cognitive Neuroscience, 25, 1678e1691. http://dx.doi.org/10.1162/jocn_a_00420. Brysbaert, M., & New, B. (2009). Moving beyond Ku cera and Francis: a critical evaluation of current word frequency norms and the introduction of a new and improved word frequency measure for American English. Behavior Research Methods, 41(4), 977e990. http://dx.doi.org/10.3758/BRM.41.4.977. Carlisle, J. F. (2000). Awareness of the structure and meaning of morphologically complex words: Impact on reading. Reading and Writing, 12, 169e190. http://dx.doi.org/10.1023/A:1008131926604. Carver, L. F., Vafaei, A., Guerra, R., Freire, A., & Phillips, S. P. (2013). Gender differences: examination of the 12-item Bem Sex Role Inventory (BSRI-12) in an older Brazilian population. PLoS ONE, 8(10), e76356. http://dx.doi.org/10.1371/journal.pone.0076356. Chen, L., Shu, H., Liu, Y., Zhao, J., & Li, P. (2007). ERP signatures of subjecteverb agreement in L2 learning. Bilingualism: Language and Cognition, 10, 161e174. http://dx.doi.org/10.1017/S136672890700291X. Clahsen, H., Felser, C., Neubauer, K., Sato, M., & Silva, R. (2010). Morphological structure in native and nonnative language processing. Language Learning, 60(1), 21e43. http://dx.doi.org/10.1111/j.1467-9922.2009.00550.x. Clark, M. D., Gilbert, G., & Anderson, M. L. (2011). Morphological knowledge and decoding skills of deaf readers. Psychology, 2, 109. http://dx.doi.org/10.4236/ psych.2011.22018. s, N., Diaz, B., & Rodriguez-Fornells, A. (2005). Morphological processing in early bilinguals: an ERP study of regular De Diego Balaguer, R., Sebasti an-Galle and irregular verb processing. Cognitive Brain Research, 25(1), 312e327. http://dx.doi.org/10.1016/j.cogbrainres.2005.06.003. Deacon, S. H., Benere, J., & Pasquarella, A. (2013). Reciprocal relationship: Children's morphological awareness and their reading accuracy across grades 2 to 3. Developmental Psychology, 49(6), 1113e1126. http://dx.doi.org/10.1037/a0029474. Deacon, S. H., Campbell, E., Tamminga, M., & Kirby, J. (2010). Seeing the harm in harmed and harmful: morphological processing in grades 4, 6, and 8. Applied Psycholinguistics, 31, 759e775. http://dx.doi.org/10.1017/S0142716410000238. Deacon, H., Parrila, R., & Kirby, J. R. (2006). Processing of derived forms in high-functioning dyslexics. Annals of Dyslexia, 56, 103e128. http://dx.doi.org/10. 1007/s11881-006-0005-3. ~ abeitia, J. A., Morris, J., & Keuleers, E. (2011). Fast morphological effects in first and second language word recognition. Journal of Diependaele, K., Dun Memory and Language, 64, 344e358. http://dx.doi.org/10.1016/j.jml.2011.01.003. Diependaele, K., Sandra, D., & Grainger, J. (2005). Masked cross-modal morphological priming: unraveling morpho-orthographic and morpho-semantic influences in early word recognition. Language and Cognitive Processes, 20, 75e114. http://dx.doi.org/10.1080/01690960444000197. Diependaele, K., Sandra, D., & Grainger, J. (2009). Semantic transparency and masked morphological priming: the case of prefixed words. Memory & Cognition, 37, 895e908. http://dx.doi.org/10.3758/MC.37.6.895. Feldman, L. B., & Bentin, S. (1994). Morphological analysis of disrupted morphemes: evidence from Hebrew. The Quarterly Journal of Experimental Psychology, 47(2), 407e435. http://dx.doi.org/10.1080/14640749408401118. Feldman, L. B., O'Connor, P. A., & del Prado Martín, F. M. (2009). Early morphological processing is morphosemantic and not simply morpho-orthographic: a violation of form-then-meaning accounts of word recognition. Psychonomic Bulletin & Review, 16(4), 684e691. http://dx.doi.org/10.3758/PBR.16.4.684. Friederici, A. D. (2002). Towards a neural basis of auditory sentence processing. Trends in Cognitive Sciences, 6(2), 78e84. http://dx.doi.org/10.1016/S13646613(00)01839-8. Friederici, A. D., & Weissenborn, J. (2007). Mapping sentence form onto meaning: the syntaxesemantic interface. Brain Research, 1146, 50e58. http://dx.doi. org/10.1016/j.brainres.2006.08.038. Gale-Ross, R., Baird, A., & Towson, S. (2009). Gender role, life satisfaction, and wellness: Androgyny in a southwestern Ontario sample. Canadian Journal on Aging, 28(2), 135e146. http://dx.doi.org/10.1017/S0714980809090187. Gor, K., & Cook, S. (2010). Nonnative processing of verbal morphology: In search of regularity. Language Learning, 60(1), 88e126. http://dx.doi.org/10.1111/j. 1467-9922.2009.00552.x. Hahne, A. (2001). What's different in second-language processing? Evidence from event-related brain potentials. Journal of Psycholinguistic Research, 30(3), 251e266. http://dx.doi.org/10.1023/A:1010490917575. Hahne, A., & Friederici, A. D. (2001). Processing a second language: late learners' comprehension mechanisms as revealed by event-related brain potentials. Bilingualism: Language and Cognition, 4, 123e141. http://dx.doi.org/10.1017/S1366728901000232. Hahne, A., Mueller, J. L., & Clahsen, H. (2006). Morphological processing in a second language: behavioral and event-related brain potential evidence for storage and decomposition. Journal of Cognitive Neuroscience, 18(1), 121e134. http://dx.doi.org/10.1162/089892906775250067. Havas, V., Todriguez-Fornells, V., & Clahsen, H. (2012). Brain potentials for derivational morphology: an ERP study of deadjectival nominalizations in Spanish. Brain & Language, 120, 332e344. http://dx.doi.org/10.1016/j.bandl.2011.10.008. Hernandez, A., & Li, P. (2007). Age of acquisition: Its neural and computational mechanisms. Psychological Bulletin, 133(4), 1e13. http://dx.doi.org/10.1037/ 0033-2909.133.4.638. Hernandez, A., Li, P., & MacWhinney, B. (2005). The emergence of competing modules in bilingualism. Trends in Cognitive Sciences, 9, 220e225. http://dx.doi. org/10.1016/j.tics.2005.03.003. Institute of Language Teaching and Research. (1986). A frequency dictionary of modern Chinese. Beijing: Beijing Language Institute Press. Janssen, U., Wiese, R., & Schlesewsky, M. (2006). Electrophysiological responses to violations of morphosyntactic and prosodic features in derived German nouns. Journal of Neurolinguistics, 19(6), 466e482. http://dx.doi.org/10.1016/j.jneuroling.2006.04.002. Johnson, J. S., & Newport, E. L. (1989). Critical period effects in second language learning: the influence of maturational state on the acquisition of English as a second language. Cognitive Psychology, 21(1), 60e99. http://dx.doi.org/10.1016/0010-0285(89)90003-0. Kielar, A., & Joanisse, M. F. (2011). The role of semantic and phonological factors in word recognition: an ERP cross-modal priming study of derivational morphology. Neuropsychologia, 49(2), 161e177. http://dx.doi.org/10.1016/j.neuropsychologia.2010.11.027. Kirby, J. R., Deacon, S. H., Bowers, P. N., Izenberg, L., Wade-Woolley, L., & Parrila, R. (2012). Children's morphological awareness and reading ability. Reading and Writing, 25(2), 389e410. http://dx.doi.org/10.1007/s11145-010-9276-5. http://dx.doi.org/10.1007/s11145-010-9276-5. Ku, Y. M., & Anderson, R. C. (2003). Development of morphological awareness in Chinese and English. Reading and Writing, 16, 399e422. http://dx.doi.org/ 10.1023/A:1024227231216. Kuo, L., & Anderson, R. C. (2006). Morphological awareness and learning to read: a cross-language perspective. Educational Psychologist, 41, 161e180. http:// dx.doi.org/10.1207/s15326985ep4103_3. Kutas, M., & Federmeier, K. D. (2011). Thirty years and counting: Finding meaning in the N400 component of the event-related brain potential (ERP). Annual Review of Psychology, 62, 621e647. http://dx.doi.org/10.1146/annurev.psych.093008.131123. Lavric, A., Clapp, A., & Rastle, K. (2007). ERP evidence of morphological analysis from orthography: a masked priming study. Journal of Cognitive Neuroscience, 19, 866e877. http://dx.doi.org/10.1162/jocn.2007.19.5.866. Lehtonen, M., Niska, H., Wande, E., Niemi, J., & Laine, M. (2006). Recognition of inflected words in a morphologically limited language: frequency effects in monolinguals and bilinguals. Journal of Psycholinguistic Research, 35(2), 121e146. http://dx.doi.org/10.1007/s10936-005-9008-1. €c, M., & Krausea, C. M. (2008). Event-related potential (ERP) responses to violations of inflectional and derivational rules Leinonen, A., Bratticoa, P., J€ arvenp€ aa of Finnish. Brain Research, 1218, 181e193. http://dx.doi.org/10.1016/j.brainres.2008.04.049. Liang, L. J., & Chen, B. G. (2014). Processing morphologically complex words in second-language learners: the effect of proficiency. Acta Psychologica, 150, 69e79. http://dx.doi.org/10.1016/j.actpsy.2014.04.009. Li, P., Bates, E., & MacWhinney, B. (1993). Processing a language without inflections: a reaction time study of sentence interpretation in Chinese. Journal of Memory and Language, 32, 169e192. http://dx.doi.org/10.1006/jmla.1993.1010. Li, P., Jin, Z., & Tan, L. (2004). Neural representations of nouns and verbs in Chinese: an fMRI study. NeuroImage, 21, 1533e1541. http://dx.doi.org/10.1016/j. neuroimage.2003.10.044.

T. Deng et al. / Journal of Neurolinguistics 37 (2016) 47e57

57

Li, L., & Wu, X. (2015). Effects of metalinguistic awareness on Reading comprehension and the mediator role of reading fluency from grades 2 to 4. PloS ONE, 10(3), e0114417. http://dx.doi.org/10.1371/journal.pone.0114417. Lück, M., Hahne, A., & Clahsen, H. (2006). Brain potentials to morphologically complex words during listening. Brain Research, 1077, 144e152. http://dx.doi. org/10.1016/j.brainres.2006.01.030. McBride-Chang, C., Tardif, T., Cho, J. R., Shu, H., Fletcher, P., Stokes, S. F., et al. (2008). What's in a word? Morphological awareness and vocabulary knowledge in three languages. Applied Psycholinguistics, 29, 437e462. http://dx.doi.org/10.1017/S014271640808020X. McBride-Chang, C., Wagner, R. K., Muse, A., Chow, B., & Hua, S. (2005). The role of morphological awareness in children's vocabulary acquisition in English. Applied Psycholinguistics, 26, 415e435. http://dx.doi.org/10.1017/S014271640505023X. McKinnon, R., Allen, M., & Osterhout, L. (2003). Morphological decomposition involving non-productive morphemes: ERP evidence. Neuroreport, 14(6), 883e886. http://dx.doi.org/10.1097/01.wnr.0000070192.28954.7e. Morgan-Short, K., Sanz, C., Steinhauer, K., & Ullman, M. T. (2010). Second language acquisition of gender agreement in explicit and implicit training conditions: an event-related potential study. Language Learning, 60, 154e193. http://dx.doi.org/10.1111/j.1467-9922.2009.00554.x. Morgan-Short, K., Steinhauer, K., Sanz, C., & Ullman, M. T. (2012). Explicit and implicit second language training differentially affect the achievement of native-like brain activation patterns. Journal of Cognitive Neuroscience, 24, 933e947. http://dx.doi.org/10.1162/jocn_a_00119. Morris, J., Franck, T., Grainger, J., & Holcomb, P. J. (2007). Semantic transparency and masked morphological priming: an ERP investigation. Psychophysiology, 44, 506e521. http://dx.doi.org/10.1111/j.1469-8986.2007.00538.x. Morris, J., & Holcomb, P. J. (2005). Event-related potentials to violations of inflectional verb morphology in English. Cognitive Brain Research, 25, 963e981. http://dx.doi.org/10.1016/j.cogbrainres.2005.09.021. Nagy, W. E., Anderson, R., Schommer, M., Scott, J., & Stallman, A. (1989). Morphological families in the internal lexicon. Reading Research Quarterly, 24, 262e282. http://dx.doi.org/10.2307/747770. Nagy, W. E., Berninger, V. W., & Abbott, R. D. (2006). Contributions of morphology beyond phonology to literacy outcomes of upper elementary and middleschool students. Journal of Educational Psychology, 98, 134e147. http://dx.doi.org/10.1037/0022-0663.98.1.134. Nagy, W. E., Carlisle, J. F., & Goodwin, A. P. (2014). Morphological knowledge and literacy acquisition. Journal of Learning Disabilities, 47, 3e12. http://dx.doi. org/10.1177/0022219413509967. NeubKauer, K., & Clahsen, H. (2009). Decomposition of inflected words in a second language. Studies in Second Language Acquisition, 31, 403e435. http://dx. doi.org/10.1017/S0272263109090354. Ojima, S., Nakata, H., & Kakigi, R. (2005). An ERP study on second language learning after childhood: effects of proficiency. Journal of Cognitive Neuroscience, 17, 1212e1228. http://dx.doi.org/10.1162/0898929055002436. Pinker, S. (1999). Words and rules: The ingredients of grammar. New York: Basic Books. Pinker, S., & Ullman, M. T. (2002). The past and future of the past tense. Trends in Cognitive Sciences, 6, 456e463. http://dx.doi.org/10.1016/S1364-6613(02) 01990-3. Pliatsikas, C., & Marinis, T. (2013). Processing of regular and irregular past tense morphology in highly proficient second language learners of English: a selfpaced reading study. Applied Psycholinguistics, 34, 943e970. http://dx.doi.org/10.1017/S0142716412000082. Silva, R., & Clahsen, H. (2008). Morphologically complex words in L1 and L2 processing: evidence from masked priming experiments in English. Bilingualism: Language and Cognition, 11, 245e260. http://dx.doi.org/10.1017/S1366728908003404. Stump, G. T. (1998). Inflection. In A. Spencer, & A. M. Zwicky (Eds.), The handbook of morphology (pp. 13e43). Oxford & Malden, MA: Blackwell. Tanner, D., McLaughlin, J., Herschensohn, J., & Osterhout, L. (2013). Individual differences reveal stages of L2 grammatical acquisition: ERP evidence. Bilingualism, Language and Cognition, 16, 367e382. http://dx.doi.org/10.1017/S1366728912000302. Ullman, M. T. (2001). The neural basis of lexicon and grammar in first and second language: the declarative/procedural model. Bilingualism: Language and Cognition, 4, 105e122. http://dx.doi.org/10.1017/S1366728901000220. Ullman, M. T. (2004). Contributions of memory circuits to language: the declarative/procedural model. Cognition, 92, 231e270. http://dx.doi.org/10.1016/j. cognition.2003.10.008. Ullman, M. T. (2005). A cognitive neuroscience perspective on second language acquisition: the declarative/procedural model. In C. Sanz (Ed.), Mind and context in adult second language acquisition: Methods, theory and practice (pp. 141e178). Washington, DC: Georgetown University Press. Ullman, M. T. (2012). The declarative/procedural model. In P. Robinson (Ed.), Routledge encyclopedia of second language acquisition. New York & London: Routledge. Ullman, M. T., Pancheva, R., Love, T., Yee, E., Swinney, D., & Hickok, G. (2005). Neural correlates of lexicon and grammar: evidence from the production, reading, and judgment of inflection in aphasia. Brain and Language, 93, 185e238. http://dx.doi.org/10.1016/j.bandl.2004.10.001. Walter, J. A., Wood, A., & D’zatko, K. W. (2009). The influence of morphological awareness on the literacy development of first-grade children. Language, Speech, and Hearing Services in Schools, 40, 286e298. http://dx.doi.org/10.1044/0161-1461(2009/08-0001). Wang, M., Cheng, C., & Chen, S. (2006). Contribution of morphological awareness to Chinese-English biliteracy acquisition. Journal of Educational Psychology, 98, 542e553. http://dx.doi.org/10.1037/0022-0663.98.3.542. Zhang, D., & Koda, K. (2014). Awareness of derivation and compounding in Chinese-English biliteracy acquisition. International Journal of Bilingual Education and Bilingualism, 17, 55e73. http://dx.doi.org/10.1080/13670050.2012.736949.