Is Vocabulary Knowledge Sufficient for Word-Meaning Inference? An Investigation of the Role of Morphological Awareness in Adult L2 Learners of Chinese

Is Vocabulary Knowledge Sufficient for Word-Meaning Inference? An Investigation of the Role of... Abstract This study examined whether vocabulary knowledge is sufficient for second language (L2) word-meaning inferencing. Specifically, it investigated the role of the learner’s sensitivity to a word’s morphological structure (referred to ‘morphological awareness’ or ‘MA’ hereafter) as additional support that enhances multi-character word-meaning inferencing in adult learners of Chinese as a second language. Two hypotheses were tested: (i) L2 vocabulary knowledge is the sole predictor of word-meaning inferencing involving morphologically simple multi-character words. (ii) L2 vocabulary knowledge and MA jointly contribute to word-meaning inferencing involving morphologically complex multi-character words. A set of paper-and-pencil tests was administered to 56 English-speaking learners of L2 Chinese, measuring their L2 vocabulary knowledge, MA, and word-meaning inferencing ability. The results are as follows: L2 vocabulary knowledge contributed directly and consistently to inferring the meanings of multi-character words. Yet, L2 vocabulary knowledge was not the sole predictor of L2 word-meaning inferencing; L2 MA made an additional unique contribution. Last, L2 MA contributed indirectly to L2 word-meaning inferencing through the mediation of L2 vocabulary knowledge. 1. INTRODUCTION One of the main purposes of foreign language instruction and adult second language (L2) learning is to equip learners with the skills to learn new words incidentally through reading, listening, and viewing. Skilled L2 readers are presumed to be adept at making inferences when encountering an unfamiliar word based on the information available in the word and the information afforded by the surrounding context in which the word appears. To do so, they rely heavily on L2 linguistic knowledge—vocabulary knowledge, in particular. Vocabulary knowledge is broadly defined as a learner’s knowledge of words, including knowing its form, meaning, as well as use (Nation 1990, 2001). Research has consistently shown that vocabulary knowledge is closely related to word-meaning inference among L2 readers (Nassaji 2004; Qian 2005; Hatami and Tavakoli 2012). Given that morphologically complex words are prevalent in reading materials for adult learners, a critical question is whether vocabulary knowledge alone is sufficient for supporting successful L2 word-meaning inferencing. Previous studies have indicated that morphological awareness (MA), learners’ sensitivity to the morphological structure of words (Koda 2000; Carlisle 2000; Verhoeven and Perfetti 2011), may compensate for low vocabulary knowledge in L2 reading development (Parel 2004; Gilbert et al. 2013). But findings varied among L1 and L2 reading research that considered the relative contributions of MA and vocabulary knowledge to reading subskills development. Some researchers held that MA make a significant contribution over and above vocabulary knowledge (e.g. L1: Carlisle 2000; Ku and Anderson 2003; Nagy et al. 2006; L2: Kieffer and Lesaux 2008; Jeon 2011), whereas others found that the contribution of MA is mediated fully or partially mediated by vocabulary knowledge (Zhang and Koda 2012; Kieffer and Lesaux 2012a; Goodwin et al. 2013; Kieffer et al. 2013; Deacon et al. 2014). Notably, the aforementioned insights have been derived mainly from studies focusing on reading comprehension as the outcome and English as the target language. It remains unclear as to how L2 MA and L2 vocabulary knowledge could jointly contribute to L2 word-meaning inferencing across typologically different languages. This study investigated the role of the learner’s sensitivity to a word’s morphological structure (referred to ‘morphological awareness/MA’ hereafter) as additional support that enhances word-meaning inferencing in adult learners of Chinese as a second language. Previous studies on L2 lexical inferencing typically followed a design that aimed to identify the proportional distributions of MA and vocabulary knowledge, as two different resources, to successful L2 word-meaning inferencing (Nassaji 2003), and little research has taken a component approach to modeling how L2 MA and L2 vocabulary knowledge may interact to predict L2 word-meaning inferencing ability. In this research, we adopted a component approach toward L2 word-meaning inferencing during reading in this research. Following Bialystok (2001) and Koda (2007), we distinguish MA (a facet of metalinguistic awareness, that is, sensitivity to the abstract structure of language) from vocabulary knowledge (knowledge of specific words), and word-meaning inferencing as a reading subskill. The objectives of the study were twofold: (i) determining whether L2 vocabulary knowledge contributes directly and consistently to inferring the meanings of different types of multi-character words, and (ii) exploring the intralingual relationship between L2 vocabulary knowledge and L2 MA in their contributions L2 word-meaning inferencing. The writing system and morphological properties of Chinese provide an ideal opportunity for this investigation: (i) Chinese employs a morphosyllabic writing system, in which each graphic symbol encodes a morpheme (its syllable) rather than a phoneme. (ii) The majority of Chinese words are written with multiple characters. (iii) The number of characters (morphemes) varies from word to word (most words consist of two to four characters), and (iv) the word boundaries and the morphological boundaries within a word are not visually marked because a space is not used in Chinese writing. What follows first includes a review of the respective roles of vocabulary knowledge and MA in L2 word-meaning inferencing, the intralingual relationship between the two, and the unique morphological properties of printed Chinese word. 2. LITERATURE REVIEW 2.1 L2 vocabulary knowledge and L2 word-meaning inferencing Word-meaning inferencing or lexical inferencing is often referred to as ‘making informed guesses to the meaning of a word in the light of all available linguistic cues in combination with the learners’ general knowledge of the world, [his/her] awareness of the co-text and [his/her] relevant linguistic knowledge’ (Haastrup 1991: 11). Previous studies of L2 word-meaning inferencing have identified L2 vocabulary knowledge as the critical predictor of inferring unknown word meanings successfully in L2 readers (Nassaji 2004; Qian 2005; Hatami and Tavakoli 2012). The relationship between vocabulary knowledge and word-meaning inferencing among L2 learners was examined with regard to the relative contributions of two subcomponents of vocabulary knowledge, namely, vocabulary size and vocabulary depth. Vocabulary size is often referred to as the quantity or the number of words one knows (Nation 2001), while vocabulary depth relates to how well one knows a word (Read 1993, 2000; Meara 1996). Although vocabulary size is often considered to play a critical role in unsupported reading, which indicates the need for a very high vocabulary coverage rate for successful reading comprehension, ranging between 95% (Liu and Nation 1985) and 98% (Hu and Nation 2000; Nation 2006; Schmitt et al. 2012), vocabulary depth was found to be more important in L2 word-meaning inferencing. For instance, Hatami and Tavakoli (2012) measured both vocabulary size and depth in Persian-speaking university learners of English as a foreign language (EFL), and compared their relative contributions to word-meaning inferencing. Their results suggested that vocabulary depth contributed more to L2 word-meaning inferencing than did vocabulary size. In other studies that focused on the impact of vocabulary depth on L2 word-meaning inferencing, a positive correlation was observed as well (Nassaji 2004; Qian 2005). Qian (2005) postulated that there might be a Matthew effect between the two, as profound vocabulary knowledge increases the chance of word-meaning inferencing success, and successful word-meaning inferencing in turn deepens L2 learners’ knowledge of words. In spite of the significance of L2 vocabulary knowledge, a question that remains is whether vocabulary knowledge alone will be sufficient for inferring unknown word meanings successfully in L2 reading. If the answer is yes, this seems to imply an all-or-nothing manner in L2 reading and learning; that is, if L2 readers cannot identify an unknown word based on their prior vocabulary knowledge, they have no other resources to infer the unknown word meaning and learn new words. Another issue is whether L2 readers’ vocabulary knowledge is sufficient for them to retrieve contextual information to infer unknown word meanings. As indicated in previous research, for L2 readers, contextual information might not be as useful as it is assumed to be because it could strengthen initial misinterpretation (Christianson and Luke 2011). On the other hand, high-quality word-internal lexical representation (a nexus of orthographical, phonological, and morphological/semantic information) will reduce reliance on context (Andrews and Bond 2009), as it has been found that word-internal morphological information is one of the resources used frequently by L2 readers in word-meaning inferencing (de Bot et al. 1997; Paribakht and Wesche 1999). Both de Bot et al. (1997) and Paribakht and Wesche (1999) examined different resources used by university intermediate-level learners’ inferencing processes in L2 English and observed that L2 learners used their knowledge of word derivations and inflections frequently to infer the meanings of unknown words. This ability to parse words and utilize word-internal morphological information should pertain to MA. 2.2 L2 MA and L2 word-meaning inferencing MA is often referred to as the ability to reflect upon and manipulate constituent morphemes in visual word processing for the purpose of meaning construction (Carlisle 2000; Kuo and Anderson 2008). Considering the amount of studies of the effects of L2 vocabulary knowledge on L2 word-meaning inferencing, the role of L2 MA is underexamined in extant literature. Yet, there has been evidence suggesting that L2 MA makes unique and independent contributions to guessing unknown word meanings during reading. For instance, Park (2004) measured first language (L1) and L2 MA of Korean-speaking English language learners at third to fifth grades and evaluated their relationships with L2 word-meaning inferencing. It was found that L2 MA not only mediated the contribution of L1 MA to L2 word-meaning inferencing, but also had a direct impact on L2 word-meaning inferencing. Park held that the utility of MA lies in that it helps with the extraction of partial information from unfamiliar words, which promotes analytical approaches to lexical processing in reading, even among L2 readers with low L2 linguistic knowledge. In other studies that did not include any MA measure, the utility of MA was inferred from the effects of unknown word morphological properties on successful L2 word-meaning inferencing in learners of higher/lower L2 linguistic knowledge (Mori and Nagy 1999; Hamada 2014). In Mori and Nagy’s (1999) study with English-speaking university learners of L2 Japanese at intermediate and preadvanced levels, they asked learners to infer the meanings of novel semantically semi-transparent compound words in a sentence. Compound words comprised familiar kanjis in three conditions (i.e. kanji/word-internal information only, contextual information only, and the combination of kanji and contextual information). It was found that: (i) more than half of L2 Japanese learners tended to integrate both word-internal and contextual information, and integrators were more successful in word-meaning inferencing than were non-integrators; and (ii) L2 linguistic knowledge correlated with the use of contextual clues but not with the use of word-internal information. Their findings with intermediate- and preadvanced-level Japanese learners indicate that MA seems to be an independent resource from contextual information extraction and less constrained by learners’ L2 linguistic knowledge. More recently, Hamada (2014) investigated whether the use of morphological/contextual information in word-meaning inferencing depends upon the learners’ L2 proficiency and the reliability of the morphological information, with university learners of L2 English at four proficiency levels (beginning, intermediate, high-intermediate, and advanced). Two types of unknown word were included in Hamada’s (2014) study: morphology reliable pseudo compounds (the known word part provided semantic information that matched sentential context) and morphology unreliable compound (the known word part provided semantic information that did not match sentential context). The findings suggested that the choice of information was influenced by the morphological reliability condition because participants performed similarly across proficiency levels in the morphology reliable condition but differently in the morphology unreliable condition (the beginning group used morphological information more often than other proficiency groups). The observations of Mori and Nagy’s (1999) and Hamada’s (2014) indicate that the utility of MA in L2 word-meaning inferencing is independent from L2 linguistic knowledge, but dependent on the morphological structure of a word. Notably, both studies were based on learners’ overall L2 proficiency and did not include any measure of L2 vocabulary knowledge. 2.3 L2 MA, L2 vocabulary knowledge, and L2 word-meaning inferencing When the intralingual relationship between MA, vocabulary knowledge, and word-meaning inferencing is considered, other than the close relationship between vocabulary knowledge and word-meaning inferencing, and that between MA and word-meaning inferencing, the relationship between MA and vocabulary knowledge should be noted as well. Previous L1 reading studies have shown that MA contributes significantly to vocabulary acquisition in different languages (including English, Chinese, and Korean) (Anglin et al. 1993; McBride-Chang et al. 2008; Kieffer and Lesaux 2012b). Anglin et al. (1993) observed a significant increase of derived word knowledge between first- and fifth-grade native English-speaking children, and found that it was through ‘morphological problem solving’ that children improved in their word learning. In a longitudinal study, McBride-Chang et al. (2008) tracked the performance of Chinese-speaking and Korean-speaking monolingual preschoolers over spans of nine months to one year, and found that early MA significantly predicted vocabulary development. Given the close relationship between MA and vocabulary knowledge, and that between vocabulary knowledge and word-meaning inferencing reviewed previously, there is a possibility that L2 MA contributes indirectly to L2 word-meaning inferencing through L2 vocabulary knowledge. However, to our knowledge, to date, there is a lack of research that directly investigates the intralingual relationship between MA and vocabulary knowledge in predicting L2 word-meaning inferencing. Notably, when examining the role of L2 MA in L2 reading, additional attention should be paid to the potential impact of word properties in the target language. In this regard, what follows is a discussion of the case of (Mandarin) Chinese, a language that has attracted relatively less attention in L2 reading acquisition. 2.4 Words, morphemes, and MA in Chinese Following Packard (2000: 12), a word is defined as ‘an independent occupant of a syntactic form class slot’. Written words in English are salient because they are separated by spaces in written texts. In Chinese, however, written texts are not word-based but character-based, with no spaces inserted between words. Emerging evidence suggests that words, as defined by Packard, do have psychological reality for both L1 and L2 reading in Chinese for a number of reasons: (i) identifying characters within a word increases accuracy as opposed to identifying characters in a string of characters that do not constitute a word (Chen 1999); (ii) inserting spaces between characters within words hinders the reading of Chinese texts, whereas inserting spaces between words does not (Bai et al. 2008); and (3) word properties have effects on Chinese reading that go beyond the characters’ properties (Li et al. 2014) (for a review, see Liu et al. 2013). Notably, words in Chinese vary in the numbers of characters that comprise them: 6% are single-character words, 72% are two-character words, 12% are three-character words, and 10% are four-character words (as cited in Li et al. 2014). Because the numbers of two-character and three-character words significantly exceed the number of single-character words, the examination of Chinese words focuses on multi-character words henceforth. Since multiple terms have been used to refer to concepts that involve meaning in Chinese, including character, morpheme, and word, it is important to note the similarities and differences among these concepts. Morphemes in Chinese fall into two categories—grammatical morphemes (e.g. 了, le, as an aspect marker) and word formation morphemes (e.g. 化, huà, resembles the affix -ize in English) (Packard 2000). As mentioned before, each character can usually represent one morpheme, but the two are not identical, since some characters can shift in pronunciation, meaning, or both from context to context (see Myers 2006). An example cited by Myers (2006) is the first morpheme of 行人 (xíngrén, ‘pedestrian’ or ‘walk-person’ literally) and the second morpheme of 银行 (yínháng, ‘bank’ or ‘silver-store’ literally), written with the same character. As well, there has been some misconceptualization, assuming that all multi-character words are compound or multi-morphemic words. For instance, each character in the two-character word 花生 has its own meanings (‘flower/spend’ for 花 and ‘born/student’ for 生, respectively) and can be used independently and functionally; yet the two-character word 花生is a lexicalized expression mapping onto one morpheme, meaning ‘peanut’, the meaning of which cannot be directly inferred from the two-component characters. Another oft-cited notion related to morpheme in written Chinese is semantic radical, which is a subcomponent within a Chinese character that denotes meaning. The present study did not consider the role of semantic radical because the focus was on multi-character rather than single-character words, and the semantic cues provided by semantic radicals are not always reliable (Shu et al. 2003; Chung and Leung 2008). Based on the review above, it is important for researchers to differentiate morphologically complex multi-character words from morphologically simple multi-character words when examining reading in Chinese. Previous studies with both L1 and L2 adult Chinese readers indicated that they are sensitive to the structural properties of multi-character words (Xu 2004; Lin et al. 2011). In a study comparing native Chinese adults’ and children’s performances in a word parsing task, Lin et al. (2011) found that native Chinese adult readers were more aware of word features and more capable of using relatively productive characters as word boundary cues. As for adult readers of L2 Chinese, Xu (2004) examined L2 Chinese MA among learners with no prior exposure to character learning in their L1 experience, and asserted that adult L2 Chinese readers had already developed sensitivity to productive morphemes in their first year of study in China. It is noteworthy that most of the previous studies were based on reading or recognizing known words. Taken together, although L2 vocabulary knowledge has been considered to be an important predictor for L2 word-meaning inferencing, whether L2 vocabulary knowledge alone will contribute consistently to inferring the meanings of unknown words is still an open question. As suggested by previous findings, L2 MA can compensate for insufficient L2 vocabulary knowledge. What merits further examination is how vocabulary knowledge and MA jointly contribute to L2 word-meaning inferencing. Moreover, there is a need for investigation in target languages that are typologically distinct from alphabetic languages. 3. THE PRESENT STUDY This study examined whether L2 vocabulary knowledge alone is sufficient for L2 word-meaning inferencing in Chinese as a second language. Of special interest is whether L2 vocabulary knowledge and L2 MA contribute jointly or differentially to inferring the meanings of different types of multi-character words in L2 Chinese. It aims to expand our understanding of the intricate relationships among L2 vocabulary knowledge, L2 MA, and L2 word-meaning inferencing, and the way in which L2 readers utilize word-internal properties and word-external/contextual information when guessing unknown word meanings during reading. 3.1 Hypotheses In the current study, two hypotheses were formulated regarding how L2 vocabulary knowledge could contribute to L2 word-meaning inferencing when accounting for the potential influence of L2 MA and unknown word characteristics. The first hypothesis was that L2 vocabulary knowledge would be the sole predictor of word-meaning inferencing involving morphologically simple multi-character words, and that there would be no additional contribution of L2 MA. The second was that L2 vocabulary knowledge and MA would jointly contribute to lexical inferencing involving morphologically complex multi-character words. 3.2 Methods 3.2.1 Participants In total, 56 English-speaking learners of L2 Chinese participated in the present study. They were all Chinese learners at an advanced level, which was determined by three criteria: (i) the instructional levels in the participants’ affiliated universities, (ii) a minimum of two years of Chinese learning experience, and (iii) reading as one of the major learning and instructional components. There were 27 female students and 29 male students, with an average age of 21.98 (SD = 4.33). They were recruited from six universities in Shanghai and Beijing, China. By the time of data collection, the participants had received an average of 3.2 years of formal Chinese education. They studied Chinese for 2–4 class hours per week, and were expected to have mastered 2,000 commonly used words and related grammar patterns. 3.2.2 Instruments, data collection, and analytical procedures A paper-and-pencil test package was distributed to the participants. They first completed the L2 word-meaning inferencing task, followed by the L2 MA tasks, L2 vocabulary knowledge tasks, and a word checklist to ensure the lexicality status of the target words in the inferencing task. The estimated time to complete the whole test was 1 hour. L2 word-meaning inferencing. In accordance with Haastrup’s (1991) definition, word-meaning inferencing refers to the ability to make appropriate predictions of unknown word meanings based on word-internal and contextual information. The word-meaning inferencing task was adapted from Mori and Nagy (1999). The participants were asked to infer the meaning of an unknown word underlined within a phrase or a short sentence with a relatively constrained context (which is consisted of less than 10 characters and only provides part-of-speech clue for the target unknown word), and to select the correct answer from four choices written in English based on four conditions: (i) an integrated answer combining both context and the meaning of word subcomponents; (ii) the meaning of the word subcomponent only; (iii) context only; and (iv) a distractor/anomalous answer. For instance, the phrase 全国覆盖面最广(quánguó fùgàimiàn zuìguǎng, ‘the nation’s widest coverage’) had the target unknown word 覆盖面 (fùgàimiàn, ‘coverage’) underlined, followed by four choices: (i) coverage, (ii) surface, (iii) net, and (iv) voice. If L2 Chinese readers only resort to their prior print vocabulary knowledge, they might select (ii) because 面 (miàn) means ‘side/facet’ and the meaning of 覆盖 (fùgài, ‘to cover’) is usually beyond advanced-level learners’ vocabulary. If they use contextual information only, they may select (iii) because the Chinese word 网 (wǎng) corresponding to ‘net’ in English is not relevant to any word-internal component of the target word 覆盖面, and option (iv) was a distractor. Only when the participants utilize both the word-internal morphological information and the word-external contextual meaning can they successfully infer the meaning of the underlined word and select (i). Ease with which the meaning of a target word can be inferred was manipulated in two ways: first by varying the morphological structure of the target word (i.e. morphologically complex/bimorphemic versus morphologically simple/monomorphemic) and second by varying base word familiarity (i.e. whether the participants knew the meaning of the base word). Other than the underlined unknown words, other lexical items in the phrase/sentence were all familiar to the participants (from Bands One and Two/lowest levels in the Grading Syllabus for Chinese Vocabulary and Chinese Characters, Chinese Proficiency Test Center 2001). In total, there were 32 items in the word-meaning inferencing task, with eight items for each word type (as shown in Table 1). One point was assigned for each correct answer. The Cronbach’s α was 0.75. Table 1: Sample items in L2 word-meaning inferencing task Word type n Example English meaning Monomorphemic multi-character words with familiar word part 8 分外 Very Monomorphemic multi-character words with unfamiliar word part 8 螺丝 Screw Bimorphemic multi-character words with familiar base 8 副业 Sideline Bimorphemic multi-character words with unfamiliar base 8 小范围 Small-scale Word type n Example English meaning Monomorphemic multi-character words with familiar word part 8 分外 Very Monomorphemic multi-character words with unfamiliar word part 8 螺丝 Screw Bimorphemic multi-character words with familiar base 8 副业 Sideline Bimorphemic multi-character words with unfamiliar base 8 小范围 Small-scale Note: Unfamiliar base word/word part is underlined. Table 1: Sample items in L2 word-meaning inferencing task Word type n Example English meaning Monomorphemic multi-character words with familiar word part 8 分外 Very Monomorphemic multi-character words with unfamiliar word part 8 螺丝 Screw Bimorphemic multi-character words with familiar base 8 副业 Sideline Bimorphemic multi-character words with unfamiliar base 8 小范围 Small-scale Word type n Example English meaning Monomorphemic multi-character words with familiar word part 8 分外 Very Monomorphemic multi-character words with unfamiliar word part 8 螺丝 Screw Bimorphemic multi-character words with familiar base 8 副业 Sideline Bimorphemic multi-character words with unfamiliar base 8 小范围 Small-scale Note: Unfamiliar base word/word part is underlined. As mentioned earlier, a post-test word checklist was distributed to participants to ensure that selected words’ properties conform to the aforementioned four conditions. The checklist required the participants to self-report their knowledge of the target words and provide the English meanings. If participants successfully identified the word meaning, it was considered as familiar; otherwise, the word was counted as unfamiliar or unknown. L2 MA. MA is a multi-faceted construct, and current measures of MA range from more language-specific to less language-specific (Koda Lűand Zhang 2014). L2 MA was measured by two tasks, segmentation and intraword structure analysis. Segmentation refers to the ability to decompose a word into smaller meaningful units; intraword structure analysis refers to readers’ sensitivity to intraword morphological structure (Koda 2000). Following Peng (2004), the segmentation task required the participants to split multi-character words into smaller words that still have meanings. The items included two-character and three-character words that fell into four categories (as illustrated in Table 2). The segmentation task included 32 items. A participant was credited with one point for each correct answer. The Cronbach’s α was 0.80. For the intraword structure analysis, the task was constructed after Ku (2001). The participants were asked to read three words that shared a character and to choose the word with a shared character meaning different from the other two. For example, 画家 (huàjiā, ‘painter’), 作家 (zuòjiā, ‘writer’), and 大家 (dàjiā, ‘everyone’) share the character 家 (jiā, meaning ‘house’ or ‘family,’ or ‘professional’ when used independently). 大家 is a monomorphemic word in which the shared character’s meaning is not transparent, whereas the former two are bimorphemic words meaning ‘professional in doing something’, and 家 jiā functions like the suffix ‘–er’ in ‘writer’ in English. It follows that, to reason about the meaning carried by the shared character, the participants needed to analyze the morphological structure within each word, to discover that the shared character in 大家 does not bear a specific meaning that contributes to the whole word meaning as in the other two words, and to choose the correct answer 大家. One point was credited for each correct answer. There were 16 items in the intraword structure analysis. The Cronbach’s α was 0.80. Table 2: Sample items in L2 MA segmentation task Word type n Example English meaning Two-character monomorphemic 8 花生 Peanut Two-character bimorphemic 8 医学 The study of medicine Three-character monomorphemic 8 对不起 Sorry Three-character bimorphemic 8 运动员 Sports player Word type n Example English meaning Two-character monomorphemic 8 花生 Peanut Two-character bimorphemic 8 医学 The study of medicine Three-character monomorphemic 8 对不起 Sorry Three-character bimorphemic 8 运动员 Sports player Table 2: Sample items in L2 MA segmentation task Word type n Example English meaning Two-character monomorphemic 8 花生 Peanut Two-character bimorphemic 8 医学 The study of medicine Three-character monomorphemic 8 对不起 Sorry Three-character bimorphemic 8 运动员 Sports player Word type n Example English meaning Two-character monomorphemic 8 花生 Peanut Two-character bimorphemic 8 医学 The study of medicine Three-character monomorphemic 8 对不起 Sorry Three-character bimorphemic 8 运动员 Sports player L2 vocabulary knowledge. Following Anderson and Freebody (1981), we focused on a learner’s knowledge of word forms and meanings, and measured L2 vocabulary knowledge by two tasks, semantic word knowledge and morpheme knowledge. The semantic word knowledge task was based on Liu (2013), which asked participants to translate Chinese words into English meaning (n = 60). The Cronbach’s α was 0.95. In a similar vein, the morpheme knowledge task asked the participants to indicate Chinese affixes’ meanings or functions in English. Two types of morphemes, grammatical morphemes (n = 4) and word formation morphemes (n = 20), were included. The word formation morphemes were selected from Zeng’s (2008) database of productive morphemes in Chinese, consisting of 8 prefixoids and 12 suffixoids. The Cronbach’s α was 0.85. Analytical procedures. The analysis plan entailed a range of statistical analyses to test the aforementioned hypotheses—whether L2 vocabulary knowledge and L2 MA will contribute jointly or differentially when L2 Chinese readers infer the meanings of different types of multi-character unknown words. First, a two-by-two repeated-measures analysis of variance (ANOVA) would be performed to examine if there are any effects of word characteristics on L2 word-meaning inferencing, with morphological structure and base word familiarity as within-subject variables and L2 word-meaning inferencing as the dependent variable. Second, if there were any effects of word characteristics, separate regression analyses would be carried out for different word types with L2 vocabulary knowledge and L2 MA as independent variables, and L2 word-meaning inferencing as dependent variables, which was aimed to investigate the way in which L2 vocabulary knowledge and L2 MA contribute to L2 word-meaning inferencing. 4. RESULTS 4.1 Descriptive statistics Eight participants who did not complete all the tasks were removed from the data set. Another three cases were removed for they were identified as outliners at the exploratory data analysis phase. Therefore, all subsequent analyses performed in the study included 45 valid cases of the 56. The descriptive statistics are displayed in Table 3. Table 3: Descriptive statistics for all variables (N = 45) Variable M SD L2 lexical inferencing (correct out of 32) 17.71 5.21 L2 MA composite (z score-based) 0.00 1.72 Segmentation (correct out of 32) 23.04 5.14 Intraword structure analysis (correct out of 16) 9.93 3.19 L2 vocabulary knowledge composite (z score-based) 0.00 1.94 Semantic word knowledge (correct out of 60) 38.33 10.93 Morpheme knowledge (correct out of 24) 13.64 4.52 Variable M SD L2 lexical inferencing (correct out of 32) 17.71 5.21 L2 MA composite (z score-based) 0.00 1.72 Segmentation (correct out of 32) 23.04 5.14 Intraword structure analysis (correct out of 16) 9.93 3.19 L2 vocabulary knowledge composite (z score-based) 0.00 1.94 Semantic word knowledge (correct out of 60) 38.33 10.93 Morpheme knowledge (correct out of 24) 13.64 4.52 Table 3: Descriptive statistics for all variables (N = 45) Variable M SD L2 lexical inferencing (correct out of 32) 17.71 5.21 L2 MA composite (z score-based) 0.00 1.72 Segmentation (correct out of 32) 23.04 5.14 Intraword structure analysis (correct out of 16) 9.93 3.19 L2 vocabulary knowledge composite (z score-based) 0.00 1.94 Semantic word knowledge (correct out of 60) 38.33 10.93 Morpheme knowledge (correct out of 24) 13.64 4.52 Variable M SD L2 lexical inferencing (correct out of 32) 17.71 5.21 L2 MA composite (z score-based) 0.00 1.72 Segmentation (correct out of 32) 23.04 5.14 Intraword structure analysis (correct out of 16) 9.93 3.19 L2 vocabulary knowledge composite (z score-based) 0.00 1.94 Semantic word knowledge (correct out of 60) 38.33 10.93 Morpheme knowledge (correct out of 24) 13.64 4.52 4.2 Effects of morphological structure and base word familiarity Table 4 illustrates the participants’ performance in L2 word-meaning inferencing. There seems to be no notable difference across the four word types. A two (monomorphemic versus bimorphemic) by two (familiar versus unfamiliar base) repeated-measures ANOVA was carried out to analyze the data, with morphological structure and base word familiarity as within-subject variables. The results suggest that L2 word-meaning inferencing was not significantly affected the interaction between the two (F1, 44 = 0.16, p = .69), not by morphological structure only (F1, 44 = 0.71, p = .40), nor by base word familiarity only (F1, 44 = 0.46, p = .50). Table 4: Means in L2 word-meaning inferencing as a function of word types (N = 45) Word type M SD Monomorphemic-familiar 4.62 1.60 Monomorphemic-unfamiliar 4.40 1.72 Bimorphemic-familiar 4.38 1.61 Bimorphemic-unfamiliar 4.31 2.03 Word type M SD Monomorphemic-familiar 4.62 1.60 Monomorphemic-unfamiliar 4.40 1.72 Bimorphemic-familiar 4.38 1.61 Bimorphemic-unfamiliar 4.31 2.03 Note: Familiar, familiar word part/base; unfamiliar, unfamiliar word part/base. Table 4: Means in L2 word-meaning inferencing as a function of word types (N = 45) Word type M SD Monomorphemic-familiar 4.62 1.60 Monomorphemic-unfamiliar 4.40 1.72 Bimorphemic-familiar 4.38 1.61 Bimorphemic-unfamiliar 4.31 2.03 Word type M SD Monomorphemic-familiar 4.62 1.60 Monomorphemic-unfamiliar 4.40 1.72 Bimorphemic-familiar 4.38 1.61 Bimorphemic-unfamiliar 4.31 2.03 Note: Familiar, familiar word part/base; unfamiliar, unfamiliar word part/base. As shown in Figure 1 below, the data are slightly skewed. Following Larson-Hall (2010), data transformation was conducted by squaring the scores. Two-way repeated-measures ANOVA was rerun with morphological structure and familiarity with base words as within-subject variables. Yet, the results remained the same, while post hoc checking found no notable violation of assumptions (i.e. normal distribution and equal variance of data, normal distribution and equal variance of residuals). Effect sizes were calculated according to Field (2009). When the main effect of morphological structure was computed, r was 0.13; for familiarity with base words, r was 0.10. Both were relatively small. Given that there were no statistical effects of the two focal word characteristics, the following analysis only focuses on the contributions of L2 vocabulary knowledge and L2 MA to L2 word-meaning inferencing across all types of words. Figure 1: View largeDownload slide Performance in L2 word-meaning inferencing differentiated by word types Note: Familiar, familiar word part/base; unfamiliar, unfamiliar word part/base Figure 1: View largeDownload slide Performance in L2 word-meaning inferencing differentiated by word types Note: Familiar, familiar word part/base; unfamiliar, unfamiliar word part/base 4.3 Relative contributions of L2 vocabulary knowledge and L2 MA As shown in Table 5, there was moderate correlation between L2 MA (composite) and L2 word-meaning inferencing (r = 0.46, p < .01), and high correlation between L2 vocabulary knowledge (composite) and L2 word-meaning inferencing (r = 0.80, p < .001). Notably, the correlation between L2 MA and L2 vocabulary knowledge was high (r = 0.72, p < .001). The potential interaction effects between L2 MA and L2 vocabulary knowledge were examined from the multiple regression model results using the composite scores of L2 MA and L2 vocabulary knowledge, respectively. The rationale is as follows. First, as indicated in Table 5, there was a high correlation between the two subcomponent of L2 vocabulary knowledge (i.e. morpheme knowledge and semantic word knowledge) (r = 0.84), which might be due to the task modality. They were both tested in the written format, and orthographic knowledge or character knowledge was required to perform both tasks. Second, with respect to the two subcomponent of L2 MA (segmentation and intraword structure analysis), the correlational patterns, either among themselves, or with L2 vocabulary knowledge, or with L2 word-meaning inferencing, were compatible in general: (i) the magnitude of correlation between the two was large (r = 0.55) according to Cohen’s (1988) bench mark (0.1 being small, 0.25 being medium, and 0.4 being large). (ii) The correlation between segmentation and L2 vocabulary knowledge (composite) (r = 0.57), and that between intraword analysis and L2 vocabulary knowledge (composite) (r = 0.70) can both be counted as large effect sizes, and (iii) the correlation between segmentation and L2 word-meaning inferencing (r = 0.32) was not notably different from that between intraword analysis and L2 word-meaning inferencing (r = 0.49). In this regard, a composite score of L2 MA was used in subsequent regression analyses. Table 5: Bivariate correlations among variables (N = 45) Measure 1 2 3 4 5 6 7 8 9 10 11 12 1 L2 word-meaning inferencing – 2 L2 WMIMF .68*** – 3 L2 WMIMU .79*** .39** – 4 L2 WMIBF .66*** .24 .35* – 5 L2 WMIBU .84*** .42** .59*** .40** – 6 L2 MA (composite) .46** .16 .36* .33* .48** – 7 L2 segmentation .32* .08 .20 .22 .42** .89*** – 8 L2 intraword analysis .49** .21 .43** .36* .42** .87*** .55*** – 9 L2 vocabulary knowledge (composite) .80*** .50*** .66*** .53*** .67*** .72*** .57*** .70*** – 10 Semantic word knowledge .80*** .48** .64*** .57*** .69*** .74*** .57*** .73*** .96*** – 11 Morpheme knowledge .73*** .49** .63*** .45** .60*** .64*** .52*** .62*** .96*** .84*** – 12 L2 MA (composite) × L2 vocabulary knowledge (composite) .40** .47** .33* .06 .33* .31** .31* .24 .36* .30* .39** – Measure 1 2 3 4 5 6 7 8 9 10 11 12 1 L2 word-meaning inferencing – 2 L2 WMIMF .68*** – 3 L2 WMIMU .79*** .39** – 4 L2 WMIBF .66*** .24 .35* – 5 L2 WMIBU .84*** .42** .59*** .40** – 6 L2 MA (composite) .46** .16 .36* .33* .48** – 7 L2 segmentation .32* .08 .20 .22 .42** .89*** – 8 L2 intraword analysis .49** .21 .43** .36* .42** .87*** .55*** – 9 L2 vocabulary knowledge (composite) .80*** .50*** .66*** .53*** .67*** .72*** .57*** .70*** – 10 Semantic word knowledge .80*** .48** .64*** .57*** .69*** .74*** .57*** .73*** .96*** – 11 Morpheme knowledge .73*** .49** .63*** .45** .60*** .64*** .52*** .62*** .96*** .84*** – 12 L2 MA (composite) × L2 vocabulary knowledge (composite) .40** .47** .33* .06 .33* .31** .31* .24 .36* .30* .39** – Notes: WMIMF, word-meaning inferencing of monomorphemic words with familiar bases; WMIMU, word-meaning inferencing of monomorphemic words with unfamiliar bases; WMIUF, word-meaning inferencing of bimorphemic words with familiar bases; WMIBF, word-meaning inferencing of bimorphemic words with unfamiliar bases. * p < .05; **p < .01; ***p < .001. Table 5: Bivariate correlations among variables (N = 45) Measure 1 2 3 4 5 6 7 8 9 10 11 12 1 L2 word-meaning inferencing – 2 L2 WMIMF .68*** – 3 L2 WMIMU .79*** .39** – 4 L2 WMIBF .66*** .24 .35* – 5 L2 WMIBU .84*** .42** .59*** .40** – 6 L2 MA (composite) .46** .16 .36* .33* .48** – 7 L2 segmentation .32* .08 .20 .22 .42** .89*** – 8 L2 intraword analysis .49** .21 .43** .36* .42** .87*** .55*** – 9 L2 vocabulary knowledge (composite) .80*** .50*** .66*** .53*** .67*** .72*** .57*** .70*** – 10 Semantic word knowledge .80*** .48** .64*** .57*** .69*** .74*** .57*** .73*** .96*** – 11 Morpheme knowledge .73*** .49** .63*** .45** .60*** .64*** .52*** .62*** .96*** .84*** – 12 L2 MA (composite) × L2 vocabulary knowledge (composite) .40** .47** .33* .06 .33* .31** .31* .24 .36* .30* .39** – Measure 1 2 3 4 5 6 7 8 9 10 11 12 1 L2 word-meaning inferencing – 2 L2 WMIMF .68*** – 3 L2 WMIMU .79*** .39** – 4 L2 WMIBF .66*** .24 .35* – 5 L2 WMIBU .84*** .42** .59*** .40** – 6 L2 MA (composite) .46** .16 .36* .33* .48** – 7 L2 segmentation .32* .08 .20 .22 .42** .89*** – 8 L2 intraword analysis .49** .21 .43** .36* .42** .87*** .55*** – 9 L2 vocabulary knowledge (composite) .80*** .50*** .66*** .53*** .67*** .72*** .57*** .70*** – 10 Semantic word knowledge .80*** .48** .64*** .57*** .69*** .74*** .57*** .73*** .96*** – 11 Morpheme knowledge .73*** .49** .63*** .45** .60*** .64*** .52*** .62*** .96*** .84*** – 12 L2 MA (composite) × L2 vocabulary knowledge (composite) .40** .47** .33* .06 .33* .31** .31* .24 .36* .30* .39** – Notes: WMIMF, word-meaning inferencing of monomorphemic words with familiar bases; WMIMU, word-meaning inferencing of monomorphemic words with unfamiliar bases; WMIUF, word-meaning inferencing of bimorphemic words with familiar bases; WMIBF, word-meaning inferencing of bimorphemic words with unfamiliar bases. * p < .05; **p < .01; ***p < .001. Hierarchical regression analysis was performed to examine the main effects of L2 vocabulary knowledge and L2 MA as well as the interaction between the two on L2 word-meaning inferencing. Two sets of blockwise regression were carried out. First, L2 word-meaning inferencing was regressed on the order of (i) L2 MA, (ii) L2 vocabulary knowledge, and then (iii) the interaction between L2 MA and L2 vocabulary knowledge. Second, the order of L2 MA and L2 vocabulary knowledge was reversed, with L2 vocabulary knowledge entered first, L2 MA second, and the interaction between L2 MA and L2 vocabulary knowledge last. Reversing the order of L2 vocabulary knowledge and L2 MA helped to examine the unique variance explained by the new variable entered to the model while accounting for the other, which thus illustrates the relative contributions of L2 vocabulary knowledge and L2 MA, and/or the possible interaction between the two to L2 word-meaning inferencing. In Table 6, the results from Model 1 suggest that, in Step 1, the main effect of L2 MA was significant, explaining 21% of the variance of L2 word-meaning inferencing (R2 = 0.21, F1, 43 = 11.27, p < .01). In Model 1 Step 2, when controlling for L2 MA, L2 vocabulary knowledge significantly explained an additional 46% of the variance (R2 = 0.67, p < .001). Altogether, in Model 1, L2 MA and L2 vocabulary knowledge explained 67% of the variance. However, the results of Model 2 indicate that there was no significant effect of L2 MA after accounting for L2 vocabulary (R2 = 0.67, p = .06). When entered first in Model 2, L2 vocabulary alone significantly predicted 64% of the variance of L2 word-meaning inferencing (R2 = 0.64, p < .001). The result that either model analysis found any statistical effect of the interaction between L2 MA and L2 vocabulary knowledge, and that the main effect of L2 MA was significant when it was entered first, yet insignificant after accounting for L2 vocabulary knowledge, seems to suggest that there was a mediation effect of L2 vocabulary knowledge in L2 word-meaning inferencing. Table 6: Hierarchical regression results with L2 word-meaning inferencing as outcome (N = 45) Model Step Variable B R2 ΔR2 ΔF Model 1 1 L2 MA 2.76** 0.21 0.21 11.27** 2 L2 MA −1.51 0.67 0.46 58.66*** L2 vocabulary knowledge 5.26*** 3 L2 MA −1.60* 0.69 0.02 2.12 L2 vocabulary knowledge 5.05*** L2 MA × L2 vocabulary knowledge 0.86 Model 2 1 L2 vocabulary knowledge 4.30*** 0.64 0.64 76.27*** 2 L2 vocabulary knowledge 5.26*** 0.67 0.03 3.80 L2 MA −1.51 3 L2 vocabulary knowledge 5.05*** 0.69 0.02 2.12 L2 MA −1.60* L2 MA × L2 vocabulary knowledge 0.86 Model Step Variable B R2 ΔR2 ΔF Model 1 1 L2 MA 2.76** 0.21 0.21 11.27** 2 L2 MA −1.51 0.67 0.46 58.66*** L2 vocabulary knowledge 5.26*** 3 L2 MA −1.60* 0.69 0.02 2.12 L2 vocabulary knowledge 5.05*** L2 MA × L2 vocabulary knowledge 0.86 Model 2 1 L2 vocabulary knowledge 4.30*** 0.64 0.64 76.27*** 2 L2 vocabulary knowledge 5.26*** 0.67 0.03 3.80 L2 MA −1.51 3 L2 vocabulary knowledge 5.05*** 0.69 0.02 2.12 L2 MA −1.60* L2 MA × L2 vocabulary knowledge 0.86 * p < .05; **p < .01; ***p < .001. Table 6: Hierarchical regression results with L2 word-meaning inferencing as outcome (N = 45) Model Step Variable B R2 ΔR2 ΔF Model 1 1 L2 MA 2.76** 0.21 0.21 11.27** 2 L2 MA −1.51 0.67 0.46 58.66*** L2 vocabulary knowledge 5.26*** 3 L2 MA −1.60* 0.69 0.02 2.12 L2 vocabulary knowledge 5.05*** L2 MA × L2 vocabulary knowledge 0.86 Model 2 1 L2 vocabulary knowledge 4.30*** 0.64 0.64 76.27*** 2 L2 vocabulary knowledge 5.26*** 0.67 0.03 3.80 L2 MA −1.51 3 L2 vocabulary knowledge 5.05*** 0.69 0.02 2.12 L2 MA −1.60* L2 MA × L2 vocabulary knowledge 0.86 Model Step Variable B R2 ΔR2 ΔF Model 1 1 L2 MA 2.76** 0.21 0.21 11.27** 2 L2 MA −1.51 0.67 0.46 58.66*** L2 vocabulary knowledge 5.26*** 3 L2 MA −1.60* 0.69 0.02 2.12 L2 vocabulary knowledge 5.05*** L2 MA × L2 vocabulary knowledge 0.86 Model 2 1 L2 vocabulary knowledge 4.30*** 0.64 0.64 76.27*** 2 L2 vocabulary knowledge 5.26*** 0.67 0.03 3.80 L2 MA −1.51 3 L2 vocabulary knowledge 5.05*** 0.69 0.02 2.12 L2 MA −1.60* L2 MA × L2 vocabulary knowledge 0.86 * p < .05; **p < .01; ***p < .001. In light of the above, a subsequent analysis was carried out to examine the mediation effect of L2 vocabulary knowledge using SPSS 19.0 and an add-on tool Process Version 2.12.1 (Hayes 2013). Following Hayes (2013), four steps were taken to examine the mediator in the regression analysis (as demonstrated in Figure 2). In Step 1 of the mediation model, the regression of L2 MA on L2 word-meaning inferencing, ignoring the mediator L2 vocabulary knowledge, was significant, B = 1.38, p < .01. Step 2 showed that the regression of L2 MA on the mediator L2 vocabulary knowledge was also significant, B = 0.81, p < .001. Step 3 of the mediation process suggested that the mediator (L2 vocabulary knowledge), controlling for L2 MA, was significant, B = 2.63, p < .001. In Step 4, the analyses revealed that, controlling for the mediator (L2 vocabulary knowledge), L2 MA was not a significant predictor of L2 word-meaning inferencing, B = −0.76, p = .058. A Sobel test was conducted and found full mediation in the model (z = 5.06, p < .001, R2 mediation effect size = 0.18). Therefore, the aforementioned analysis indicated that L2 vocabulary knowledge fully mediated the relationship between L2 MA and L2 word-meaning inferencing, the indirect effect of L2 MA accounted for about 18% of the variance of L2 word-meaning inferencing. Figure 2: View largeDownload slide Steps to test mediation effect in regression analysis (adapted from Hayes, 2013) Note: IV, independent variable; M, mediator; DV, dependent variable Figure 2: View largeDownload slide Steps to test mediation effect in regression analysis (adapted from Hayes, 2013) Note: IV, independent variable; M, mediator; DV, dependent variable 5. DISCUSSION The results of this study add to our understanding of the respective roles of L2 vocabulary knowledge and L2 MA, as well as their interrelationship in L2 word-meaning inferencing. To reiterate, previous research has considered L2 vocabulary knowledge as the predominant predictor of successful L2 word-meaning inferencing, and there has been limited attention to how L2 vocabulary knowledge and L2 MA jointly contribute to the inferencing of different types of unknown words. Among those that did examine the relative contributions of vocabulary knowledge and MA in L2 reading, the focus was on reading comprehension as the outcome, and seemed to indicate there are both direct and indirect contributions of L2 MA via L2 vocabulary knowledge (Kieffer and Lesaux 2012a; Kieffer et al. 2013). Aligned with previous research on L2 word-meaning inferencing, L2 vocabulary knowledge is important in this study because it contributed directly and consistently to inferencing the meanings of different types of multi-character words. However, L2 vocabulary knowledge was not the sole significant predictor of L2 word-meaning inferencing; L2 MA made an additional unique contribution. L2 vocabulary knowledge and L2 MA together explained around 67% in the variance of L2 word-meaning inferencing; L2 MA contributed indirectly to L2 word-meaning inferencing through the mediation of L2 vocabulary knowledge, which accounted for about 18% of the variance in L2 word-meaning inferencing. Collectively, to answer the question whether vocabulary knowledge alone is sufficient for L2 word-meaning inferencing, the findings of this study seem to indicate that L2 vocabulary knowledge is necessary, but in itself insufficient, in predicting L2 word-meaning inferencing ability. It is certainly not surprising to find the direct effect of L2 vocabulary knowledge on L2 word-meaning inferencing, since this study measured both semantic word knowledge and morpheme knowledge of adult L2 Chinese readers, and it was more likely for those with larger vocabulary size as well as better insights of words to be successful in identifying unknown word meanings. On the other hand, the indirect effect of L2 MA mediated by L2 vocabulary knowledge merits further discussion. As mentioned earlier, previous studies disagreed about the relative contributions of L2 MA and L2 vocabulary knowledge to L2 reading outcomes (mostly reading comprehension): some found a significant direct effect of MA over and above vocabulary knowledge (Kieffer and Lesaux 2008; Jeon 2011), some found an indirect effect of MA through vocabulary knowledge (Zhang and Koda 2012; Goodwin et al. 2013), and still others observed both direct and indirect effects of L2 MA with full or partial mediation of vocabulary knowledge (Kieffer and Lesaux 2012a; Kieffer et al. 2013). In a study with Chinese university EFL students, Zhang and Koda (2012) observed no significant direct effect of MA when considering vocabulary knowledge, but found a full mediation effect of vocabulary knowledge. They postulated that part of the reason might be that they measured multiple dimensions of vocabulary knowledge (vocabulary size and vocabulary depth), and thus it strengthened the contribution of vocabulary knowledge and limited that of MA, whereas some of previous studies measured vocabulary size only (Kieffer and Lesaux 2008, 2012a; Jeon 2011; Kieffer et al. 2013). Corroborating the results presented by Zhang and Koda (2012), this study measured the different dimensions of vocabulary knowledge (i.e. semantic word knowledge and morpheme knowledge) of adult L2 Chinese readers, which might amplify the significance of vocabulary knowledge in L2 word-meaning inferencing and limit that of MA. Another possibility for not observing any significant direct effect of MA beyond vocabulary knowledge is the way in which MA was measured. Different from this study, Kieffer and colleagues (Kieffer and Lesaux 2012a; Kieffer et al. 2013) observed both direct and indirect effects of MA on reading comprehension in English among adolescent Spanish-speaking ELLs. For one thing, they only measured (reading) vocabulary size; for another, their measure of MA in the target language (i.e. English) was modeled after Carlisle’s (2000) at the sentence level. For example, the participants were provided with a word with a derivational suffix (e.g. complexity) and asked to extract the base word (e.g. complex) to complete a sentence (e.g. The problem is ___ .) (adapted from Kieffer et al. 2013). This kind of measure might indicate the reader’s ability to detect syntactic signals (e.g. syntactic information conveyed by suffixes) (Nagy 2007), and in doing so, MA can make a direct contribution to reading comprehension for its facilitation effect on syntactic processing. However, this syntactic aspect might be language-specific. In the present study, L2 MA in Chinese was measured at the word level. In Chinese, it is unlikely that readers will determine the syntactic status of a stand-alone morphologically complex word as in English, for a lot of the syntactic markers cannot be found (e.g. case and number markers in nouns, as well as tense and aspect markers in verbs, Li and Thompson 1981). Rather, the syntactic category of a word is determined mainly by word order in the sentence. In view of above, the respective measurement of L2 vocabulary knowledge and L2 MA might have contributed to current findings that the contribution of L2 MA to L2 word-meaning inferencing was indirect and fully mediated by L2 vocabulary knowledge. 6. CONCLUSIONS AND LIMITATIONS The present study investigated whether L2 vocabulary knowledge alone is sufficient for L2 word-meaning inferencing. Specifically, it examined the contributions of L2 vocabulary knowledge and L2 MA in inferring the meanings of different types of multi-character unknown words in L2 Chinese. The results suggested that L2 vocabulary knowledge was necessary but insufficient for successful L2 word-meaning inferencing, since (i) the interrelationship between L2 vocabulary knowledge and L2 word-meaning inferencing was consistent across different types of multi-character words, at least not influenced by morphological structure and familiarity with base words, (ii) L2 vocabulary knowledge played an important role due to its direct and mediation effects, and (iii) L2 MA made an additional contribution through the mediation of L2 vocabulary knowledge. The findings of this study add to our understanding of the contribution of L2 MA to L2 word-meaning inferencing after accounting for L2 vocabulary knowledge, as well as the complex intralingual relationship between L2 MA and vocabulary knowledge in L2 word-meaning inferencing. Resonating with previous research on later reading skill development, this study did not observe any direct contribution of L2 MA in adult L2 readers of Chinese, but an indirect contribution of L2 MA through the mediation of L2 vocabulary knowledge. Also, this study was among the first to examine potential differences in inferring the meanings of morphologically complex and simple words in Chinese as a second language. Given that many of the Chinese written words are multi-character and vary in their morphological structure, it is postulated that L2 MA is especially useful for L2 Chinese readers because it helps them to segment words into smaller meaningful units and extract meanings from unknown lexical units in reading Chinese. However, only a very small, but not statistically significant, impact of word effects was observed. It is noted that the present study has some limitations to be addressed in future research. First, it only measured the accuracy of word-meaning inferencing. To examine potential word effects, examining both accuracy and speed of word-meaning inferencing might be a more sensitive and appropriate measure. Second, due to the relatively small sample size, mean- and correlation-based comparisons were made in the analysis. To examine the interaction between L2 MA and vocabulary knowledge in predicting L2 word-meaning inferencing, more sophisticated statistical analyses (e.g. path analysis or structural equation modeling) with a larger participant pool and item pool is needed for future replication studies. Third, other word characteristics might also need to be considered in future research. Following Reichle and Perfetti (2003), this study adopted a relatively straightforward form (orthography)-meaning approach and focused on two word characteristics when investigating L2 Chinese reading: morphological structure and base word familiarity. But it did not control strictly for other word characteristics, for example, word length and word syntactic category (see Dronjic 2011). Finally, future studies might consider including control variables like orthographic knowledge. For instance, Everson (1998) found a very strong convergence between non-heritage Chinese learners’ ability to name written words and their knowledge of the meanings of those words. When investigating how L2 MA contributed to L2 word-meaning inferencing, this study measured MA in the written format, which differed from previous studies with monolingual children that typically measured MA in oral tasks. Given that the target population in the present study (i.e. adult L2 readers of Chinese) started learning to speak and read in the target language concurrently, using written tasks was perceived to be more appropriate and pertaining to their learning experience. In future research, there is a need to examine the potential effect of orthographic knowledge in relation to MA and vocabulary knowledge in L2 word-meaning inferencing in learners with different proficiency levels and literacy experiences in the target language. Sihui Ke received her PhD in Second Language Acquisition at Carnegie Mellon University. Her research has focused on second language reading and biliteracy development, and foreign language assessment. Currently, she is involved in studies of the contributions of morphological awareness to adult and child second language reading. Address for correspondence: Room G419, School of Humanities and Social Science, Harbin Institute of Technology (Shenzhen), Xili University Town, HIT Campus, Nanshan District, Shenzhen 518055, P. R. China. <echoecho.ke@gmail.com> or <kesihui@hit.edu.cn> Keiko Koda is a professor in the Department of Modern Languages at Carnegie Mellon University. Her research interests include second language reading, biliteracy development, and foreign language instruction and assessment. REFERENCES Anderson R. C. , Freebody P. . 1981 . ‘Vocabulary knowledge’ in Guthrie J. T. (ed.): Comprehension and Teaching: Research Reviews . International Reading Association , pp. 77 – 117 . Andrews S. , Bond R. . 2009 . ‘ Lexical expertise and reading skill: Bottom-up and top-down processing of lexical ambiguity ,’ Reading and Writing: An Interdisciplinary Journal 22 / 6 : 687 – 711 . Google Scholar CrossRef Search ADS Anglin J. M. , Miller G. A. , Wakefield P. C. . 1993 . ‘ Vocabulary development: A morphological analysis ,’ Monographs of the Society for Research in Child Development 58 / 10 : 1 – 187 . Google Scholar CrossRef Search ADS Bai X. , Yan G. , Zang C. , Liversedge S. P. , Rayner K. . 2008 . ‘ Reading spaced and unspaced Chinese text: Evidence from eye movements ,’ Journal of Experimental Psychology: Human Perception and Performance 34 / 5 : 1277 – 87 . Google Scholar CrossRef Search ADS PubMed Bialystok E. 2001 . Bilingualism in Development . Cambridge University Press . Google Scholar CrossRef Search ADS Carlisle J. F. 2000 . ‘ Awareness of the structure and meaning of morphologically complex words: Impact on reading ,’ Reading and Writing: An Interdisciplinary Journal 12 : 169 – 90 . Google Scholar CrossRef Search ADS Chen J. 1999 . ‘Word recognition during the reading of Chinese sentences: Evidence from studying the word superiority effect’ in Wang J. , Inhoff A. W. , Chen H.-C. (eds): Reading Chinese Script: A Cognitive Analysis . Erlbaum , pp. 239 – 56 . Chinese Proficiency Test Center . 2001 . Grading Syllabus for Chinese Vocabulary and Chinese Characters [in Chinese]. Economy and Science Press . Christianson K. , Luke S. G., . 2011 . ‘ Context strengthens initial misinterpretations of text ,’ Scientific Studies of Reading 15 / 2 : 136 – 66 . Google Scholar CrossRef Search ADS Chung F. H -K. , Leung M. -T., . 2008 . ‘ Data analysis of Chinese characters in primary school corpora of Hong Kong and mainland China: Preliminary theoretical interpretations ,’ Clinical Linguistics and Phonetics 22 / 4–5 : 379 – 89 . Google Scholar CrossRef Search ADS PubMed Cohen J. 1988 . Statistical Power Analysis for the Behavioral Sciences , 2nd edn Erlbaum . de Bot K. , Paribakht T. S. , Wesche M. B., . 1997 . ‘ Toward a lexical processing model for the study of second language vocabulary acquisition ,’ Studies in Second Language Acquisition 19 / 3 : 309 – 29 . Google Scholar CrossRef Search ADS Deacon S. H. , Kieffer M. J. , Laroche A., . 2014 . ‘ The relation between morphological awareness and reading comprehension: Evidence from mediation and longitudinal models ,’ Scientific Studies of Reading 18 : 432 – 51 . Google Scholar CrossRef Search ADS Dronjic V. 2011 . ‘ Mandarin Chinese compounds, their representation, and processing in the visual modality ,’ Writing Systems Research 3 / 1 : 5 – 21 . Google Scholar CrossRef Search ADS Everson M. E. 1998 . ‘ Word recognition among learners of Chinese as a foreign language: Investigating the relationship between naming and knowing ,’ Modern Language Journal 82 / 2 : 194 – 204 . Google Scholar CrossRef Search ADS Field A. 2009 . Discovering Statistics Using SPSS , 3rd edn Sage . Gilbert J. K. , Goodwin A. P. , Compton D. L. , Kearns D. M. . 2013 . ‘ Multisyllabic word reading as a moderator of morphological awareness and reading comprehension ,’ Journal of Learning Disabilities 47 / 1 : 34 – 43 . Google Scholar CrossRef Search ADS PubMed Goodwin A. P. , Huggins A. C. , Carlo M. S. , August D. , Calderon M., . ( 2013 ). ‘ Minding morphology: How morphological awareness relates to reading for English language learners ,’ Reading and Writing: An Interdisciplinary Journal 26 / 9 : 1387 – 141 . Google Scholar CrossRef Search ADS Haastrup K. 1991 . Lexical Inferencing Procedures or Talking about Words . Gunter Narr Verlag . Hamada M. 2014 . ‘ The role of morphological and contextual information in L2 lexical inference ,’ The Modern Language Journal 98 / 4 : 992 – 1005 . Google Scholar CrossRef Search ADS Hatami S. , Tavakoli M., . 2012 . ‘ The role of depth versus breadth of vocabulary knowledge in success and ease in L2 lexical inferencing ,’ TESL Canada Journal 31 / 1 : 1 – 21 . Hayes A. F. 2013 . Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-based Approach . Guilford Press . Hu M. , Nation I. S. P., . 2000 . ‘ Unknown vocabulary density and reading comprehension ,’ Reading in a Foreign Language 13 : 403 – 30 . Jeon E. H. 2011 . ‘ Contribution of Morphological awareness to second-language reading comprehension ,’ The Modern Language Journal 95 / 2 : 217 – 35 . Google Scholar CrossRef Search ADS Kieffer M. J. , Biancarosa G. , Mancilla-Martinez J., . 2013 . ‘ Roles of morphological awareness in the reading comprehension of Spanish-speaking language minority learners: Exploring partial mediation by vocabulary and reading fluency ,’ Applied Psycholinguistics 34 / 04 : 697 – 725 . Google Scholar CrossRef Search ADS Kieffer M. J. , Lesaux N. K. , 2008 . ‘ The role of derivational morphology in the reading comprehension of Spanish-speaking English language learners ,’ Reading and Writing: An Interdisciplinary Journal 21 / 8 : 783 – 804 . Google Scholar CrossRef Search ADS Kieffer M. J. , Lesaux N. K., . 2012a . ‘ Direct and indirect roles of morphological awareness in the English reading comprehension of native English, Spanish, Filipino, and Vietnamese speakers ,’ Language Learning 62 / 4 : 1170 – 204 . Google Scholar CrossRef Search ADS Kieffer M. J. , Lesaux N. K., . 2012b . ‘ Development of morphological awareness and vocabulary knowledge in Spanish-speaking language minority learners: A parallel process latent growth curve model ,’ Applied Psycholinguistics 33 : 23 – 54 . Google Scholar CrossRef Search ADS Koda, K. 2000 . ‘ Cross-linguistic variations in L2 morphological awareness ,’ Applied Psycholinguistics 21 : 297 – 320 . Koda, K. 2007 . ‘ Reading and language learning: cross-linguistic constraints on second language reading development ,’ Language Learning 57 / S1 : 1 – 44 . Koda K. , Lű C. , Zhang D., . 2014 . ‘L1-induced facilitation in biliteracy development in Chinese and English’ in Chen X. , Wang Q. , Luo Y. C. (eds): Reading Development and Difficulties in Monolingual and Bilingual Chinese Children . Springer . Ku Y.-M. 2001 . ‘Development of morphological awareness in Chinese and English,’ Unpublished doctoral dissertation, University of Illinois at Urbana-Champaign. Ku Y. -M. , Anderson R. C., . 2003 . ‘ Development of morphological awareness in Chinese and English ,’ Reading and Writing: An Interdisciplinary Journal 16 : 399 – 422 . Google Scholar CrossRef Search ADS Kuo L. -J. , Anderson R. C., . 2008 . ‘Conceptual and methodological issues in comparing metalinguistic awareness across languages’ in Koda K. , Zehler A. (eds): Learning to Read Across Languages . Routledge , pp. 39 – 67 . Larson-Hall J. 2010 . A Guide to Doing Statistic in Second Language Research Using SPSS . Routledge . Li C. N. , Thompson S. A. . 1981 . Mandarin Chinese: A Functional Reference Grammar . University of California Press . Li X. S. , Bicknell K. , Liu P. P. , Wei W. , Rayner K., . 2014 . ‘ Reading is fundamentally similar across disparate writing systems: A systematic characterization of how words and characters influence eye movements in Chinese reading ,’ Journal of Experimental Psychology General 143 / 2 : 895 – 913 . Google Scholar CrossRef Search ADS PubMed Lin T. -J. , Anderson R. J. , Ku Y. -M. , Christianson K. , Packard J. L., . 2011 . ‘ Chinese children’s concept of word ,’ Writing Systems Research 3 / 1 : 41 – 57 . Google Scholar CrossRef Search ADS Liu N. , Nation P., . 1985 . ‘ Factors affecting guessing vocabulary in context ,’ RECL Journal 16 : 33 – 42 . Liu P. -P. , Li W. -J. , Lin N. , Li X. -S. . 2013 . ‘ Do Chinese readers follow the national standard rules for word segmentation during reading? ,’ PLoS One 8 / 2 : e55440 . Google Scholar CrossRef Search ADS PubMed Liu Y. 2013 . ‘L1 reading ability, L2 linguistic knowledge, and L2 semantic gap filling ability,’ Unpublished doctoral dissertation, Carnegie Mellon University. Mcbride-Chang C. , Tardif T. , Cho J. -R. , Shu H. , Fletcher P. , Stokes S. F. , Wong A. , Leung K. . 2008 . ‘ What’s in a word? Morphological awareness and vocabulary knowledge in three languages ,’ Applied Psycholinguistics , 0329 : 437 – 62 . Meara P. 1996 . ‘The dimensions of lexical competence’ in Brown G , Malmkjaer K. , Williams J. (eds): Competence and Performance in Language Learning . Cambridge University Press Mori Y. , Nagy W. . 1999 . ‘ Integration of information from context and word elements in interpreting novel kanji compounds ,’ Reading Research Quarterly 34 / 1 : 80 – 101 . Google Scholar CrossRef Search ADS Myers J. 2006 . ‘Processing Chinese compounds: A survey of the literature’ in Libben G. , Jarema G. (eds): The Representation and Processing of Compound Words . Oxford University Press , pp. 169 – 96 . Nagy W. 2007 . ‘Metalinguistic awareness and the vocabulary-comprehension connection’ in Wagner R. , Muse A. , Tannenbaum K. (eds): Vocabulary Acquisition: Implications for Reading Comprehension . Guilford Press , pp. 53 – 77 . Nagy W. , Berninger V. W. , Abbott R. D., . 2006 . ‘ Contributions of morphology beyond phonology to literacy outcomes of upper elementary and middle-school students ,’ Journal of Educational Psychology 98 / 1 : 134 – 47 . Google Scholar CrossRef Search ADS Nassaji H. 2003 . ‘ L2 vocabulary learning from context: Strategies, knowledge sources, and their relationship with success in L2 lexical inferencing ,’ TESOL Quarterly 37 / 4 : 645 – 70 . Google Scholar CrossRef Search ADS Nassaji H. 2004 . ‘ The relationship between depth of vocabulary knowledge and L2 learners' lexical inferencing strategy use and success ,’ The Canadian Modern Language Review 61 / 1 : 107 – 34 . Google Scholar CrossRef Search ADS Nation I. S. P. 1990 . Teaching and Learning Vocabulary . Newbury House . Nation I. S. P. 2001 . Learning Vocabulary in Another Language . Cambridge University Press . Google Scholar CrossRef Search ADS Nation I. S. P. 2006 . ‘ How large a vocabulary is needed for reading and listening? ,’ The Canadian Modern Language Review 63 / 1 : 59 – 81. . Google Scholar CrossRef Search ADS Packard J. L. 2000 . The Morphology of Chinese . Cambridge University Press . Google Scholar CrossRef Search ADS Parel R. 2004 . ‘ The impact of lexical inferencing strategies on second language reading proficiency ,’ Reading and Writing: An Interdisciplinary Journal 17 / 6 : 847 – 73 . Google Scholar CrossRef Search ADS Paribakht T. S. , Wesche M., . 1999 . ‘' Incidental' and instructed L2 vocabulary acquisition through reading: An introspective study ,’ Studies in Second Language Acquisition 21 / 2 : 195 – 224 . Google Scholar CrossRef Search ADS Park E. C. 2004 . ‘The Relationship between Morphological Awareness and Lexical Inference Ability for English Language Learning Children with Korean First-language Background,’ Unpublished doctoral dissertation, Carnegie Mellon University. Peng R. -Y. 2004 . ‘ Even words are right, odd ones are odd: Explaining word segmentation inconsistency among Chinese readers ,’ Chinese Journal of Psychology 46 / 1 : 49 – 55 . Qian D. D. 2005 . ‘ Demystifying lexical inferencing: The role of aspects of vocabulary knowledge ,’ TESL Canada Journal 22 / 2 : 34 – 54 . Google Scholar CrossRef Search ADS Read J. 1993 . ‘ The development of a new measure of L2 vocabulary knowledge ,’ Language Testing 10 / 3 : 355 – 71 . Google Scholar CrossRef Search ADS Read J. 2000 . Assessing Vocabulary . Cambridge University Press . Google Scholar CrossRef Search ADS Reichle E. D. , Perfetti C., . 2003 . ‘ Morphology in word identification: A word-experience model that account for morpheme frequency effect ,’ Scientific Studies of Reading 7 / 3 : 219 – 37 . Google Scholar CrossRef Search ADS Schmitt N. , Jiang X. , Grabe W. . 2012 . ‘The percentage of words known in a text and reading comprehension ,’ Modern Language Journal 95 / 1 : 26 – 43 . Google Scholar CrossRef Search ADS Shu H. , Chen X. , Anderson R. C. , Wu N. , Xuan Y., . 2003 . ‘ Properties of school Chinese: Implications for learning to read ,’ Child Development 74 / 1 : 27 – 47 . http://dx.doi.org/10.1023/A:1025334204364. Google Scholar CrossRef Search ADS PubMed Verhoeven L. , Perfetti C. A., . 2011 . ‘ Morphological processing in reading acquisition: A cross-linguistic perspective ,’ Applied Psycholinguistics 32 / 3 : 457 – 66 . Google Scholar CrossRef Search ADS Xu X. 2004 . ‘The morphological awareness of overseas students in the acquisition of Chinese compound words,’ Unpublished master thesis, Beijing Language and Culture University. Zeng L. 2008 . ‘ Quantitative and qualitative study of Chinese semi-affixes ,’ World Chinese Language Teaching 4 : 75 – 87 . Zhang, D. and Koda, K . 2012 . ‘ Contribution of morphological awareness and lexical inferencing ability to L2 vocabulary knowledge and reading comprehension among advanced EFL learners: Testing direct and indirect effects ,’ Reading and Writing 25 / 5 : 1195 – 216 . © Oxford University Press 2017 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Applied Linguistics Oxford University Press

Is Vocabulary Knowledge Sufficient for Word-Meaning Inference? An Investigation of the Role of Morphological Awareness in Adult L2 Learners of Chinese

Applied Linguistics , Volume Advance Article – Nov 16, 2017

Loading next page...
 
/lp/ou_press/is-vocabulary-knowledge-sufficient-for-word-meaning-inference-an-PKyAszbXtw
Publisher
Oxford University Press
Copyright
© Oxford University Press 2017
ISSN
0142-6001
eISSN
1477-450X
D.O.I.
10.1093/applin/amx040
Publisher site
See Article on Publisher Site

Abstract

Abstract This study examined whether vocabulary knowledge is sufficient for second language (L2) word-meaning inferencing. Specifically, it investigated the role of the learner’s sensitivity to a word’s morphological structure (referred to ‘morphological awareness’ or ‘MA’ hereafter) as additional support that enhances multi-character word-meaning inferencing in adult learners of Chinese as a second language. Two hypotheses were tested: (i) L2 vocabulary knowledge is the sole predictor of word-meaning inferencing involving morphologically simple multi-character words. (ii) L2 vocabulary knowledge and MA jointly contribute to word-meaning inferencing involving morphologically complex multi-character words. A set of paper-and-pencil tests was administered to 56 English-speaking learners of L2 Chinese, measuring their L2 vocabulary knowledge, MA, and word-meaning inferencing ability. The results are as follows: L2 vocabulary knowledge contributed directly and consistently to inferring the meanings of multi-character words. Yet, L2 vocabulary knowledge was not the sole predictor of L2 word-meaning inferencing; L2 MA made an additional unique contribution. Last, L2 MA contributed indirectly to L2 word-meaning inferencing through the mediation of L2 vocabulary knowledge. 1. INTRODUCTION One of the main purposes of foreign language instruction and adult second language (L2) learning is to equip learners with the skills to learn new words incidentally through reading, listening, and viewing. Skilled L2 readers are presumed to be adept at making inferences when encountering an unfamiliar word based on the information available in the word and the information afforded by the surrounding context in which the word appears. To do so, they rely heavily on L2 linguistic knowledge—vocabulary knowledge, in particular. Vocabulary knowledge is broadly defined as a learner’s knowledge of words, including knowing its form, meaning, as well as use (Nation 1990, 2001). Research has consistently shown that vocabulary knowledge is closely related to word-meaning inference among L2 readers (Nassaji 2004; Qian 2005; Hatami and Tavakoli 2012). Given that morphologically complex words are prevalent in reading materials for adult learners, a critical question is whether vocabulary knowledge alone is sufficient for supporting successful L2 word-meaning inferencing. Previous studies have indicated that morphological awareness (MA), learners’ sensitivity to the morphological structure of words (Koda 2000; Carlisle 2000; Verhoeven and Perfetti 2011), may compensate for low vocabulary knowledge in L2 reading development (Parel 2004; Gilbert et al. 2013). But findings varied among L1 and L2 reading research that considered the relative contributions of MA and vocabulary knowledge to reading subskills development. Some researchers held that MA make a significant contribution over and above vocabulary knowledge (e.g. L1: Carlisle 2000; Ku and Anderson 2003; Nagy et al. 2006; L2: Kieffer and Lesaux 2008; Jeon 2011), whereas others found that the contribution of MA is mediated fully or partially mediated by vocabulary knowledge (Zhang and Koda 2012; Kieffer and Lesaux 2012a; Goodwin et al. 2013; Kieffer et al. 2013; Deacon et al. 2014). Notably, the aforementioned insights have been derived mainly from studies focusing on reading comprehension as the outcome and English as the target language. It remains unclear as to how L2 MA and L2 vocabulary knowledge could jointly contribute to L2 word-meaning inferencing across typologically different languages. This study investigated the role of the learner’s sensitivity to a word’s morphological structure (referred to ‘morphological awareness/MA’ hereafter) as additional support that enhances word-meaning inferencing in adult learners of Chinese as a second language. Previous studies on L2 lexical inferencing typically followed a design that aimed to identify the proportional distributions of MA and vocabulary knowledge, as two different resources, to successful L2 word-meaning inferencing (Nassaji 2003), and little research has taken a component approach to modeling how L2 MA and L2 vocabulary knowledge may interact to predict L2 word-meaning inferencing ability. In this research, we adopted a component approach toward L2 word-meaning inferencing during reading in this research. Following Bialystok (2001) and Koda (2007), we distinguish MA (a facet of metalinguistic awareness, that is, sensitivity to the abstract structure of language) from vocabulary knowledge (knowledge of specific words), and word-meaning inferencing as a reading subskill. The objectives of the study were twofold: (i) determining whether L2 vocabulary knowledge contributes directly and consistently to inferring the meanings of different types of multi-character words, and (ii) exploring the intralingual relationship between L2 vocabulary knowledge and L2 MA in their contributions L2 word-meaning inferencing. The writing system and morphological properties of Chinese provide an ideal opportunity for this investigation: (i) Chinese employs a morphosyllabic writing system, in which each graphic symbol encodes a morpheme (its syllable) rather than a phoneme. (ii) The majority of Chinese words are written with multiple characters. (iii) The number of characters (morphemes) varies from word to word (most words consist of two to four characters), and (iv) the word boundaries and the morphological boundaries within a word are not visually marked because a space is not used in Chinese writing. What follows first includes a review of the respective roles of vocabulary knowledge and MA in L2 word-meaning inferencing, the intralingual relationship between the two, and the unique morphological properties of printed Chinese word. 2. LITERATURE REVIEW 2.1 L2 vocabulary knowledge and L2 word-meaning inferencing Word-meaning inferencing or lexical inferencing is often referred to as ‘making informed guesses to the meaning of a word in the light of all available linguistic cues in combination with the learners’ general knowledge of the world, [his/her] awareness of the co-text and [his/her] relevant linguistic knowledge’ (Haastrup 1991: 11). Previous studies of L2 word-meaning inferencing have identified L2 vocabulary knowledge as the critical predictor of inferring unknown word meanings successfully in L2 readers (Nassaji 2004; Qian 2005; Hatami and Tavakoli 2012). The relationship between vocabulary knowledge and word-meaning inferencing among L2 learners was examined with regard to the relative contributions of two subcomponents of vocabulary knowledge, namely, vocabulary size and vocabulary depth. Vocabulary size is often referred to as the quantity or the number of words one knows (Nation 2001), while vocabulary depth relates to how well one knows a word (Read 1993, 2000; Meara 1996). Although vocabulary size is often considered to play a critical role in unsupported reading, which indicates the need for a very high vocabulary coverage rate for successful reading comprehension, ranging between 95% (Liu and Nation 1985) and 98% (Hu and Nation 2000; Nation 2006; Schmitt et al. 2012), vocabulary depth was found to be more important in L2 word-meaning inferencing. For instance, Hatami and Tavakoli (2012) measured both vocabulary size and depth in Persian-speaking university learners of English as a foreign language (EFL), and compared their relative contributions to word-meaning inferencing. Their results suggested that vocabulary depth contributed more to L2 word-meaning inferencing than did vocabulary size. In other studies that focused on the impact of vocabulary depth on L2 word-meaning inferencing, a positive correlation was observed as well (Nassaji 2004; Qian 2005). Qian (2005) postulated that there might be a Matthew effect between the two, as profound vocabulary knowledge increases the chance of word-meaning inferencing success, and successful word-meaning inferencing in turn deepens L2 learners’ knowledge of words. In spite of the significance of L2 vocabulary knowledge, a question that remains is whether vocabulary knowledge alone will be sufficient for inferring unknown word meanings successfully in L2 reading. If the answer is yes, this seems to imply an all-or-nothing manner in L2 reading and learning; that is, if L2 readers cannot identify an unknown word based on their prior vocabulary knowledge, they have no other resources to infer the unknown word meaning and learn new words. Another issue is whether L2 readers’ vocabulary knowledge is sufficient for them to retrieve contextual information to infer unknown word meanings. As indicated in previous research, for L2 readers, contextual information might not be as useful as it is assumed to be because it could strengthen initial misinterpretation (Christianson and Luke 2011). On the other hand, high-quality word-internal lexical representation (a nexus of orthographical, phonological, and morphological/semantic information) will reduce reliance on context (Andrews and Bond 2009), as it has been found that word-internal morphological information is one of the resources used frequently by L2 readers in word-meaning inferencing (de Bot et al. 1997; Paribakht and Wesche 1999). Both de Bot et al. (1997) and Paribakht and Wesche (1999) examined different resources used by university intermediate-level learners’ inferencing processes in L2 English and observed that L2 learners used their knowledge of word derivations and inflections frequently to infer the meanings of unknown words. This ability to parse words and utilize word-internal morphological information should pertain to MA. 2.2 L2 MA and L2 word-meaning inferencing MA is often referred to as the ability to reflect upon and manipulate constituent morphemes in visual word processing for the purpose of meaning construction (Carlisle 2000; Kuo and Anderson 2008). Considering the amount of studies of the effects of L2 vocabulary knowledge on L2 word-meaning inferencing, the role of L2 MA is underexamined in extant literature. Yet, there has been evidence suggesting that L2 MA makes unique and independent contributions to guessing unknown word meanings during reading. For instance, Park (2004) measured first language (L1) and L2 MA of Korean-speaking English language learners at third to fifth grades and evaluated their relationships with L2 word-meaning inferencing. It was found that L2 MA not only mediated the contribution of L1 MA to L2 word-meaning inferencing, but also had a direct impact on L2 word-meaning inferencing. Park held that the utility of MA lies in that it helps with the extraction of partial information from unfamiliar words, which promotes analytical approaches to lexical processing in reading, even among L2 readers with low L2 linguistic knowledge. In other studies that did not include any MA measure, the utility of MA was inferred from the effects of unknown word morphological properties on successful L2 word-meaning inferencing in learners of higher/lower L2 linguistic knowledge (Mori and Nagy 1999; Hamada 2014). In Mori and Nagy’s (1999) study with English-speaking university learners of L2 Japanese at intermediate and preadvanced levels, they asked learners to infer the meanings of novel semantically semi-transparent compound words in a sentence. Compound words comprised familiar kanjis in three conditions (i.e. kanji/word-internal information only, contextual information only, and the combination of kanji and contextual information). It was found that: (i) more than half of L2 Japanese learners tended to integrate both word-internal and contextual information, and integrators were more successful in word-meaning inferencing than were non-integrators; and (ii) L2 linguistic knowledge correlated with the use of contextual clues but not with the use of word-internal information. Their findings with intermediate- and preadvanced-level Japanese learners indicate that MA seems to be an independent resource from contextual information extraction and less constrained by learners’ L2 linguistic knowledge. More recently, Hamada (2014) investigated whether the use of morphological/contextual information in word-meaning inferencing depends upon the learners’ L2 proficiency and the reliability of the morphological information, with university learners of L2 English at four proficiency levels (beginning, intermediate, high-intermediate, and advanced). Two types of unknown word were included in Hamada’s (2014) study: morphology reliable pseudo compounds (the known word part provided semantic information that matched sentential context) and morphology unreliable compound (the known word part provided semantic information that did not match sentential context). The findings suggested that the choice of information was influenced by the morphological reliability condition because participants performed similarly across proficiency levels in the morphology reliable condition but differently in the morphology unreliable condition (the beginning group used morphological information more often than other proficiency groups). The observations of Mori and Nagy’s (1999) and Hamada’s (2014) indicate that the utility of MA in L2 word-meaning inferencing is independent from L2 linguistic knowledge, but dependent on the morphological structure of a word. Notably, both studies were based on learners’ overall L2 proficiency and did not include any measure of L2 vocabulary knowledge. 2.3 L2 MA, L2 vocabulary knowledge, and L2 word-meaning inferencing When the intralingual relationship between MA, vocabulary knowledge, and word-meaning inferencing is considered, other than the close relationship between vocabulary knowledge and word-meaning inferencing, and that between MA and word-meaning inferencing, the relationship between MA and vocabulary knowledge should be noted as well. Previous L1 reading studies have shown that MA contributes significantly to vocabulary acquisition in different languages (including English, Chinese, and Korean) (Anglin et al. 1993; McBride-Chang et al. 2008; Kieffer and Lesaux 2012b). Anglin et al. (1993) observed a significant increase of derived word knowledge between first- and fifth-grade native English-speaking children, and found that it was through ‘morphological problem solving’ that children improved in their word learning. In a longitudinal study, McBride-Chang et al. (2008) tracked the performance of Chinese-speaking and Korean-speaking monolingual preschoolers over spans of nine months to one year, and found that early MA significantly predicted vocabulary development. Given the close relationship between MA and vocabulary knowledge, and that between vocabulary knowledge and word-meaning inferencing reviewed previously, there is a possibility that L2 MA contributes indirectly to L2 word-meaning inferencing through L2 vocabulary knowledge. However, to our knowledge, to date, there is a lack of research that directly investigates the intralingual relationship between MA and vocabulary knowledge in predicting L2 word-meaning inferencing. Notably, when examining the role of L2 MA in L2 reading, additional attention should be paid to the potential impact of word properties in the target language. In this regard, what follows is a discussion of the case of (Mandarin) Chinese, a language that has attracted relatively less attention in L2 reading acquisition. 2.4 Words, morphemes, and MA in Chinese Following Packard (2000: 12), a word is defined as ‘an independent occupant of a syntactic form class slot’. Written words in English are salient because they are separated by spaces in written texts. In Chinese, however, written texts are not word-based but character-based, with no spaces inserted between words. Emerging evidence suggests that words, as defined by Packard, do have psychological reality for both L1 and L2 reading in Chinese for a number of reasons: (i) identifying characters within a word increases accuracy as opposed to identifying characters in a string of characters that do not constitute a word (Chen 1999); (ii) inserting spaces between characters within words hinders the reading of Chinese texts, whereas inserting spaces between words does not (Bai et al. 2008); and (3) word properties have effects on Chinese reading that go beyond the characters’ properties (Li et al. 2014) (for a review, see Liu et al. 2013). Notably, words in Chinese vary in the numbers of characters that comprise them: 6% are single-character words, 72% are two-character words, 12% are three-character words, and 10% are four-character words (as cited in Li et al. 2014). Because the numbers of two-character and three-character words significantly exceed the number of single-character words, the examination of Chinese words focuses on multi-character words henceforth. Since multiple terms have been used to refer to concepts that involve meaning in Chinese, including character, morpheme, and word, it is important to note the similarities and differences among these concepts. Morphemes in Chinese fall into two categories—grammatical morphemes (e.g. 了, le, as an aspect marker) and word formation morphemes (e.g. 化, huà, resembles the affix -ize in English) (Packard 2000). As mentioned before, each character can usually represent one morpheme, but the two are not identical, since some characters can shift in pronunciation, meaning, or both from context to context (see Myers 2006). An example cited by Myers (2006) is the first morpheme of 行人 (xíngrén, ‘pedestrian’ or ‘walk-person’ literally) and the second morpheme of 银行 (yínháng, ‘bank’ or ‘silver-store’ literally), written with the same character. As well, there has been some misconceptualization, assuming that all multi-character words are compound or multi-morphemic words. For instance, each character in the two-character word 花生 has its own meanings (‘flower/spend’ for 花 and ‘born/student’ for 生, respectively) and can be used independently and functionally; yet the two-character word 花生is a lexicalized expression mapping onto one morpheme, meaning ‘peanut’, the meaning of which cannot be directly inferred from the two-component characters. Another oft-cited notion related to morpheme in written Chinese is semantic radical, which is a subcomponent within a Chinese character that denotes meaning. The present study did not consider the role of semantic radical because the focus was on multi-character rather than single-character words, and the semantic cues provided by semantic radicals are not always reliable (Shu et al. 2003; Chung and Leung 2008). Based on the review above, it is important for researchers to differentiate morphologically complex multi-character words from morphologically simple multi-character words when examining reading in Chinese. Previous studies with both L1 and L2 adult Chinese readers indicated that they are sensitive to the structural properties of multi-character words (Xu 2004; Lin et al. 2011). In a study comparing native Chinese adults’ and children’s performances in a word parsing task, Lin et al. (2011) found that native Chinese adult readers were more aware of word features and more capable of using relatively productive characters as word boundary cues. As for adult readers of L2 Chinese, Xu (2004) examined L2 Chinese MA among learners with no prior exposure to character learning in their L1 experience, and asserted that adult L2 Chinese readers had already developed sensitivity to productive morphemes in their first year of study in China. It is noteworthy that most of the previous studies were based on reading or recognizing known words. Taken together, although L2 vocabulary knowledge has been considered to be an important predictor for L2 word-meaning inferencing, whether L2 vocabulary knowledge alone will contribute consistently to inferring the meanings of unknown words is still an open question. As suggested by previous findings, L2 MA can compensate for insufficient L2 vocabulary knowledge. What merits further examination is how vocabulary knowledge and MA jointly contribute to L2 word-meaning inferencing. Moreover, there is a need for investigation in target languages that are typologically distinct from alphabetic languages. 3. THE PRESENT STUDY This study examined whether L2 vocabulary knowledge alone is sufficient for L2 word-meaning inferencing in Chinese as a second language. Of special interest is whether L2 vocabulary knowledge and L2 MA contribute jointly or differentially to inferring the meanings of different types of multi-character words in L2 Chinese. It aims to expand our understanding of the intricate relationships among L2 vocabulary knowledge, L2 MA, and L2 word-meaning inferencing, and the way in which L2 readers utilize word-internal properties and word-external/contextual information when guessing unknown word meanings during reading. 3.1 Hypotheses In the current study, two hypotheses were formulated regarding how L2 vocabulary knowledge could contribute to L2 word-meaning inferencing when accounting for the potential influence of L2 MA and unknown word characteristics. The first hypothesis was that L2 vocabulary knowledge would be the sole predictor of word-meaning inferencing involving morphologically simple multi-character words, and that there would be no additional contribution of L2 MA. The second was that L2 vocabulary knowledge and MA would jointly contribute to lexical inferencing involving morphologically complex multi-character words. 3.2 Methods 3.2.1 Participants In total, 56 English-speaking learners of L2 Chinese participated in the present study. They were all Chinese learners at an advanced level, which was determined by three criteria: (i) the instructional levels in the participants’ affiliated universities, (ii) a minimum of two years of Chinese learning experience, and (iii) reading as one of the major learning and instructional components. There were 27 female students and 29 male students, with an average age of 21.98 (SD = 4.33). They were recruited from six universities in Shanghai and Beijing, China. By the time of data collection, the participants had received an average of 3.2 years of formal Chinese education. They studied Chinese for 2–4 class hours per week, and were expected to have mastered 2,000 commonly used words and related grammar patterns. 3.2.2 Instruments, data collection, and analytical procedures A paper-and-pencil test package was distributed to the participants. They first completed the L2 word-meaning inferencing task, followed by the L2 MA tasks, L2 vocabulary knowledge tasks, and a word checklist to ensure the lexicality status of the target words in the inferencing task. The estimated time to complete the whole test was 1 hour. L2 word-meaning inferencing. In accordance with Haastrup’s (1991) definition, word-meaning inferencing refers to the ability to make appropriate predictions of unknown word meanings based on word-internal and contextual information. The word-meaning inferencing task was adapted from Mori and Nagy (1999). The participants were asked to infer the meaning of an unknown word underlined within a phrase or a short sentence with a relatively constrained context (which is consisted of less than 10 characters and only provides part-of-speech clue for the target unknown word), and to select the correct answer from four choices written in English based on four conditions: (i) an integrated answer combining both context and the meaning of word subcomponents; (ii) the meaning of the word subcomponent only; (iii) context only; and (iv) a distractor/anomalous answer. For instance, the phrase 全国覆盖面最广(quánguó fùgàimiàn zuìguǎng, ‘the nation’s widest coverage’) had the target unknown word 覆盖面 (fùgàimiàn, ‘coverage’) underlined, followed by four choices: (i) coverage, (ii) surface, (iii) net, and (iv) voice. If L2 Chinese readers only resort to their prior print vocabulary knowledge, they might select (ii) because 面 (miàn) means ‘side/facet’ and the meaning of 覆盖 (fùgài, ‘to cover’) is usually beyond advanced-level learners’ vocabulary. If they use contextual information only, they may select (iii) because the Chinese word 网 (wǎng) corresponding to ‘net’ in English is not relevant to any word-internal component of the target word 覆盖面, and option (iv) was a distractor. Only when the participants utilize both the word-internal morphological information and the word-external contextual meaning can they successfully infer the meaning of the underlined word and select (i). Ease with which the meaning of a target word can be inferred was manipulated in two ways: first by varying the morphological structure of the target word (i.e. morphologically complex/bimorphemic versus morphologically simple/monomorphemic) and second by varying base word familiarity (i.e. whether the participants knew the meaning of the base word). Other than the underlined unknown words, other lexical items in the phrase/sentence were all familiar to the participants (from Bands One and Two/lowest levels in the Grading Syllabus for Chinese Vocabulary and Chinese Characters, Chinese Proficiency Test Center 2001). In total, there were 32 items in the word-meaning inferencing task, with eight items for each word type (as shown in Table 1). One point was assigned for each correct answer. The Cronbach’s α was 0.75. Table 1: Sample items in L2 word-meaning inferencing task Word type n Example English meaning Monomorphemic multi-character words with familiar word part 8 分外 Very Monomorphemic multi-character words with unfamiliar word part 8 螺丝 Screw Bimorphemic multi-character words with familiar base 8 副业 Sideline Bimorphemic multi-character words with unfamiliar base 8 小范围 Small-scale Word type n Example English meaning Monomorphemic multi-character words with familiar word part 8 分外 Very Monomorphemic multi-character words with unfamiliar word part 8 螺丝 Screw Bimorphemic multi-character words with familiar base 8 副业 Sideline Bimorphemic multi-character words with unfamiliar base 8 小范围 Small-scale Note: Unfamiliar base word/word part is underlined. Table 1: Sample items in L2 word-meaning inferencing task Word type n Example English meaning Monomorphemic multi-character words with familiar word part 8 分外 Very Monomorphemic multi-character words with unfamiliar word part 8 螺丝 Screw Bimorphemic multi-character words with familiar base 8 副业 Sideline Bimorphemic multi-character words with unfamiliar base 8 小范围 Small-scale Word type n Example English meaning Monomorphemic multi-character words with familiar word part 8 分外 Very Monomorphemic multi-character words with unfamiliar word part 8 螺丝 Screw Bimorphemic multi-character words with familiar base 8 副业 Sideline Bimorphemic multi-character words with unfamiliar base 8 小范围 Small-scale Note: Unfamiliar base word/word part is underlined. As mentioned earlier, a post-test word checklist was distributed to participants to ensure that selected words’ properties conform to the aforementioned four conditions. The checklist required the participants to self-report their knowledge of the target words and provide the English meanings. If participants successfully identified the word meaning, it was considered as familiar; otherwise, the word was counted as unfamiliar or unknown. L2 MA. MA is a multi-faceted construct, and current measures of MA range from more language-specific to less language-specific (Koda Lűand Zhang 2014). L2 MA was measured by two tasks, segmentation and intraword structure analysis. Segmentation refers to the ability to decompose a word into smaller meaningful units; intraword structure analysis refers to readers’ sensitivity to intraword morphological structure (Koda 2000). Following Peng (2004), the segmentation task required the participants to split multi-character words into smaller words that still have meanings. The items included two-character and three-character words that fell into four categories (as illustrated in Table 2). The segmentation task included 32 items. A participant was credited with one point for each correct answer. The Cronbach’s α was 0.80. For the intraword structure analysis, the task was constructed after Ku (2001). The participants were asked to read three words that shared a character and to choose the word with a shared character meaning different from the other two. For example, 画家 (huàjiā, ‘painter’), 作家 (zuòjiā, ‘writer’), and 大家 (dàjiā, ‘everyone’) share the character 家 (jiā, meaning ‘house’ or ‘family,’ or ‘professional’ when used independently). 大家 is a monomorphemic word in which the shared character’s meaning is not transparent, whereas the former two are bimorphemic words meaning ‘professional in doing something’, and 家 jiā functions like the suffix ‘–er’ in ‘writer’ in English. It follows that, to reason about the meaning carried by the shared character, the participants needed to analyze the morphological structure within each word, to discover that the shared character in 大家 does not bear a specific meaning that contributes to the whole word meaning as in the other two words, and to choose the correct answer 大家. One point was credited for each correct answer. There were 16 items in the intraword structure analysis. The Cronbach’s α was 0.80. Table 2: Sample items in L2 MA segmentation task Word type n Example English meaning Two-character monomorphemic 8 花生 Peanut Two-character bimorphemic 8 医学 The study of medicine Three-character monomorphemic 8 对不起 Sorry Three-character bimorphemic 8 运动员 Sports player Word type n Example English meaning Two-character monomorphemic 8 花生 Peanut Two-character bimorphemic 8 医学 The study of medicine Three-character monomorphemic 8 对不起 Sorry Three-character bimorphemic 8 运动员 Sports player Table 2: Sample items in L2 MA segmentation task Word type n Example English meaning Two-character monomorphemic 8 花生 Peanut Two-character bimorphemic 8 医学 The study of medicine Three-character monomorphemic 8 对不起 Sorry Three-character bimorphemic 8 运动员 Sports player Word type n Example English meaning Two-character monomorphemic 8 花生 Peanut Two-character bimorphemic 8 医学 The study of medicine Three-character monomorphemic 8 对不起 Sorry Three-character bimorphemic 8 运动员 Sports player L2 vocabulary knowledge. Following Anderson and Freebody (1981), we focused on a learner’s knowledge of word forms and meanings, and measured L2 vocabulary knowledge by two tasks, semantic word knowledge and morpheme knowledge. The semantic word knowledge task was based on Liu (2013), which asked participants to translate Chinese words into English meaning (n = 60). The Cronbach’s α was 0.95. In a similar vein, the morpheme knowledge task asked the participants to indicate Chinese affixes’ meanings or functions in English. Two types of morphemes, grammatical morphemes (n = 4) and word formation morphemes (n = 20), were included. The word formation morphemes were selected from Zeng’s (2008) database of productive morphemes in Chinese, consisting of 8 prefixoids and 12 suffixoids. The Cronbach’s α was 0.85. Analytical procedures. The analysis plan entailed a range of statistical analyses to test the aforementioned hypotheses—whether L2 vocabulary knowledge and L2 MA will contribute jointly or differentially when L2 Chinese readers infer the meanings of different types of multi-character unknown words. First, a two-by-two repeated-measures analysis of variance (ANOVA) would be performed to examine if there are any effects of word characteristics on L2 word-meaning inferencing, with morphological structure and base word familiarity as within-subject variables and L2 word-meaning inferencing as the dependent variable. Second, if there were any effects of word characteristics, separate regression analyses would be carried out for different word types with L2 vocabulary knowledge and L2 MA as independent variables, and L2 word-meaning inferencing as dependent variables, which was aimed to investigate the way in which L2 vocabulary knowledge and L2 MA contribute to L2 word-meaning inferencing. 4. RESULTS 4.1 Descriptive statistics Eight participants who did not complete all the tasks were removed from the data set. Another three cases were removed for they were identified as outliners at the exploratory data analysis phase. Therefore, all subsequent analyses performed in the study included 45 valid cases of the 56. The descriptive statistics are displayed in Table 3. Table 3: Descriptive statistics for all variables (N = 45) Variable M SD L2 lexical inferencing (correct out of 32) 17.71 5.21 L2 MA composite (z score-based) 0.00 1.72 Segmentation (correct out of 32) 23.04 5.14 Intraword structure analysis (correct out of 16) 9.93 3.19 L2 vocabulary knowledge composite (z score-based) 0.00 1.94 Semantic word knowledge (correct out of 60) 38.33 10.93 Morpheme knowledge (correct out of 24) 13.64 4.52 Variable M SD L2 lexical inferencing (correct out of 32) 17.71 5.21 L2 MA composite (z score-based) 0.00 1.72 Segmentation (correct out of 32) 23.04 5.14 Intraword structure analysis (correct out of 16) 9.93 3.19 L2 vocabulary knowledge composite (z score-based) 0.00 1.94 Semantic word knowledge (correct out of 60) 38.33 10.93 Morpheme knowledge (correct out of 24) 13.64 4.52 Table 3: Descriptive statistics for all variables (N = 45) Variable M SD L2 lexical inferencing (correct out of 32) 17.71 5.21 L2 MA composite (z score-based) 0.00 1.72 Segmentation (correct out of 32) 23.04 5.14 Intraword structure analysis (correct out of 16) 9.93 3.19 L2 vocabulary knowledge composite (z score-based) 0.00 1.94 Semantic word knowledge (correct out of 60) 38.33 10.93 Morpheme knowledge (correct out of 24) 13.64 4.52 Variable M SD L2 lexical inferencing (correct out of 32) 17.71 5.21 L2 MA composite (z score-based) 0.00 1.72 Segmentation (correct out of 32) 23.04 5.14 Intraword structure analysis (correct out of 16) 9.93 3.19 L2 vocabulary knowledge composite (z score-based) 0.00 1.94 Semantic word knowledge (correct out of 60) 38.33 10.93 Morpheme knowledge (correct out of 24) 13.64 4.52 4.2 Effects of morphological structure and base word familiarity Table 4 illustrates the participants’ performance in L2 word-meaning inferencing. There seems to be no notable difference across the four word types. A two (monomorphemic versus bimorphemic) by two (familiar versus unfamiliar base) repeated-measures ANOVA was carried out to analyze the data, with morphological structure and base word familiarity as within-subject variables. The results suggest that L2 word-meaning inferencing was not significantly affected the interaction between the two (F1, 44 = 0.16, p = .69), not by morphological structure only (F1, 44 = 0.71, p = .40), nor by base word familiarity only (F1, 44 = 0.46, p = .50). Table 4: Means in L2 word-meaning inferencing as a function of word types (N = 45) Word type M SD Monomorphemic-familiar 4.62 1.60 Monomorphemic-unfamiliar 4.40 1.72 Bimorphemic-familiar 4.38 1.61 Bimorphemic-unfamiliar 4.31 2.03 Word type M SD Monomorphemic-familiar 4.62 1.60 Monomorphemic-unfamiliar 4.40 1.72 Bimorphemic-familiar 4.38 1.61 Bimorphemic-unfamiliar 4.31 2.03 Note: Familiar, familiar word part/base; unfamiliar, unfamiliar word part/base. Table 4: Means in L2 word-meaning inferencing as a function of word types (N = 45) Word type M SD Monomorphemic-familiar 4.62 1.60 Monomorphemic-unfamiliar 4.40 1.72 Bimorphemic-familiar 4.38 1.61 Bimorphemic-unfamiliar 4.31 2.03 Word type M SD Monomorphemic-familiar 4.62 1.60 Monomorphemic-unfamiliar 4.40 1.72 Bimorphemic-familiar 4.38 1.61 Bimorphemic-unfamiliar 4.31 2.03 Note: Familiar, familiar word part/base; unfamiliar, unfamiliar word part/base. As shown in Figure 1 below, the data are slightly skewed. Following Larson-Hall (2010), data transformation was conducted by squaring the scores. Two-way repeated-measures ANOVA was rerun with morphological structure and familiarity with base words as within-subject variables. Yet, the results remained the same, while post hoc checking found no notable violation of assumptions (i.e. normal distribution and equal variance of data, normal distribution and equal variance of residuals). Effect sizes were calculated according to Field (2009). When the main effect of morphological structure was computed, r was 0.13; for familiarity with base words, r was 0.10. Both were relatively small. Given that there were no statistical effects of the two focal word characteristics, the following analysis only focuses on the contributions of L2 vocabulary knowledge and L2 MA to L2 word-meaning inferencing across all types of words. Figure 1: View largeDownload slide Performance in L2 word-meaning inferencing differentiated by word types Note: Familiar, familiar word part/base; unfamiliar, unfamiliar word part/base Figure 1: View largeDownload slide Performance in L2 word-meaning inferencing differentiated by word types Note: Familiar, familiar word part/base; unfamiliar, unfamiliar word part/base 4.3 Relative contributions of L2 vocabulary knowledge and L2 MA As shown in Table 5, there was moderate correlation between L2 MA (composite) and L2 word-meaning inferencing (r = 0.46, p < .01), and high correlation between L2 vocabulary knowledge (composite) and L2 word-meaning inferencing (r = 0.80, p < .001). Notably, the correlation between L2 MA and L2 vocabulary knowledge was high (r = 0.72, p < .001). The potential interaction effects between L2 MA and L2 vocabulary knowledge were examined from the multiple regression model results using the composite scores of L2 MA and L2 vocabulary knowledge, respectively. The rationale is as follows. First, as indicated in Table 5, there was a high correlation between the two subcomponent of L2 vocabulary knowledge (i.e. morpheme knowledge and semantic word knowledge) (r = 0.84), which might be due to the task modality. They were both tested in the written format, and orthographic knowledge or character knowledge was required to perform both tasks. Second, with respect to the two subcomponent of L2 MA (segmentation and intraword structure analysis), the correlational patterns, either among themselves, or with L2 vocabulary knowledge, or with L2 word-meaning inferencing, were compatible in general: (i) the magnitude of correlation between the two was large (r = 0.55) according to Cohen’s (1988) bench mark (0.1 being small, 0.25 being medium, and 0.4 being large). (ii) The correlation between segmentation and L2 vocabulary knowledge (composite) (r = 0.57), and that between intraword analysis and L2 vocabulary knowledge (composite) (r = 0.70) can both be counted as large effect sizes, and (iii) the correlation between segmentation and L2 word-meaning inferencing (r = 0.32) was not notably different from that between intraword analysis and L2 word-meaning inferencing (r = 0.49). In this regard, a composite score of L2 MA was used in subsequent regression analyses. Table 5: Bivariate correlations among variables (N = 45) Measure 1 2 3 4 5 6 7 8 9 10 11 12 1 L2 word-meaning inferencing – 2 L2 WMIMF .68*** – 3 L2 WMIMU .79*** .39** – 4 L2 WMIBF .66*** .24 .35* – 5 L2 WMIBU .84*** .42** .59*** .40** – 6 L2 MA (composite) .46** .16 .36* .33* .48** – 7 L2 segmentation .32* .08 .20 .22 .42** .89*** – 8 L2 intraword analysis .49** .21 .43** .36* .42** .87*** .55*** – 9 L2 vocabulary knowledge (composite) .80*** .50*** .66*** .53*** .67*** .72*** .57*** .70*** – 10 Semantic word knowledge .80*** .48** .64*** .57*** .69*** .74*** .57*** .73*** .96*** – 11 Morpheme knowledge .73*** .49** .63*** .45** .60*** .64*** .52*** .62*** .96*** .84*** – 12 L2 MA (composite) × L2 vocabulary knowledge (composite) .40** .47** .33* .06 .33* .31** .31* .24 .36* .30* .39** – Measure 1 2 3 4 5 6 7 8 9 10 11 12 1 L2 word-meaning inferencing – 2 L2 WMIMF .68*** – 3 L2 WMIMU .79*** .39** – 4 L2 WMIBF .66*** .24 .35* – 5 L2 WMIBU .84*** .42** .59*** .40** – 6 L2 MA (composite) .46** .16 .36* .33* .48** – 7 L2 segmentation .32* .08 .20 .22 .42** .89*** – 8 L2 intraword analysis .49** .21 .43** .36* .42** .87*** .55*** – 9 L2 vocabulary knowledge (composite) .80*** .50*** .66*** .53*** .67*** .72*** .57*** .70*** – 10 Semantic word knowledge .80*** .48** .64*** .57*** .69*** .74*** .57*** .73*** .96*** – 11 Morpheme knowledge .73*** .49** .63*** .45** .60*** .64*** .52*** .62*** .96*** .84*** – 12 L2 MA (composite) × L2 vocabulary knowledge (composite) .40** .47** .33* .06 .33* .31** .31* .24 .36* .30* .39** – Notes: WMIMF, word-meaning inferencing of monomorphemic words with familiar bases; WMIMU, word-meaning inferencing of monomorphemic words with unfamiliar bases; WMIUF, word-meaning inferencing of bimorphemic words with familiar bases; WMIBF, word-meaning inferencing of bimorphemic words with unfamiliar bases. * p < .05; **p < .01; ***p < .001. Table 5: Bivariate correlations among variables (N = 45) Measure 1 2 3 4 5 6 7 8 9 10 11 12 1 L2 word-meaning inferencing – 2 L2 WMIMF .68*** – 3 L2 WMIMU .79*** .39** – 4 L2 WMIBF .66*** .24 .35* – 5 L2 WMIBU .84*** .42** .59*** .40** – 6 L2 MA (composite) .46** .16 .36* .33* .48** – 7 L2 segmentation .32* .08 .20 .22 .42** .89*** – 8 L2 intraword analysis .49** .21 .43** .36* .42** .87*** .55*** – 9 L2 vocabulary knowledge (composite) .80*** .50*** .66*** .53*** .67*** .72*** .57*** .70*** – 10 Semantic word knowledge .80*** .48** .64*** .57*** .69*** .74*** .57*** .73*** .96*** – 11 Morpheme knowledge .73*** .49** .63*** .45** .60*** .64*** .52*** .62*** .96*** .84*** – 12 L2 MA (composite) × L2 vocabulary knowledge (composite) .40** .47** .33* .06 .33* .31** .31* .24 .36* .30* .39** – Measure 1 2 3 4 5 6 7 8 9 10 11 12 1 L2 word-meaning inferencing – 2 L2 WMIMF .68*** – 3 L2 WMIMU .79*** .39** – 4 L2 WMIBF .66*** .24 .35* – 5 L2 WMIBU .84*** .42** .59*** .40** – 6 L2 MA (composite) .46** .16 .36* .33* .48** – 7 L2 segmentation .32* .08 .20 .22 .42** .89*** – 8 L2 intraword analysis .49** .21 .43** .36* .42** .87*** .55*** – 9 L2 vocabulary knowledge (composite) .80*** .50*** .66*** .53*** .67*** .72*** .57*** .70*** – 10 Semantic word knowledge .80*** .48** .64*** .57*** .69*** .74*** .57*** .73*** .96*** – 11 Morpheme knowledge .73*** .49** .63*** .45** .60*** .64*** .52*** .62*** .96*** .84*** – 12 L2 MA (composite) × L2 vocabulary knowledge (composite) .40** .47** .33* .06 .33* .31** .31* .24 .36* .30* .39** – Notes: WMIMF, word-meaning inferencing of monomorphemic words with familiar bases; WMIMU, word-meaning inferencing of monomorphemic words with unfamiliar bases; WMIUF, word-meaning inferencing of bimorphemic words with familiar bases; WMIBF, word-meaning inferencing of bimorphemic words with unfamiliar bases. * p < .05; **p < .01; ***p < .001. Hierarchical regression analysis was performed to examine the main effects of L2 vocabulary knowledge and L2 MA as well as the interaction between the two on L2 word-meaning inferencing. Two sets of blockwise regression were carried out. First, L2 word-meaning inferencing was regressed on the order of (i) L2 MA, (ii) L2 vocabulary knowledge, and then (iii) the interaction between L2 MA and L2 vocabulary knowledge. Second, the order of L2 MA and L2 vocabulary knowledge was reversed, with L2 vocabulary knowledge entered first, L2 MA second, and the interaction between L2 MA and L2 vocabulary knowledge last. Reversing the order of L2 vocabulary knowledge and L2 MA helped to examine the unique variance explained by the new variable entered to the model while accounting for the other, which thus illustrates the relative contributions of L2 vocabulary knowledge and L2 MA, and/or the possible interaction between the two to L2 word-meaning inferencing. In Table 6, the results from Model 1 suggest that, in Step 1, the main effect of L2 MA was significant, explaining 21% of the variance of L2 word-meaning inferencing (R2 = 0.21, F1, 43 = 11.27, p < .01). In Model 1 Step 2, when controlling for L2 MA, L2 vocabulary knowledge significantly explained an additional 46% of the variance (R2 = 0.67, p < .001). Altogether, in Model 1, L2 MA and L2 vocabulary knowledge explained 67% of the variance. However, the results of Model 2 indicate that there was no significant effect of L2 MA after accounting for L2 vocabulary (R2 = 0.67, p = .06). When entered first in Model 2, L2 vocabulary alone significantly predicted 64% of the variance of L2 word-meaning inferencing (R2 = 0.64, p < .001). The result that either model analysis found any statistical effect of the interaction between L2 MA and L2 vocabulary knowledge, and that the main effect of L2 MA was significant when it was entered first, yet insignificant after accounting for L2 vocabulary knowledge, seems to suggest that there was a mediation effect of L2 vocabulary knowledge in L2 word-meaning inferencing. Table 6: Hierarchical regression results with L2 word-meaning inferencing as outcome (N = 45) Model Step Variable B R2 ΔR2 ΔF Model 1 1 L2 MA 2.76** 0.21 0.21 11.27** 2 L2 MA −1.51 0.67 0.46 58.66*** L2 vocabulary knowledge 5.26*** 3 L2 MA −1.60* 0.69 0.02 2.12 L2 vocabulary knowledge 5.05*** L2 MA × L2 vocabulary knowledge 0.86 Model 2 1 L2 vocabulary knowledge 4.30*** 0.64 0.64 76.27*** 2 L2 vocabulary knowledge 5.26*** 0.67 0.03 3.80 L2 MA −1.51 3 L2 vocabulary knowledge 5.05*** 0.69 0.02 2.12 L2 MA −1.60* L2 MA × L2 vocabulary knowledge 0.86 Model Step Variable B R2 ΔR2 ΔF Model 1 1 L2 MA 2.76** 0.21 0.21 11.27** 2 L2 MA −1.51 0.67 0.46 58.66*** L2 vocabulary knowledge 5.26*** 3 L2 MA −1.60* 0.69 0.02 2.12 L2 vocabulary knowledge 5.05*** L2 MA × L2 vocabulary knowledge 0.86 Model 2 1 L2 vocabulary knowledge 4.30*** 0.64 0.64 76.27*** 2 L2 vocabulary knowledge 5.26*** 0.67 0.03 3.80 L2 MA −1.51 3 L2 vocabulary knowledge 5.05*** 0.69 0.02 2.12 L2 MA −1.60* L2 MA × L2 vocabulary knowledge 0.86 * p < .05; **p < .01; ***p < .001. Table 6: Hierarchical regression results with L2 word-meaning inferencing as outcome (N = 45) Model Step Variable B R2 ΔR2 ΔF Model 1 1 L2 MA 2.76** 0.21 0.21 11.27** 2 L2 MA −1.51 0.67 0.46 58.66*** L2 vocabulary knowledge 5.26*** 3 L2 MA −1.60* 0.69 0.02 2.12 L2 vocabulary knowledge 5.05*** L2 MA × L2 vocabulary knowledge 0.86 Model 2 1 L2 vocabulary knowledge 4.30*** 0.64 0.64 76.27*** 2 L2 vocabulary knowledge 5.26*** 0.67 0.03 3.80 L2 MA −1.51 3 L2 vocabulary knowledge 5.05*** 0.69 0.02 2.12 L2 MA −1.60* L2 MA × L2 vocabulary knowledge 0.86 Model Step Variable B R2 ΔR2 ΔF Model 1 1 L2 MA 2.76** 0.21 0.21 11.27** 2 L2 MA −1.51 0.67 0.46 58.66*** L2 vocabulary knowledge 5.26*** 3 L2 MA −1.60* 0.69 0.02 2.12 L2 vocabulary knowledge 5.05*** L2 MA × L2 vocabulary knowledge 0.86 Model 2 1 L2 vocabulary knowledge 4.30*** 0.64 0.64 76.27*** 2 L2 vocabulary knowledge 5.26*** 0.67 0.03 3.80 L2 MA −1.51 3 L2 vocabulary knowledge 5.05*** 0.69 0.02 2.12 L2 MA −1.60* L2 MA × L2 vocabulary knowledge 0.86 * p < .05; **p < .01; ***p < .001. In light of the above, a subsequent analysis was carried out to examine the mediation effect of L2 vocabulary knowledge using SPSS 19.0 and an add-on tool Process Version 2.12.1 (Hayes 2013). Following Hayes (2013), four steps were taken to examine the mediator in the regression analysis (as demonstrated in Figure 2). In Step 1 of the mediation model, the regression of L2 MA on L2 word-meaning inferencing, ignoring the mediator L2 vocabulary knowledge, was significant, B = 1.38, p < .01. Step 2 showed that the regression of L2 MA on the mediator L2 vocabulary knowledge was also significant, B = 0.81, p < .001. Step 3 of the mediation process suggested that the mediator (L2 vocabulary knowledge), controlling for L2 MA, was significant, B = 2.63, p < .001. In Step 4, the analyses revealed that, controlling for the mediator (L2 vocabulary knowledge), L2 MA was not a significant predictor of L2 word-meaning inferencing, B = −0.76, p = .058. A Sobel test was conducted and found full mediation in the model (z = 5.06, p < .001, R2 mediation effect size = 0.18). Therefore, the aforementioned analysis indicated that L2 vocabulary knowledge fully mediated the relationship between L2 MA and L2 word-meaning inferencing, the indirect effect of L2 MA accounted for about 18% of the variance of L2 word-meaning inferencing. Figure 2: View largeDownload slide Steps to test mediation effect in regression analysis (adapted from Hayes, 2013) Note: IV, independent variable; M, mediator; DV, dependent variable Figure 2: View largeDownload slide Steps to test mediation effect in regression analysis (adapted from Hayes, 2013) Note: IV, independent variable; M, mediator; DV, dependent variable 5. DISCUSSION The results of this study add to our understanding of the respective roles of L2 vocabulary knowledge and L2 MA, as well as their interrelationship in L2 word-meaning inferencing. To reiterate, previous research has considered L2 vocabulary knowledge as the predominant predictor of successful L2 word-meaning inferencing, and there has been limited attention to how L2 vocabulary knowledge and L2 MA jointly contribute to the inferencing of different types of unknown words. Among those that did examine the relative contributions of vocabulary knowledge and MA in L2 reading, the focus was on reading comprehension as the outcome, and seemed to indicate there are both direct and indirect contributions of L2 MA via L2 vocabulary knowledge (Kieffer and Lesaux 2012a; Kieffer et al. 2013). Aligned with previous research on L2 word-meaning inferencing, L2 vocabulary knowledge is important in this study because it contributed directly and consistently to inferencing the meanings of different types of multi-character words. However, L2 vocabulary knowledge was not the sole significant predictor of L2 word-meaning inferencing; L2 MA made an additional unique contribution. L2 vocabulary knowledge and L2 MA together explained around 67% in the variance of L2 word-meaning inferencing; L2 MA contributed indirectly to L2 word-meaning inferencing through the mediation of L2 vocabulary knowledge, which accounted for about 18% of the variance in L2 word-meaning inferencing. Collectively, to answer the question whether vocabulary knowledge alone is sufficient for L2 word-meaning inferencing, the findings of this study seem to indicate that L2 vocabulary knowledge is necessary, but in itself insufficient, in predicting L2 word-meaning inferencing ability. It is certainly not surprising to find the direct effect of L2 vocabulary knowledge on L2 word-meaning inferencing, since this study measured both semantic word knowledge and morpheme knowledge of adult L2 Chinese readers, and it was more likely for those with larger vocabulary size as well as better insights of words to be successful in identifying unknown word meanings. On the other hand, the indirect effect of L2 MA mediated by L2 vocabulary knowledge merits further discussion. As mentioned earlier, previous studies disagreed about the relative contributions of L2 MA and L2 vocabulary knowledge to L2 reading outcomes (mostly reading comprehension): some found a significant direct effect of MA over and above vocabulary knowledge (Kieffer and Lesaux 2008; Jeon 2011), some found an indirect effect of MA through vocabulary knowledge (Zhang and Koda 2012; Goodwin et al. 2013), and still others observed both direct and indirect effects of L2 MA with full or partial mediation of vocabulary knowledge (Kieffer and Lesaux 2012a; Kieffer et al. 2013). In a study with Chinese university EFL students, Zhang and Koda (2012) observed no significant direct effect of MA when considering vocabulary knowledge, but found a full mediation effect of vocabulary knowledge. They postulated that part of the reason might be that they measured multiple dimensions of vocabulary knowledge (vocabulary size and vocabulary depth), and thus it strengthened the contribution of vocabulary knowledge and limited that of MA, whereas some of previous studies measured vocabulary size only (Kieffer and Lesaux 2008, 2012a; Jeon 2011; Kieffer et al. 2013). Corroborating the results presented by Zhang and Koda (2012), this study measured the different dimensions of vocabulary knowledge (i.e. semantic word knowledge and morpheme knowledge) of adult L2 Chinese readers, which might amplify the significance of vocabulary knowledge in L2 word-meaning inferencing and limit that of MA. Another possibility for not observing any significant direct effect of MA beyond vocabulary knowledge is the way in which MA was measured. Different from this study, Kieffer and colleagues (Kieffer and Lesaux 2012a; Kieffer et al. 2013) observed both direct and indirect effects of MA on reading comprehension in English among adolescent Spanish-speaking ELLs. For one thing, they only measured (reading) vocabulary size; for another, their measure of MA in the target language (i.e. English) was modeled after Carlisle’s (2000) at the sentence level. For example, the participants were provided with a word with a derivational suffix (e.g. complexity) and asked to extract the base word (e.g. complex) to complete a sentence (e.g. The problem is ___ .) (adapted from Kieffer et al. 2013). This kind of measure might indicate the reader’s ability to detect syntactic signals (e.g. syntactic information conveyed by suffixes) (Nagy 2007), and in doing so, MA can make a direct contribution to reading comprehension for its facilitation effect on syntactic processing. However, this syntactic aspect might be language-specific. In the present study, L2 MA in Chinese was measured at the word level. In Chinese, it is unlikely that readers will determine the syntactic status of a stand-alone morphologically complex word as in English, for a lot of the syntactic markers cannot be found (e.g. case and number markers in nouns, as well as tense and aspect markers in verbs, Li and Thompson 1981). Rather, the syntactic category of a word is determined mainly by word order in the sentence. In view of above, the respective measurement of L2 vocabulary knowledge and L2 MA might have contributed to current findings that the contribution of L2 MA to L2 word-meaning inferencing was indirect and fully mediated by L2 vocabulary knowledge. 6. CONCLUSIONS AND LIMITATIONS The present study investigated whether L2 vocabulary knowledge alone is sufficient for L2 word-meaning inferencing. Specifically, it examined the contributions of L2 vocabulary knowledge and L2 MA in inferring the meanings of different types of multi-character unknown words in L2 Chinese. The results suggested that L2 vocabulary knowledge was necessary but insufficient for successful L2 word-meaning inferencing, since (i) the interrelationship between L2 vocabulary knowledge and L2 word-meaning inferencing was consistent across different types of multi-character words, at least not influenced by morphological structure and familiarity with base words, (ii) L2 vocabulary knowledge played an important role due to its direct and mediation effects, and (iii) L2 MA made an additional contribution through the mediation of L2 vocabulary knowledge. The findings of this study add to our understanding of the contribution of L2 MA to L2 word-meaning inferencing after accounting for L2 vocabulary knowledge, as well as the complex intralingual relationship between L2 MA and vocabulary knowledge in L2 word-meaning inferencing. Resonating with previous research on later reading skill development, this study did not observe any direct contribution of L2 MA in adult L2 readers of Chinese, but an indirect contribution of L2 MA through the mediation of L2 vocabulary knowledge. Also, this study was among the first to examine potential differences in inferring the meanings of morphologically complex and simple words in Chinese as a second language. Given that many of the Chinese written words are multi-character and vary in their morphological structure, it is postulated that L2 MA is especially useful for L2 Chinese readers because it helps them to segment words into smaller meaningful units and extract meanings from unknown lexical units in reading Chinese. However, only a very small, but not statistically significant, impact of word effects was observed. It is noted that the present study has some limitations to be addressed in future research. First, it only measured the accuracy of word-meaning inferencing. To examine potential word effects, examining both accuracy and speed of word-meaning inferencing might be a more sensitive and appropriate measure. Second, due to the relatively small sample size, mean- and correlation-based comparisons were made in the analysis. To examine the interaction between L2 MA and vocabulary knowledge in predicting L2 word-meaning inferencing, more sophisticated statistical analyses (e.g. path analysis or structural equation modeling) with a larger participant pool and item pool is needed for future replication studies. Third, other word characteristics might also need to be considered in future research. Following Reichle and Perfetti (2003), this study adopted a relatively straightforward form (orthography)-meaning approach and focused on two word characteristics when investigating L2 Chinese reading: morphological structure and base word familiarity. But it did not control strictly for other word characteristics, for example, word length and word syntactic category (see Dronjic 2011). Finally, future studies might consider including control variables like orthographic knowledge. For instance, Everson (1998) found a very strong convergence between non-heritage Chinese learners’ ability to name written words and their knowledge of the meanings of those words. When investigating how L2 MA contributed to L2 word-meaning inferencing, this study measured MA in the written format, which differed from previous studies with monolingual children that typically measured MA in oral tasks. Given that the target population in the present study (i.e. adult L2 readers of Chinese) started learning to speak and read in the target language concurrently, using written tasks was perceived to be more appropriate and pertaining to their learning experience. In future research, there is a need to examine the potential effect of orthographic knowledge in relation to MA and vocabulary knowledge in L2 word-meaning inferencing in learners with different proficiency levels and literacy experiences in the target language. Sihui Ke received her PhD in Second Language Acquisition at Carnegie Mellon University. Her research has focused on second language reading and biliteracy development, and foreign language assessment. Currently, she is involved in studies of the contributions of morphological awareness to adult and child second language reading. Address for correspondence: Room G419, School of Humanities and Social Science, Harbin Institute of Technology (Shenzhen), Xili University Town, HIT Campus, Nanshan District, Shenzhen 518055, P. R. China. <echoecho.ke@gmail.com> or <kesihui@hit.edu.cn> Keiko Koda is a professor in the Department of Modern Languages at Carnegie Mellon University. Her research interests include second language reading, biliteracy development, and foreign language instruction and assessment. REFERENCES Anderson R. C. , Freebody P. . 1981 . ‘Vocabulary knowledge’ in Guthrie J. T. (ed.): Comprehension and Teaching: Research Reviews . International Reading Association , pp. 77 – 117 . Andrews S. , Bond R. . 2009 . ‘ Lexical expertise and reading skill: Bottom-up and top-down processing of lexical ambiguity ,’ Reading and Writing: An Interdisciplinary Journal 22 / 6 : 687 – 711 . Google Scholar CrossRef Search ADS Anglin J. M. , Miller G. A. , Wakefield P. C. . 1993 . ‘ Vocabulary development: A morphological analysis ,’ Monographs of the Society for Research in Child Development 58 / 10 : 1 – 187 . Google Scholar CrossRef Search ADS Bai X. , Yan G. , Zang C. , Liversedge S. P. , Rayner K. . 2008 . ‘ Reading spaced and unspaced Chinese text: Evidence from eye movements ,’ Journal of Experimental Psychology: Human Perception and Performance 34 / 5 : 1277 – 87 . Google Scholar CrossRef Search ADS PubMed Bialystok E. 2001 . Bilingualism in Development . Cambridge University Press . Google Scholar CrossRef Search ADS Carlisle J. F. 2000 . ‘ Awareness of the structure and meaning of morphologically complex words: Impact on reading ,’ Reading and Writing: An Interdisciplinary Journal 12 : 169 – 90 . Google Scholar CrossRef Search ADS Chen J. 1999 . ‘Word recognition during the reading of Chinese sentences: Evidence from studying the word superiority effect’ in Wang J. , Inhoff A. W. , Chen H.-C. (eds): Reading Chinese Script: A Cognitive Analysis . Erlbaum , pp. 239 – 56 . Chinese Proficiency Test Center . 2001 . Grading Syllabus for Chinese Vocabulary and Chinese Characters [in Chinese]. Economy and Science Press . Christianson K. , Luke S. G., . 2011 . ‘ Context strengthens initial misinterpretations of text ,’ Scientific Studies of Reading 15 / 2 : 136 – 66 . Google Scholar CrossRef Search ADS Chung F. H -K. , Leung M. -T., . 2008 . ‘ Data analysis of Chinese characters in primary school corpora of Hong Kong and mainland China: Preliminary theoretical interpretations ,’ Clinical Linguistics and Phonetics 22 / 4–5 : 379 – 89 . Google Scholar CrossRef Search ADS PubMed Cohen J. 1988 . Statistical Power Analysis for the Behavioral Sciences , 2nd edn Erlbaum . de Bot K. , Paribakht T. S. , Wesche M. B., . 1997 . ‘ Toward a lexical processing model for the study of second language vocabulary acquisition ,’ Studies in Second Language Acquisition 19 / 3 : 309 – 29 . Google Scholar CrossRef Search ADS Deacon S. H. , Kieffer M. J. , Laroche A., . 2014 . ‘ The relation between morphological awareness and reading comprehension: Evidence from mediation and longitudinal models ,’ Scientific Studies of Reading 18 : 432 – 51 . Google Scholar CrossRef Search ADS Dronjic V. 2011 . ‘ Mandarin Chinese compounds, their representation, and processing in the visual modality ,’ Writing Systems Research 3 / 1 : 5 – 21 . Google Scholar CrossRef Search ADS Everson M. E. 1998 . ‘ Word recognition among learners of Chinese as a foreign language: Investigating the relationship between naming and knowing ,’ Modern Language Journal 82 / 2 : 194 – 204 . Google Scholar CrossRef Search ADS Field A. 2009 . Discovering Statistics Using SPSS , 3rd edn Sage . Gilbert J. K. , Goodwin A. P. , Compton D. L. , Kearns D. M. . 2013 . ‘ Multisyllabic word reading as a moderator of morphological awareness and reading comprehension ,’ Journal of Learning Disabilities 47 / 1 : 34 – 43 . Google Scholar CrossRef Search ADS PubMed Goodwin A. P. , Huggins A. C. , Carlo M. S. , August D. , Calderon M., . ( 2013 ). ‘ Minding morphology: How morphological awareness relates to reading for English language learners ,’ Reading and Writing: An Interdisciplinary Journal 26 / 9 : 1387 – 141 . Google Scholar CrossRef Search ADS Haastrup K. 1991 . Lexical Inferencing Procedures or Talking about Words . Gunter Narr Verlag . Hamada M. 2014 . ‘ The role of morphological and contextual information in L2 lexical inference ,’ The Modern Language Journal 98 / 4 : 992 – 1005 . Google Scholar CrossRef Search ADS Hatami S. , Tavakoli M., . 2012 . ‘ The role of depth versus breadth of vocabulary knowledge in success and ease in L2 lexical inferencing ,’ TESL Canada Journal 31 / 1 : 1 – 21 . Hayes A. F. 2013 . Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-based Approach . Guilford Press . Hu M. , Nation I. S. P., . 2000 . ‘ Unknown vocabulary density and reading comprehension ,’ Reading in a Foreign Language 13 : 403 – 30 . Jeon E. H. 2011 . ‘ Contribution of Morphological awareness to second-language reading comprehension ,’ The Modern Language Journal 95 / 2 : 217 – 35 . Google Scholar CrossRef Search ADS Kieffer M. J. , Biancarosa G. , Mancilla-Martinez J., . 2013 . ‘ Roles of morphological awareness in the reading comprehension of Spanish-speaking language minority learners: Exploring partial mediation by vocabulary and reading fluency ,’ Applied Psycholinguistics 34 / 04 : 697 – 725 . Google Scholar CrossRef Search ADS Kieffer M. J. , Lesaux N. K. , 2008 . ‘ The role of derivational morphology in the reading comprehension of Spanish-speaking English language learners ,’ Reading and Writing: An Interdisciplinary Journal 21 / 8 : 783 – 804 . Google Scholar CrossRef Search ADS Kieffer M. J. , Lesaux N. K., . 2012a . ‘ Direct and indirect roles of morphological awareness in the English reading comprehension of native English, Spanish, Filipino, and Vietnamese speakers ,’ Language Learning 62 / 4 : 1170 – 204 . Google Scholar CrossRef Search ADS Kieffer M. J. , Lesaux N. K., . 2012b . ‘ Development of morphological awareness and vocabulary knowledge in Spanish-speaking language minority learners: A parallel process latent growth curve model ,’ Applied Psycholinguistics 33 : 23 – 54 . Google Scholar CrossRef Search ADS Koda, K. 2000 . ‘ Cross-linguistic variations in L2 morphological awareness ,’ Applied Psycholinguistics 21 : 297 – 320 . Koda, K. 2007 . ‘ Reading and language learning: cross-linguistic constraints on second language reading development ,’ Language Learning 57 / S1 : 1 – 44 . Koda K. , Lű C. , Zhang D., . 2014 . ‘L1-induced facilitation in biliteracy development in Chinese and English’ in Chen X. , Wang Q. , Luo Y. C. (eds): Reading Development and Difficulties in Monolingual and Bilingual Chinese Children . Springer . Ku Y.-M. 2001 . ‘Development of morphological awareness in Chinese and English,’ Unpublished doctoral dissertation, University of Illinois at Urbana-Champaign. Ku Y. -M. , Anderson R. C., . 2003 . ‘ Development of morphological awareness in Chinese and English ,’ Reading and Writing: An Interdisciplinary Journal 16 : 399 – 422 . Google Scholar CrossRef Search ADS Kuo L. -J. , Anderson R. C., . 2008 . ‘Conceptual and methodological issues in comparing metalinguistic awareness across languages’ in Koda K. , Zehler A. (eds): Learning to Read Across Languages . Routledge , pp. 39 – 67 . Larson-Hall J. 2010 . A Guide to Doing Statistic in Second Language Research Using SPSS . Routledge . Li C. N. , Thompson S. A. . 1981 . Mandarin Chinese: A Functional Reference Grammar . University of California Press . Li X. S. , Bicknell K. , Liu P. P. , Wei W. , Rayner K., . 2014 . ‘ Reading is fundamentally similar across disparate writing systems: A systematic characterization of how words and characters influence eye movements in Chinese reading ,’ Journal of Experimental Psychology General 143 / 2 : 895 – 913 . Google Scholar CrossRef Search ADS PubMed Lin T. -J. , Anderson R. J. , Ku Y. -M. , Christianson K. , Packard J. L., . 2011 . ‘ Chinese children’s concept of word ,’ Writing Systems Research 3 / 1 : 41 – 57 . Google Scholar CrossRef Search ADS Liu N. , Nation P., . 1985 . ‘ Factors affecting guessing vocabulary in context ,’ RECL Journal 16 : 33 – 42 . Liu P. -P. , Li W. -J. , Lin N. , Li X. -S. . 2013 . ‘ Do Chinese readers follow the national standard rules for word segmentation during reading? ,’ PLoS One 8 / 2 : e55440 . Google Scholar CrossRef Search ADS PubMed Liu Y. 2013 . ‘L1 reading ability, L2 linguistic knowledge, and L2 semantic gap filling ability,’ Unpublished doctoral dissertation, Carnegie Mellon University. Mcbride-Chang C. , Tardif T. , Cho J. -R. , Shu H. , Fletcher P. , Stokes S. F. , Wong A. , Leung K. . 2008 . ‘ What’s in a word? Morphological awareness and vocabulary knowledge in three languages ,’ Applied Psycholinguistics , 0329 : 437 – 62 . Meara P. 1996 . ‘The dimensions of lexical competence’ in Brown G , Malmkjaer K. , Williams J. (eds): Competence and Performance in Language Learning . Cambridge University Press Mori Y. , Nagy W. . 1999 . ‘ Integration of information from context and word elements in interpreting novel kanji compounds ,’ Reading Research Quarterly 34 / 1 : 80 – 101 . Google Scholar CrossRef Search ADS Myers J. 2006 . ‘Processing Chinese compounds: A survey of the literature’ in Libben G. , Jarema G. (eds): The Representation and Processing of Compound Words . Oxford University Press , pp. 169 – 96 . Nagy W. 2007 . ‘Metalinguistic awareness and the vocabulary-comprehension connection’ in Wagner R. , Muse A. , Tannenbaum K. (eds): Vocabulary Acquisition: Implications for Reading Comprehension . Guilford Press , pp. 53 – 77 . Nagy W. , Berninger V. W. , Abbott R. D., . 2006 . ‘ Contributions of morphology beyond phonology to literacy outcomes of upper elementary and middle-school students ,’ Journal of Educational Psychology 98 / 1 : 134 – 47 . Google Scholar CrossRef Search ADS Nassaji H. 2003 . ‘ L2 vocabulary learning from context: Strategies, knowledge sources, and their relationship with success in L2 lexical inferencing ,’ TESOL Quarterly 37 / 4 : 645 – 70 . Google Scholar CrossRef Search ADS Nassaji H. 2004 . ‘ The relationship between depth of vocabulary knowledge and L2 learners' lexical inferencing strategy use and success ,’ The Canadian Modern Language Review 61 / 1 : 107 – 34 . Google Scholar CrossRef Search ADS Nation I. S. P. 1990 . Teaching and Learning Vocabulary . Newbury House . Nation I. S. P. 2001 . Learning Vocabulary in Another Language . Cambridge University Press . Google Scholar CrossRef Search ADS Nation I. S. P. 2006 . ‘ How large a vocabulary is needed for reading and listening? ,’ The Canadian Modern Language Review 63 / 1 : 59 – 81. . Google Scholar CrossRef Search ADS Packard J. L. 2000 . The Morphology of Chinese . Cambridge University Press . Google Scholar CrossRef Search ADS Parel R. 2004 . ‘ The impact of lexical inferencing strategies on second language reading proficiency ,’ Reading and Writing: An Interdisciplinary Journal 17 / 6 : 847 – 73 . Google Scholar CrossRef Search ADS Paribakht T. S. , Wesche M., . 1999 . ‘' Incidental' and instructed L2 vocabulary acquisition through reading: An introspective study ,’ Studies in Second Language Acquisition 21 / 2 : 195 – 224 . Google Scholar CrossRef Search ADS Park E. C. 2004 . ‘The Relationship between Morphological Awareness and Lexical Inference Ability for English Language Learning Children with Korean First-language Background,’ Unpublished doctoral dissertation, Carnegie Mellon University. Peng R. -Y. 2004 . ‘ Even words are right, odd ones are odd: Explaining word segmentation inconsistency among Chinese readers ,’ Chinese Journal of Psychology 46 / 1 : 49 – 55 . Qian D. D. 2005 . ‘ Demystifying lexical inferencing: The role of aspects of vocabulary knowledge ,’ TESL Canada Journal 22 / 2 : 34 – 54 . Google Scholar CrossRef Search ADS Read J. 1993 . ‘ The development of a new measure of L2 vocabulary knowledge ,’ Language Testing 10 / 3 : 355 – 71 . Google Scholar CrossRef Search ADS Read J. 2000 . Assessing Vocabulary . Cambridge University Press . Google Scholar CrossRef Search ADS Reichle E. D. , Perfetti C., . 2003 . ‘ Morphology in word identification: A word-experience model that account for morpheme frequency effect ,’ Scientific Studies of Reading 7 / 3 : 219 – 37 . Google Scholar CrossRef Search ADS Schmitt N. , Jiang X. , Grabe W. . 2012 . ‘The percentage of words known in a text and reading comprehension ,’ Modern Language Journal 95 / 1 : 26 – 43 . Google Scholar CrossRef Search ADS Shu H. , Chen X. , Anderson R. C. , Wu N. , Xuan Y., . 2003 . ‘ Properties of school Chinese: Implications for learning to read ,’ Child Development 74 / 1 : 27 – 47 . http://dx.doi.org/10.1023/A:1025334204364. Google Scholar CrossRef Search ADS PubMed Verhoeven L. , Perfetti C. A., . 2011 . ‘ Morphological processing in reading acquisition: A cross-linguistic perspective ,’ Applied Psycholinguistics 32 / 3 : 457 – 66 . Google Scholar CrossRef Search ADS Xu X. 2004 . ‘The morphological awareness of overseas students in the acquisition of Chinese compound words,’ Unpublished master thesis, Beijing Language and Culture University. Zeng L. 2008 . ‘ Quantitative and qualitative study of Chinese semi-affixes ,’ World Chinese Language Teaching 4 : 75 – 87 . Zhang, D. and Koda, K . 2012 . ‘ Contribution of morphological awareness and lexical inferencing ability to L2 vocabulary knowledge and reading comprehension among advanced EFL learners: Testing direct and indirect effects ,’ Reading and Writing 25 / 5 : 1195 – 216 . © Oxford University Press 2017

Journal

Applied LinguisticsOxford University Press

Published: Nov 16, 2017

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off