Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Morphological representation and reading comprehension in early elementary English language learners O'Toole, Gillian Laura 2018

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2018_may_o'toole_gillian .pdf [ 8.34MB ]
Metadata
JSON: 24-1.0365823.json
JSON-LD: 24-1.0365823-ld.json
RDF/XML (Pretty): 24-1.0365823-rdf.xml
RDF/JSON: 24-1.0365823-rdf.json
Turtle: 24-1.0365823-turtle.txt
N-Triples: 24-1.0365823-rdf-ntriples.txt
Original Record: 24-1.0365823-source.json
Full Text
24-1.0365823-fulltext.txt
Citation
24-1.0365823.ris

Full Text

MORPHOLOGICAL REPRESENTATION AND READING COMPREHENSION IN EARLY ELEMENTARY ENGLISH LANGUAGE LEARNERS by  Gillian Laura O’Toole  B.A., Simon Fraser University, 2014  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF SCIENCE in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Audiology and Speech Sciences)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  April 2018  © Gillian Laura O’Toole, 2018   ii Abstract Reading is a complex cognitive-linguistic endeavor involving multiple levels of linguistic knowledge. The relative contribution of different levels of linguistic awareness may depend on the individual linguistic processes of a child’s linguistic background. A child learning English as a second language may face differences as they become proficient in their educational language which is essential for academic success. The purpose of the study is to gain insight into the various levels of linguistic knowledge associated with reading comprehension in children who are English language learners (ELLs). This study examined morphological knowledge in relation to reading comprehension between grade two students who learned English as first language (EL1) and ELLs. The participants completed tasks of English phonological awareness, word reading, vocabulary, and reading comprehension. Further, the participants were assessed on their knowledge of English morphology through an experimental task based on a distributed connectionist approach: the Morphological Awareness Semantic Task (MAST). This task requires students to process sentences with different morphological and semantic conditions: form only, low semantic, moderate semantic, high semantic, and semantic only. No significant differences between language groups on any measures were found, except on receptive vocabulary. Statistically significant correlations were found between phonological awareness and word reading for the ELL group, and between word reading and reading comprehension in the EL1 group, but not between morphological representations and reading. In an analysis of the five different MAST conditions, the groups did not display any statistically significant differences between them, with some notable differences: the EL1 group tended to score higher than the ELL group on all conditions, except for the high semantic condition. The overall findings of the   iii current study suggest that ELLs are able to achieve similar scores in reading comprehension to that of native English speakers, with minor differences in their morphological representations of English words. Despite having a smaller vocabulary size than native English speaking students, ELLs appear to be relying on different linguistic skills to decode and derive meaning in a morphologically complex task.      iv Lay Summary Understanding how children with different language backgrounds develop reading comprehension in the early elementary years can help inform how to support future academic success. Investigating reading comprehension from a theoretical framework arguing that children are supported in literacy through the integration of all levels of language – from sounds and how they relate to letters, how words are formed, their role within a sentence, and they construct meaning. Learning more about language development provides insight into the cognitive processes involved in bilingualism and second language learners, who are very common in the elementary schools of the lower mainland.               v Preface This is original, unpublished, independent work by the author Gillian Laura O’Toole. Data collection was completed at the University of Toronto and the members of Dr. Becky Chen’s Laboratory. Gillian Laura O’Toole, with the help of Dr. Stefka Marinova-Todd, was responsible for research questions, data analysis, and manuscript composition.                     vi Table of Contents  Abstract ...................................................................................................................................... ii Lay Summary ............................................................................................................................ iv Preface ....................................................................................................................................... v List of Tables .......................................................................................................................... viiii List of Figures ........................................................................................................................... ix Acknowledgements .................................................................................................................... x Dedication ................................................................................................................................. xi Chapter 1: Introduction ............................................................................................................... 1 1.1 Purpose ............................................................................................................................. 1 1.2 Models of Reading ............................................................................................................ 2 1.3 Metalinguistic Awareness and Knowledge ........................................................................ 3 1.3.1 Monolingual English Speakers and Readers ............................................................... 3 1.3.2 Monolingual Speakers and Readers of Other Languages............................................. 6 1.3.3 Bilingual Speakers and Readers .................................................................................. 8     1.4 Summary of Previous Research ....................................................................................... 11 1.5 Morphological Representation......................................................................................... 12 1.6 Morphological Awareness Semantic Task ....................................................................... 16 1.7 Research Questions ......................................................................................................... 18 Chapter 2: Methods .................................................................................................................. 20 2.1 Participants ..................................................................................................................... 20   vii 2.2 Procedure ........................................................................................................................ 23 2.2.1 Measures .................................................................................................................. 23 2.2.2 Data Analysis ........................................................................................................... 27 Chapter 3: Results ..................................................................................................................... 28 Chapter 4: Discussion ............................................................................................................... 34 4.1 English Morphological Representation of ELL participants compared to EL1s ................ 35 4.2 English Morphological Represenatation in relation to English Reading Comprehension .. 40 4.3 Limitations ...................................................................................................................... 42 4.4 Directions for Future Research ........................................................................................ 43 4.5 Significance of Findings.................................................................................................. 44 Bibliography ............................................................................................................................. 45 Appendix .................................................................................................................................. 50 1.1 MAST Measure............................................................................................................... 50              viii List of Tables Table 2.1 Participants ............................................................................................................... 20 Table 2.2 Participant Demographics ......................................................................................... 21 Table 2.3 Guardian 1 Demographics ......................................................................................... 21 Table 2.4 Other Relevant Demographics ................................................................................... 22 Table 2.5 MAST Conditions ..................................................................................................... 24 Table 2.6 MAST Conditions Description .................................................................................. 25 Table 3.1 Standard Scores (SS) and Percent Correct (%) Results .............................................. 28 Table 3.2 Independent Samples t-Tests ..................................................................................... 29 Table 3.3 Correlations and 2-tailed Significance Levels ............................................................ 30        ix List of Figures Figure 3.1 ELL Scatterplot........................................................................................................ 30 Figure 3.2 EL1 Scatterplot ........................................................................................................ 31 Figure 3.3 Percentage Correct of MAST response trends organized by language group with standard error bars .................................................................................................................... 32            x Acknowledgements I would like to acknowledge Dr. Stefka Marinova-Todd for having mentored and supported me throughout the process of my thesis. I would also like to acknowledge her kindness, encouragement, and the independence she trusted me to have.   Dr. Paola Colozzo, my second committee member, provided helpful and insightful feedback and contributing a unique perspective regarding her ideas for our project.   Dr. Becky Chen and her students at the University of Toronto were immensely helpful with providing support through answering questions I had about their data sets and procedures.   Finally, I would like to acknowledge my friends and family for their unconditional support.            xi Dedication I dedicate my thesis to my family, Gerard, Linda, and Michael O’Toole, who have supported me in every way throughout my education. Without them I would have not had the opportunity to pursue Speech Language Pathology and work in a field I am passionate about. I also dedicate my thesis to my classmates and professors who have also guided me throughout my time at the University of British Columbia. Finally, a special thank you to my partner, Ethan Sadowski, and my classmates Alex Friesen, Courtney Lawrence, and Kelsea Ross for supporting me throughout our education together.                  1 Chapter 1: Introduction  1.1 Purpose Reading is a complex cognitive-linguistic endeavor that requires the ability to decode and comprehend written material (Kamhi & Catts, 2012). Reading comprehension involves multiple levels of linguistic knowledge and awareness, including phonological awareness, morphological awareness, orthographic awareness, knowledge of sentence structure, and vocabulary. The relative contribution of each of these factors may depend on the individual linguistic processes of different languages. English, for example, is language that uses morphophonology. In the construction of words, phonemes interact with morphological units which may change the way they are expressed, which is reflected in the English alphabetic orthography. Not all languages express sound and meaning relationships in these processes. As such, a child learning English as a second language may face differences as they become proficient in English reading comprehension, a skill that is essential for academic success. In Canada, as well as around the world, many children are receiving education and are learning to read in a language other than their first language (L1). In British Columbia, 15.3% of elementary students are learning English as a second language (British Columbia Ministry of Education, 2017). Vancouver elementary schools have an even higher statistic, where 27.9% of elementary students are English language learners (ELLs) (British Columbia Ministry of Education, 2017). Within the school system, the Language Education Policy of British Columbia states “…all students are expected to achieve proficiency in the English language” (British Columbia Ministry of Education, 2004 Revised), highlighting the importance of each student becoming capable in their language of instruction, regardless of their language   2 background. The purpose of the current study is to gain insight into the various levels of linguistic knowledge and awareness associated with reading comprehension in children who are learning English as a second language. Further, this study will examine morphological representation in relation to reading comprehension – through a task which reflects a distributed connectionist perspective attempting to account for the various levels of linguistic knowledge that contribute to reading comprehension – between ELLs and those who have learned English as their first language (EL1s).   1.2 Models of Reading  Currently, there are three cognitive models of reading acknowledged in the development of literacy: the (1) Top-Down Model, the (2) Bottom-Up Model, and (3) the Interactive Model. The Top-Down model proposes that, as a child is reading, they are using their prior knowledge to test hypotheses about what they think the text means (Vacca, Vacca, Gove, Burkey, Lenhart & McKeon, 2009). The Bottom-Up Model claims that children decode the written text, from analyzing the sounds and letters, up to words and sentences, and eventually text level information (Vacca, Vacca, Gove, Burkey, Lenhart & McKeon, 2009). Finally, in the most widely accepted model, the Interactive Model accounts for both top-down and bottom-up processes. Children are said to integrate their prior knowledge with their print knowledge, and make hypotheses about what they think the text means based on what they know about phonology, morphology, syntax, and semantics (Vacca, Vacca, Gove, Burkey, Lenhart & McKeon, 2009).      3 1.3 Metalinguistic Awareness and Knowledge 1.3.1 Monolingual English Speakers and Readers Morphology manifests itself in different ways across languages and orthographies. A particular language may use a high amount of compounding morphology, for example Mandarin, in which complex words are formed via combining existing words together. An example of this process in English is with the word “firewood”, which is comprised of two independent morphological units which are words (“fire” and “wood”) that carry meaning when on their own (Haspelmath & Sims, 2010). Typically, English is a morphophonemic language and uses derivational and inflectional morphemes attached to a root word to change its meaning. For example, the word “fire” may be inflectionally modified with plural “-s” or derivationally with “-ed” to yield different meanings.   Previous studies that have examined English-speaking students’ phonological awareness, morphological awareness, vocabulary, and reading comprehension have generally found that phonological awareness predicts reading at the word level, while morphological awareness predicts reading at the word level and beyond (Deacon & Kirby, 2004; McBride-Chang, Wagner, Muse, Chow & Shu, 2005). To investigate linguistic awareness and expressive vocabulary knowledge McBride-Chang et al. (2005) probed morphological awareness, using a connectionist perspective, through two tasks: a structure awareness task and morpheme identification task. Knowledge of morphological structures was accessed by having children decide which target morpheme corresponded best to an image between two choices. An example of this is the child deciding whether a picture with the colour “blue” or a boy who “blew” out a candle best fit with the target morpheme “blue” in blueberry. The participants also identified morphemes through observing a sequence of two to four pictures   4 and choosing the correct object or concept name that fit the sequences. In this task, the researchers described the sequence to the child. For example, the examiner described the sequence “Early in the morning, we can see the sun coming up. This is called a sunrise. At night, we might also see the moon coming up. What could we call this?” (pg. 423), to which the correct answer would be “moonrise” (McBride-Chang et al., 2005). The combined measure of morphological awareness was found to be a good predictor of expressive vocabulary in monolingual native English speaking Kindergarteners and second graders in the correlational study (McBride-Chang et al., 2005). McBride-Chang et al. (2005) further stress two significant considerations suggested from their findings: morphological awareness is a “cognitive construct separable from phonological processing and reading skills” (pg. 428), and awareness of morpheme identification and knowledge of morphological structures’ important role in the development of vocabulary – a skill paramount in the development literacy. With regards to reading comprehension, Deacon and Kirby’s (2004) four-year longitudinal study (across grades two to five) measured morphological awareness through a Sentence Analogy Task, where the children heard a priming sentence and subsequently had to fill in the same morphological markers into the following test phrase. Reading comprehension in this study was measured through various subtests of the Woodcock Reading Mastery Tests-Revised, including word pronunciation, single word reading, pseudoword reading, and passages (Deacon & Kirby, 2004). Morphological awareness was found to rarely contribute to single-word reading, however, it did play a significant role in reading development (Deacon & Kirby, 2004). More specifically, in their hierarchical regression analysis, morphological awareness continued to show an effect on reading comprehension, when phonological awareness and intelligence – both verbal and nonverbal –   5 were controlled for. The results remained true even three years after the initial assessments, suggesting that morphological awareness is a good predictor of different aspects of reading development well into a child’s elementary education (Deacon & Kirby, 2004). Their results were consistent with findings from Carlisle’s (1995) study, where it was argued that morphological awareness makes the greatest impact on reading comprehension (Deacon & Kirby, 2004). Berninger, Abbott, Nagy & Carlisle (2010) examined phonological awareness, morphological awareness, orthographic knowledge and oral vocabulary in early elementary students across grades three to six. The Comes From and Signals tasks, as well as tasks that probe decomposition and derivation were used to assess morphological awareness. Within this task, participants determined if a complex word was derived from the original base word provided, and chose the appropriate derived (derivationally or inflectionally) word, and decompose complex words into their base forms. Finally, participants filled in a blank space within the test phrases with the words they productively derived from a base provided. Phonological knowledge was probed through syllable deletion, phoneme deletion, rime deletion, and the nonword memory subtest of the Comprehensive Test of Phonological Processing – 2 (CTOPP-2). Participants’ orthographic knowledge was assessed through receptive and expressive coding activities in which the students listened to both real and pseudowords, and either wrote out or chose the word match on a screen their letter order. Finally, oral vocabulary was assessed through the students’ ability to provide the definition for a particular word. Although reading comprehension past the word level was not assessed, the results of the study displayed important findings related to the development of morphological awareness in developing readers. The outcomes of the study showed the steepest growth in morphological awareness between grades one and three, with evidence of   6 continued growth past grade four (Berninger et al., 2010). The authors argued that all types of metalinguistic awareness must work together in reading development, and phonological awareness is not the most significant predictor as previously thought (Berninger et al. 2010). These findings are significant in that they highlight how important morphological awareness is in learning to read English, particularly in the early elementary grades.  1.3.2 Monolingual Speakers and Readers of Other Languages  Reading skills and morphological awareness of monolingual students whose L1s are a language other than English have also been investigated, although the research is limited. The majority of the research in the area of metalinguistic awareness has typically focused on monolingual’s knowledge of language and its relationship to reading and literacy.  Across languages, inflectional morphological knowledge has been found to be acquired before derivational morphology and compound morphology in English, French, Chinese, Russian, and Turkish (Kuo & Anderson, 2010). Inflectional morphology is a means of constructing words to fulfill their syntactic function (Haspelmath & Sims, 2010), for example adding third person singular suffix “–s” to the verb “walk” to indicate person and tense. Derivational morphology on the other hand often changes or adds to the meaning of a word (Haspelmath & Sims, 2010), such as adding the suffix –ness to “happy” to derive the word “happiness”. Morphological knowledge appears to continue to develop throughout the school years, and is closely linked to other forms of metalinguistic awareness in other languages (e.g., phonological and syntactic), and has been found to predict future reading abilities in children’s L1s such as French, Russian, Chinese, and Turkish in cross-linguistic research (Kuo & Anderson, 2010).    7 An investigation of French monolinguals examined students from Kindergarten to grade four on measures of morphological awareness, phonological awareness, and reading comprehension at the sentence level (Casalis & Luis-Alexandre, 2000). Reading comprehension was measured through a French reading task that involves reading a sentence and choosing, out of four options, which image best represents the phrase. Morphological and phonological awareness were examined through various subtests of experimental measures. Morphological knowledge was reflected in tasks where the participants’ ability to derive meaning from complex words and select an image that best represented that word, complete a sentence with both real and pseudowords, orally segment complex words into their morphological constituents, and produce various inflected and derived forms of different verb tenses and nouns, were tested. Tasks involving rhyme choice, phoneme and syllable deletion probed phonological awareness.  The results indicate that there appears to be a strong link between phonological and morphological awareness, as well as between morphological awareness and reading abilities (Casalis & Luis-Alexandre, 2000). French morphological awareness scores were correlated with decoding and reading comprehension in the second grade (Casalis & Luis-Alexandre, 2000). Further, in a review of sixteen studies that have investigated morphology and literacy, it was consistently found that morphological awareness likely contributes to reading development through knowledge of morphemic structure, spelling, and through the meaning of written words across languages: English, Chinese, Danish, Dutch, French, and Norwegian (Carlisle, McBride-Chang, Nagy & Nunes, 2010).  The research that has focused on children who are native speakers of languages including Chinese, Mandarin, and Korean has found similar results to the existing   8 monolingual research regarding morphological awareness. Chinese syllable awareness was found to directly predict character writing and recognition, but it did not predict higher level literacy skills such as reading comprehension or reading fluency (Pan, Song, Su, McBride, Liu, Zhang, Li & Shu, 2016). Consistent with research conducted in English, morphological awareness appears to be related to literacy skills beyond the word level, as measured by tasks involving creating new words via compounding (e.g., “a tree that grows shoes” is called a “shoetree”) and producing new compounded variants of a word from the target character within the word. Additional research in East Asian languages suggests the relationship between morphological awareness and literacy in Chinese is highly influenced by vocabulary, as well as morphological awareness being a greater predictor of vocabulary knowledge compared to phonological awareness (Tong, Tong & McBride, 2016; McBride-Chang, Tardif, Cho, Shu, Fletcher, Stokes, Wong & Leung, 2008). These results indicate that morphology plays a distinct role in the development of literacy skills. Overall, the research regarding East Asian languages suggests that morphological awareness is an effective predictor of vocabulary in those languages.  1.3.3 Bilingual Speakers and Readers Regarding these studies that did consider multilingual learners, there is evidence to support a transfer of morphological knowledge across languages between English and French, Chinese and English, as well as Korean and English.  Children in elementary school who are learning to read in a language different from their L1 may be affected by the differences in morphology between their languages, as they are constantly having to decode words and meanings in distinctive ways, switching between speaking or reading in their L1 and subsequent languages throughout the day. As the   9 acquisition of biliteracy may pose additional cognitive demands to a developing reader – due to differences in the process of phonology, morphology, syntax, and orthography – it is important to gain an understanding of students’ knowledge of these processes to explore not just the differences, but commonalities as well.  In a study of Chinese-English primary school students, phonological awareness, morphological awareness, oral vocabulary, word reading, and reading comprehension in both languages were investigated. The students’ knowledge of English (L2) compounding morphology contributed to Chinese (L1) character reading and reading comprehension (Wang, Cheng & Chen, 2006). The results suggest that morphological knowledge in a L2 may support a transfer of knowledge in the metalinguistic awareness of a students’ L1, despite different orthographic representation of words and sounds (Wang, Cheng & Chen, 2006).  Deacon, Wade-Wooley and Kirby (2007) studied the morphological knowledge in two languages of native English speaking students who were enrolled in French immersion. In this study, English morphological awareness was probed through a sentence analogy test, in which the students made new forms of a sentence involving manipulations of tense. When given the example phrases “Tom held the puppy” and “Tom holds the puppy”, the students would provide the same change to the test item “Tom held the fish”. French morphological awareness was measured via a similar tense manipulation analogy task. Significant relationships between early (grade one) measures of English morphological awareness were found with both English and French reading. French morphological awareness, however, was only significantly related to French reading. Interestingly, at a later time (grade three), measures of French morphological awareness were significantly related to both English and French reading. This research highlights that morphological awareness in a L2 appears to   10 grow with experience with print in two languages as well as with general knowledge of the L2.  Investigation into the cross-language transfer of morphological awareness, research suggests that this metalinguistic ability manifests similarly in a person’s subsequent languages. In a study of the cross-linguistic transfer of Chinese and English morphological awareness, Yeh, Joshi and Ji (2015) found that students in grades seven and nine displayed similar levels of morphological awareness in both languages. Consistently, it has been found that, between Chinese and English, as well as between Korean and English, there were correlations between morphological awareness in one language and reading development in the other language, despite different orthographic and morphological systems (Wang, Cheng & Chen, 2006; Pasquarella, Chen, Lam, Luo & Ramirez, 2011; Ke & Xiao, 2015; Cho, Chiu & McBride-Chang, 2011). Moreover, morphological awareness, reading, and spelling in a second language have also been suggested to influenced by the morphological structure of the language used at home (Marinova-Todd, Siegel & Mazabel, 2013). The authors examined 244 grade six ELL students from Canada and compared them with 888 English monolingual speakers on measures of word reading, reading comprehension, spelling, and different levels of linguistic awareness. Interestingly, it was found that transparency of morphological processes – how easily the meaning of a complex word can be derived from its’ structure (Saiegh-Haddad & Geva, 2008) –  in the home language also appear to influence the degree to which the ELL student performs on an English morphological awareness task (Marinova-Todd, Siegel & Mazabel, 2013). English morphological awareness was also found to independently contribute to English reading and spelling skills above phonological awareness (Marinova-Todd, Siegel & Mazabel, 2013). Of the ELL groups, the students with Slavic   11 backgrounds tended to score the highest on English metalinguistic awareness measures compared to the other L1 backgrounds such as Korean, Persian, and Romance languages (Marinova-Todd, Siegel & Mazabel, 2013). More specifically, on the morphological awareness measure that used pseudowords to probe participants’ knowledge, the participants with Persian and Korean backgrounds scored the lowest (Marinova-Todd, Siegel & Mazabel, 2013). With the morphological task that used real words, the participants with Korean language backgrounds scored significantly lower than all other ELL groups (Marinova-Todd, Siegel & Mazabel, 2013). Overall, this research highlights the individual differences between languages and how morphological knowledge is different depending on one’s linguistic experiences.   1.4 Summary of Previous Metalinguistic Research Past research in the area of morphological awareness and reading comprehension has typically focused on monolinguals, with limited investigation into native speakers of English compared to those who are ELLs. So far, the existing work on English speaking students suggests that morphological awareness is linked to vocabulary and reading comprehension at the sentence level, and above. Across languages, morphological awareness appears to develop into the middle elementary school years, with the steepest growth within the early elementary years as children are introduced to literacy. So far, it is also suggested that morphological awareness plays a large part in reading development, particularly in literacy activities beyond the single-word stage. With regards to bilingual studies, there is an observed transfer of morphological knowledge across languages that also influences reading comprehension in one’s subsequent languages. There is also evidence suggesting that some languages may have a greater supporting role for morphological knowledge compared to   12 others depending on the learner’s L1. Because exposure to more than one language is common in developing readers, investigating the commonalities and differences in the processing of morphology is needed. Finally, it is interesting for morphological knowledge to be explored in relation to reading comprehension from a distributed connectionist perspective because it is a skill that requires the integration of multiple levels of linguistic skills, and is consistent with the interactive model of reading (Vacca, Vacca, Gove, Burkey, Lenhart & McKeon, 2009) 1.5 Morphological Representation  The concept of morphological representation has been proposed in previous studies examining the priming effects of words with different morphological conditions (e.g., different degrees of semantic relatedness between targets and primes) in the investigation of morphological processing (Taft & Nguyen-Hoan, 2010; Hong, Dronjic, Chen, Li, Cheng & Wu, 2017). Typically, the few papers that have discussed morphological representation aimed to gain a greater understanding of mental lexical organization through priming studies. Researchers have taken one of two approaches to morphological representation in the mental lexicon: (1) claiming that morphologically complex words are represented separately from their bases and are not decomposed into their individual morphological parts (Henderson, Wallis & Knight, 1984; Manelis & Tharp, 1977), or (2) that there are base representations (first accessed through word processing) that are linked to their morphologically more complex variants through the decomposition of the morphological constituents (Frost, Deutsch & Forster, 1997). Researchers who take the distributed connectionist approach to morphological processing proposed that there is not a specific level of accessing morphology, but that there is a representation of the relationships between semantics,   13 phonology, and orthography (Frost, Deutsch & Forster, 1997; Schiff, Raveh & Fighel, 2011). The current study is not directly looking to organize the English mental lexicon of those who are ELLs, although it will utilize a task which is based on the notion that morphological representations are distributed across the different domains of language, namely the Morphological Awareness Semantic Task (MAST), to investigate morphological representations between those who are ELLs and those who have learned English as their first language.  Existing literature that has taken the distributed connectionist approach to morphological processing is limited and has typically been used in lexical priming studies with adults. Taft & Nguyen-Hoan (2010) examined the priming of semantically and morphologically related words through a task involving three morphological conditions: high/moderate semantically related words, low semantically related words, and form only related words, to test how semantics and morphology interact with regards to how much priming occurs with the varying conditions. Because the methodology of the task was designed to measure priming effects in a lexical decision task, the relevant findings (regarding morphological processing from a distributed connectionist point of view) will be described. In the high/moderate semantically related condition it was argued that the word “hunt” primes the word “hunter” through the intermediate level of linguistic processing. This is because “hunter” is a morphological variant of “hunt”, and is also semantically related. An example of the low semantically related condition, would be the two words “corn” and “corner”. The word “corn” only somewhat primes the word “corner”, because they are semantically unrelated, however, “-er” is a real-world morphological marker. Finally, in the form only condition, the word “turn” does not prime the word “turnip”, as the two are   14 semantically unrelated and because “-ip” is not a morphological marker that carries any meaning. It appears that the priming effects of the interaction of morphology and semantics accounts for the lemma level of processing: the link between the form and the function of a word. Taft & Nguyen-Hoan (2010) compared the bias that is created by morphemically related primes to semantically related primes (both masked) in an investigation of “lemma” level representations. For example, three conditions of word primes and targets were compared: morphemic (e.g., prime: sticky, target: stick), semantic (e.g., prime: glue, target: stick), and unrelated. (e.g., prime: phrase, target: stick). The lemma level is suggested to mediate form and function of morphological representations. Taft & Nguyen-Hoan’s (2010) results revealed that the meaning of an ambiguous word can become more prominent “by virtue of the prior masked presentation of a word derived from that meaning” (pg. 287), a change not observed when the mask and target word are only related semantically. This phenomenon supports lemma level access as there is no priming when form on its own serves as a prime (as with homographs), nor with solely semantic relationships between the ambiguous words (Taft & Nguyen-Hoan, 2010). Investigation into morphological representations promotes the importance of the connections of multiple levels of linguistic information involved in reading comprehension, as it highlights the relationship of semantics and morphology as graded. Relevant to the present study, examining morphological representations in relation to reading comprehension should consider a distributed connectionist view as it is consistent with the Interactive Model of reading, accounting for the simultaneous processing of multiple levels of linguistic information.  In further support, Hong, Dronjic, Chen, Li, Cheng & Wu (2017) adopted the view of morphological processing from a distributed connectionist perspective of morphological   15 knowledge in relation to literacy. In their study, the authors investigated Chinese morphological awareness and the role of meaning, orthography, and phonology on reading comprehension in primary school children. The participants in the study completed linguistic tasks (e.g., phonological awareness, character naming, vocabulary) as well as a morphological judgement task, which required the participants to determine whether the first morphemes of two words presented orally had the same meaning (e.g., “yè wǎn” meaning ‘evening’ compared to “yè jiān”, meaning ‘night’). Interestingly, it was found that children’s performance on a morphological judgement task was a significant unique predictor of their reading comprehension (Hong, Dronjic, Chen, Li, Cheng & Wu, 2017). Morphological knowledge, phonological knowledge, syntactic knowledge semantics, and orthography have an impact on reading comprehension because morphology not only follows phonological rules, but carries meaning and has a syntactic function within a phrase.  Other researchers who have approached morphological knowledge from the distributed connectionist point of view have investigated the relationship between reading abilities and morphological representations of grade four and seven Hebrew speaking students. Schiff, Raveh & Fighel (2011) found greater reading skills in the grade six sample to be related to more abstract and complex morphological representations, as measured through a priming task that included complex Hebrew words that were related either morphologically (M+), semantically (S+), neither (M–S–), or both (M+S+). Such abstract morphological representations appear to be less dependent on semantics compared to less abstract representations, which are more dependent on the semantic relationship of a word’s root (Schiff, Raveh & Fighel, 2011). Further, the authors found the grade four students, less experienced readers, tended to rely more heavily on semantics, while the grade seven   16 students tended to decode words based on the morphological structure of words (Schiff, Raveh & Fighel, 2011). These results suggest that readers initially rely more on semantics and their vocabulary knowledge, and with more experience, begin to rely on morphological knowledge and word structure. 1.6 Morphological Awareness Semantic Task The Morphological Awareness Semantic Task (MAST) used in this study will be used to probe the participants’ morphological representations: how they connect the form (phonology) and meaning (semantics) of printed words. Consistent with the Interactive Model of reading, the MAST has been developed to take a distributed connectionist view of morphological awareness, an appropriate task for assessing metalinguistic knowledge in relation to reading comprehension. A similar task was used to research the priming effects of the interactions of semantic and phonological relatedness between words in adults (Gonnerman, Seidenberg & Anderson, 2007). The authors used a cross-modal lexical decision task reflecting five different morphological and semantic conditions (form only: “coffee-fee”, low semantic: “rehearse-hearse”, moderate semantic: “midstream-stream”, high semantic: “preheat-heat”, and semantic only: “destiny-fate”). Within this task, used to promote semantic level processing, the participants made decisions regarding word similarities after having been first exposed to a prime (auditory) followed by the target word (visual) (Gonnerman, Seidenberg & Anderson, 2007). The task was found to elicit graded effects of priming, predicted by semantic and phonological overlap. The results of the study support a graded relationship between words, rather than representing them as discrete units, consistent with the distributed connectionist approach (Gonnerman, Seidenberg & Andersen, 2007). Finally, the authors describe how “morphological structures are not discrete units but   17 rather reflect degrees of semantic and phonological similarity across words” (pg. 338) (Gonnerman, Seidenberg & Andersen, 2007).  This task has only been administered to children in one study so far, in the context the development of morphological representations in children across grades three to five. Quémart, Gonnerman, Downing & Deacon (2017) used a cross-modal priming activity to probe morphological knowledge, comparing the localist and distributed connectionist viewpoints. Using the same methodology as Gonnerman et al. (2007), the degree of similarity across form and meaning of primes and targets were organized into the following category of relationships: form only (spin-spinach), low semantic (belly-bell), moderate semantic (late-lately), high semantic (bold-boldly), and semantic only (garbage-trash). From the localist account, the authors predicted that if “morphological priming effects result from the activation of a specific level of morphological representation, priming effects should be an all-or-none phenomenon” (pg. 3). If priming appears to result from “high form–meaning consistency across morphologically related words, then we predict graded priming effects as a function of the degree of convergence between form and meaning” (pg. 3) (Quémart et al., 2017). The primes were presented auditorally to the children through headphones, with the target item then presented on a computer screen after which the participants had to decide if the word was real or a pseudoword.  The results of the study revealed that the degree of priming between prime and target was dependent on their relationship of form and meaning, evident with faster response times indicative of  greater priming (Quémart et al., 2017). Effect of priming, measured in milliseconds, tended to be larger in the high semantic condition compared to the moderate condition, with the latter greater than the low semantic condition. Between the low semantic and form only conditions, the effect of priming was not   18 statistically significantly different, as these conditions shared a similar effects of priming. Finally, the high semantic condition displayed a larger priming effect compared to the semantic only condition. Overall, the results were consistent with previous priming studies conducted with adults in Gonnerman et al.’s (2007) study. This study highlights the distributed connectionist account of morphological processing and the co-activation of the multiple levels of information required in reading, over the localist perspective. Both form and meaning appear to have an influence on the processing of words.  The current study will not be examining the priming effects of different morphological structures, but will be considering the effect of morphological representation on reading comprehension through a similar task used in priming studies. It is important to consider mapping orthographic representation of morphological structures to form and meaning, particularly when relating it to reading comprehension. Children who are learning to read must not only process the orthographic representation and its phonological form, but extract its true meaning. My aim is to determine if such representation of morphology contributes to reading comprehension, and how that may be different or similar in those who are learning to read in their second language.  1.7 Research Questions  This study will investigate the following research questions:  (1) How does the English morphological representation of ELL participants compare to those who speak English as a first language?  (2) How does English morphological representation relate to English reading comprehension in those who are ELL compared to EL1 control?    19 To further investigate different levels of linguistic awareness, this study will also consider how the ELL participants’ other English metalinguistic awareness skills relate to their English reading comprehension compared to EL1 control. In light of previous research, it is hypothesized that the two groups will perform similarly on measures of morphological representation and reading comprehension – reflecting a correlational relationship in both groups.    20 Chapter 2: Methods  2.1 Participants The participants in the current study were recruited from elementary schools in the Toronto area as part of a larger longitudinal study on bilingualism. The sample for the present study were children in grade 2.  There were a total of 43 participants in the study, 26 were native speakers of English (EL1), while 17 were native speakers of various languages (ELLs) other than English, in addition to receiving instruction in English at school. The participants’ language backgrounds are described in Table 2.1 below. Language groups were defined through the parent report of the child’s first language. The EL1 group was defined by the parent reporting English as the child’s L1. The ELL group was defined by the parent reporting a language other than English as the child’s L1.    Group    L1  Frequency  EL1 (n=26) English  26  ELL (n=17)  Arabic  1  Bangla 1 Bengali 3 Cantonese 3 Chinese 1 Hindi 1 Mandingo 1 Tamil 4 Tibetan 2   Table 2.1 Participants’ First Language.    21 Background Questionnaire. Each participant’s parent (either Mother or Father, assigned as Guardian 1) completed a background questionnaire which was used to gather information regarding general demographics, L1, guardian’s L1, fluency, information about the children’s siblings, and presence/absence of other medical diagnoses related to language and learning.  Participants who had a formal diagnosis of either Autism Spectrum Disorder, speech or language disorder, or learning disability were not included in the sample.   Group Age in Months Gender Age child began speaking English in months  Total  93.89 (87.20-100.00)  Male (n=22) Female (n=21)  N/A EL1 94.22 (87.30-100.00) Male (n=13) Female (n=13) 13.73 (8-24)  ELL 93.45 (87.20-99.40) Male (n=9) Female (n=8)  35.57 (12-60)  Table 2.2 Participant Demographics.  Group Parent responding Parent’s fluency in English Frequency speaks English with child Frequency speaks other language with child      EL1 Mother n=19 Father n=1  Missing n=0   3.67  3.75  1.58  ELL Mother n=12 Father n=2 Missing n=5   2.41  2.5  3.35 Table 2.3 Guardian 1 Demographics.    22 Group Maternal Level of Education Number of English books in home How often child reads books in English  Other language books in home      EL1 2.35  3.16  1.92  0.45  ELL 1.79  2.0 1.78  1.29 Table 2.4 Other Relevant Demographics.   The participants in this study had an overall mean age of 93.89 months (87.20-100.00), with the EL1 group having a mean of 94.22 months (87.30-100.00). There were 13 males and 13 females in the EL1 group. The mean age the EL1 group participants began speaking English was 13.73 months (8.00-24.00). The ELL group had a mean age of 93.55 months (87.20-99.40), consisting of 9 males and 8 females, and were reported to have begun speaking English at 35.57 months (12.00-60.00).  The Guardian 1 (either the Mother or the Father of each child) reported their fluency in English (0=not fluent, 1=limited fluency, 2=somewhat fluent, 3=quite fluent, 4=very fluent), how often they speak English with their child (0=never, 1=seldom, 2=50%, 3=usually, 4=almost always), as well how often they speak another language with their child (0=never, 1=seldom, 2=50%, 3=usually, 4=almost always). The Guardian 1s for the EL1 group rated themselves as “quite-very fluent” in English, speaking English to their child “usually-almost always”, and speaking to their child in another language “seldom-50%”. For the ELL group, their Guardian reported to be “somewhat-quite fluent” in English, on average. They also reported speaking English with their child “50% of the time-usually” and speaking another language with their child “usually-almost always”.  Other relevant demographics that were compared were maternal level of education (0=Primary, 1=Secondary, 2=College, 3=University degree, 4=University-Master’s,   23 5=University-PhD), number of English books in the house (0=fewer than 10, 1=10-25, 2=25-50, 3=50-100, 5=more than 100), how often the child reads books in English (0=never/almost never, 1=at least once a week, 2=every day), as well as how many books in other languages there were in the home (0=fewer than 10, 1=10-25, 2=25-50, 3=50-100, 5=more than 100).   For the EL1 group, the reported mean level maternal education was a College or University degree. Respondents also stated that they had between 50 and 100 English books in the home, the children read English books almost every day, and had fewer than 10 books of other languages at home. For the ELL group, the reported mean level of maternal education was a Secondary or College degree, they had between 1 and 25 English books in the home, the children read English books almost every day, and had between 10 and 50 books of other languages at home.  2.2 Procedure 2.2.1 Measures  In addition to the background questionnaire mentioned above, the following measures were administered to each participant:  Morphological Representation. The Morphological Awareness Semantic Task is an experimental measure that aims to access morphological knowledge based on how the participants rated morphological constructions and their semantic resemblance in a sentence. It was adapted from Carlisle’s (1995) kindergarten morphological awareness task to examine how children in early elementary grades relate phonology and semantics, as well as develop morphological representation. The participants were required to rate sentences on a Likert scale of 1-5 to quantify whether a sentence sounded “Silly” (1), “Sort of Silly” (2), “Not Sure” (3), “Sort of Makes Sense” (4), or “Makes Sense” (5). There were three practice items   24 and a verbal reminder from the research assistants, to orient the participants to the meaning of the faces. The task consisted of 40 total items: Forms A and B each had 20 items, and were administered at two separate time points within two weeks of each other. The examiner would instruct the participants to “point to a face for this sentence” and subsequently read aloud each sentence to the child and. Finally, the participants would point to whichever face they thought corresponded to the sentence. There were five morphological conditions within this task. The conditions and their correct responses are described in table 2.5 below:   Morphological Condition Example Number of items across forms A and B Correct Response Form Only “When you turn you are a turnip”  9 Silly  Low Semantic “A person who will wait is a waiter”  8 Silly  Moderate Semantic “If you are direct, then you walk directly”  8 Sort of makes sense/ Sort of silly High Semantic “If you sing, you are a singer”   7 Makes sense Semantic Only “If you fight, you argue” 8 Makes sense Table 2.5 MAST Conditions.  The five morphological conditions and the criteria used by the task developers at the University of Toronto to assign each sentence to a category based on the formal and semantic relationships between word pairs are described in Table 2.6 below. A sentence was classified in the form only condition when the word pairs within the sentence did not have any semantic relationship and when the added orthographic information (e.g., “-ip”) to the original base word (e.g., “turn”), was not a morphological maker. A sentence was considered to fit in the low semantic condition if there was a meaningful morphological marker added to the base   25 word, however, the morphological marker (e.g., “-er”) did not make sense within the semantics of the sentence (e.g., “wait” and “waiter”). A sentence was assigned to the moderate semantic condition if the two words were semantically related, had a viable morphological marker that carried meaning and was semantically related to the base word (e.g., between “directly” and “direct”), however the semantics within the sentence did not make sense. Sentences were considered high semantic when the base and morphologically derived word were semantically related, the morphological marker carried meaning, and made sense within the whole phrase. Sentences were assigned to the semantic only condition if the word pairs within the sentence were only related by meaning and not orthographically or based on spoken form.    Formal relationship between word pairs Meaningful morphological marker Semantic relationship between word pairs Semantic congruency throughout sentence Form Only *    Low Semantic * *   Moderate Semantic * * *  High Semantic * * * * Semantic Only   * * Table 2.6 MAST Conditions description.   Phonological Awareness. The Comprehensive Test of Phonological Processing – 2 (CTOPP-2; Wagner, Torgesen, Rashotte & Pearson, 2013) consists of several subtests that tap phonological awareness and verbal memory. To probe phonological awareness in this study, the participants were administered the Elision subtest. The task measures the ability to   26 delete parts of words which result in new, existing words. For example, the person administering the subtest instructs the participant to say “toothbrush”, and then to say “toothbrush” without saying “tooth”, where the correct response would be “brush”. Subsequent trials involve phonemes where participants are instructed to say “cup” without saying /k/, or say “tan” without saying /t/.  Vocabulary. The Peabody Picture Vocabulary Test (PPVT-IV; Dunn & Dunn, 2007) was administered to assess the participants’ vocabulary and semantic knowledge. The tester asks the children to point to different words, presented orally, in a flip book. The children must point to the picture that corresponds with the word provided, out of a choice of three or four images.  Word Reading. To probe the participants’ English word reading abilities, the Letter-Word identification subtest of the Woodcock-Johnson III Edition (Woodcock, McGrew & Mather, 2001a) was administered. This task assesses the participants’ oral reading skills through their pronunciation accuracy as they read aloud words that increase in difficulty. Test administrators pointed to a word in the test booklet and ask “What word is this?”, to which the participants read each word aloud. Researchers were allowed to prompt the participants to “first read the word silently and then say the word smoothly” once if participants read dysfluently: either letter by letter or pronouncing each syllable separately. The participants in the current study began at item 17 (suggested starting point for grade two), with the test consisting of 76 items. Fluent pronunciations were coded with a “1” and incorrect or dysfluent pronunciations were coded with a “0”. Students were not penalized for misarticulating the words or pronouncing them with dialectical differences.      27 Reading Comprehension. Finally, the Gates-MacGintie Reading Test fourth edition (GMRT; MacGinitie, MacGinitie, Maria, Dreyer & Hughes, 2000), Level 1 Form S was used to measure reading comprehension. The GMRT is a reading comprehension test that was administered to the children in a group setting, within a 35-minute time slot with two to three research assistants present. For each of the 11 trial stories, the children were be presented with a series of four sets of phrases with three images beside each, with one image being the correct one corresponding to the phrases. The child must select the picture that corresponds the best with the sentence presented. For example, in a series about a girl named “Ruth”, the phrases describe how much she loves cars: playing with her toy cars, helping her father clean his car, reading stories about cars with her father, and how she wants to fix cars as a job when she grows up. In this story, the first set of phrases reads: “Ruth likes cars. She has lots of toy cars in her room”. The three images depict (1) a bedroom with car posters on the walls with toy cars on the floor, (2) a bedroom with posters of a house and a mountain with dolls and books on the floor, and (3) Ruth and her father driving in a car. The correct image for the participant to pick would be the image of a bedroom with car posters and toys.  2.2.2 Data Analysis   For the data analysis of this study, a series of t-tests and One-Way Analysis of Variances (ANOVAs) were conducted to compare groups on various measures, as well as a repeated measures ANOVA to assess the different conditions on the MAST. Pearson Correlations were used to measure the degree of relationships between variables.       28 Chapter 3: Results Table 3.1 below displays the means and standard deviations of the participants’ results for the tasks, based on language group. For the data analysis, standard scores and percent correct were used.  Measure  EL1 Mean (SD) ELL Mean (SD) Morphological Representation (%) 53.37 (9.92) 50.00 (13.58) Phonological Awareness (SS) 10.88 (2.42) 10.12 (2.80) Word Reading (SS) 112.48 (12.34) 113.29 (9.59) Vocabulary (SS) Reading Comprehension (%) **110.35 (10.95) 45.92 (29.25) **99.18 (11.07) 41.69 (24.98)  Note. Standardized means (standard deviations) for each standardized test are described; CTOPP-2 m= 10 (3), WJ-III m=100 (15), PPVT-IV m=100 (15), GMRT-IV Level 1 m=32.5 (7.0). Table 3.1 Standard Scores (SS) and Percent Correct (%) Results. **Significant differences at p=.01 level.   To account for the multiple t-tests being performed simultaneously on five measures, the Bonferroni correction was employed to adjust the critical p-value (a = .01), reducing the likelihood of making a Type I error. Because the Levene’s Test for Equality of Variance was not significant equal variances were assumed. There were no significant differences between groups on reading comprehension (t(36)=.939, p=.371), phonological awareness (t(40)=.940, p=.353), word reading (t(40)=-.229, p=.820), or on the morphological representation measure (t(41)=.939, p=.353). There was, however, a significant difference between the group means on the vocabulary measure (t(41)=3.257, p=.002) whereby the EL1 group (m=110.35) scored higher than the ELL group (m=99.18). it is important to note that the ELL group was within the average range. Results from the t-test are presented in Table 3.2 below. It is important to   29 note that on standardized measures, both groups performed near or above standardized means.   Measure   Significance (2-tailed) Mean Difference Morphological Representation .353 3.37 Phonological Awareness .353 .76 Word Reading .820 -.81 Vocabulary **.002 11.17 Reading Comprehension .657 4.23 Table 3.2 Independent Samples t-Tests. **Significant values at p=.01 level.    Pearson Correlation were used to determine the relationships between the different variables, presented in Table 3.2, and Figures 3.1 and 3.2 below. The scores above the diagonal line are for the ELL group and the scores below the line are for the EL1 group.  A medium size significant correlation between phonological awareness and word reading (r=.645, n=17, p=.005) was observed for the ELL group. For the EL1 group, the only significantly correlated measures were word reading and reading comprehension (r=.625, n=24, p=.001), which displayed a large correlation.           30 Task 1 2 3 4 5 1. Morphological Representation - .160(.539 -341(.181) .304(.236) -.242(.425) 2. Phonological Awareness .101(.631) - **.645(.005) -.333(.192) -.121(.694) 3. Word Reading -.032(.879) .494(.012) - -.363(.153) .024(.938) 4. Vocabulary -.100(.625) .165(.432) .109(.603) - .198(.518) 5. Reading Comprehension .143(.495) .411(.046) **.625(.001) .445(.026) - Table 3.3 Correlations and 2-tailed significance levels. **Significant at the 0.01 level. Scores above the diagonal line represent the ELL group and the scores below the diagonal line represent the EL1 group.        Figure 3.1 ELL scatterplot. X axis represents word reading and Y axis represents phonological awareness.  ELL Phonological Awareness ELL Word Reading    31      Figure 3.2 EL1 scatterplot. X axis represents word reading and Y axis represents reading comprehension.    Figure 3.3 below describes the group totals for each task condition organized by language group. The mean scores represented in Figure 3.3 were calculated as percent correct for the number of questions for each morphological condition.   EL1 Reading Comprehension EL1 Word Reading   32  Figure 3.3 Percentage correct of MAST response trends organized by language group with standard error bars.   A One-Way ANOVA with language group as the independent variable and MAST conditions as the dependent variable revealed no statistically significant differences between groups on each of the MAST conditions: form only (F(1,41)=.770, p=.385), low semantic (F(1,41)=.676, p=.416), moderate semantic (F(1,41)=.954, p=.335), high semantic (F(1,41)=2.11, p=.153), and semantic only (F(1,41)=.165, p=.687). Visual examination of Figure 3.1, reveals that the participants in both groups performed highest on the high semantic condition (M=65.45%), the semantic only condition (M=60.66%), and the form only condition (M=59.17%). Both language groups performed lowest on the low semantic (M=45.41%) and the moderate semantic (M=30.49%) conditions. Although the manipulations across conditions lead to similar result and patterns for both groups, upon closer examination it becomes evident that the EL1 group had similarly high mean scores for the form only (m=62.39%), the semantic only (m=61.79%), and the high semantic conditions (m=60.99%), whereas the ELL group showed more differences, with the highest mean scores 0102030405060708090100Form Only Low Semantic ModerateSemanticHigh Semantic Semantic OnlyEL1ELL  33 for the high semantic condition (m=72.27%), the semantic only condition (m=58.93%), and the form only condition (m=54.25%).   A one-way repeated measures ANOVA was conducted to compare the effect of the degree of morphological and semantic overlap (conditions) on the students’ total performance on the MAST. The results showed that mean performance on the MAST differed significantly between conditions (F (2.77, 116.46) = 14.62, p<.001). Post hoc tests using the Bonferroni correction revealed that students performed significantly lower on low semantic (M= 45.41%) versus form only (M= 59.17%) (p<.000) and versus high semantic (M=65.45%) (p<.001) conditions. Further, students performed significantly lower on the moderate semantic condition (M=30.49%) relative to form only (M=59.17%) (p<.000), high semantic (M=65.45%) (p=.000), and semantic only (M=60.66%) (p<.000) conditions.     34 Chapter 4: Discussion  Previous research in the area of metalinguistic awareness and reading comprehension suggests that morphological awareness is connected to reading comprehension in English speaking students (Deacon & Kirby, 2004; Berninger et al. 2010) as well as in other monolingual populations (Casalis & Luis-Alexandre, 2000; Carlisle, McBride-Chang, Nagy & Nunes, 2010). Morphological knowledge continues to develop into the middle elementary school years (Berninger et al., 2010). Bilingual studies have found similar levels of morphological awareness between a child’s L1 and L2 (Wang, Cheng & Chen, 2006; Pasquarella, Chen, Lam, Luo & Ramirez, 2011, Ke & Xiao, 2015; Cho, Chiu & McBride-Chang, 2011). The concept of morphological representation is relatively new to the field of applied linguistics, and was first introduced in order to account for the graded nature between semantics and morphology, and their relationship to phonological and syntactical levels of reading comprehension. This study utilized an experimental morphological representation measure, the MAST, which has been developed to account for the various metalinguistic levels that contribute to reading comprehension from the distributed connectionist perspective. The aim of this study was to gain insight into the various levels of linguistic knowledge associated with reading comprehension in ELLs as compared to EL1s with a focus on morphological representations. The hypothesis that the EL1 and ELL groups would perform similarly on the measures of morphological representation and reading comprehension was confirmed. It was further confirmed that morphological representation and reading comprehension were correlated, but only for the EL1 group, not the ELL group. With regards to other measures of linguistic knowledge, the only significant difference   35 between language groups was on the vocabulary measure. The overall findings of the current study suggest that ELLs are able to achieve similar scores in reading comprehension to that of native English speakers, with minor differences in their morphological representations of English words. Despite having a smaller vocabulary size than native English speaking students, ELLs appear to be relying on different linguistic skills to decode and derive meaning.   4.1 English Morphological Representation of ELL participants compared to EL1s: When the English morphological representations of the ELLs and EL1s were compared, there were no significant differences between the groups on the MAST, neither for the overall score or for the various conditions. Consistent with previous research, it is possible that the equivalent performance of the ELLs was due to their ability to transfer knowledge across languages for this particular task (Cho, Chiu & McBride-Chang, 2011; Deacon, Wade-Wooley & Kirby, 2007; Ke & Xiao, 2015; Pasquarella, Chen, Lam, Luo & Ramirez, 2011; Wang, Cheng & Chen, 2006; Yeh, Joshi & Ji, 2015). However, since skills in the L1s of the ELL participants were not measured in the current study, a definite conclusion regarding transfer of knowledge could not be made. It is also possible that the participants in the ELL group had very strong skills in English, because they were exposed to it from an early age (about three years old). Moreover, previous research has found that, among other levels of metalinguistic awareness, there appears to be language transfer of morphological knowledge despite varying orthographic forms between L1 and L2 (Cho, Chiu & McBride-Chang, 2011; Deacon, Wade-Wooley & Kirby, 2007; Ke & Xiao, 2015; Pasquarella, Chen, Lam, Luo & Ramirez, 2011; Wang, Cheng & Chen, 2006; Yeh, Joshi & Ji, 2015). In the current study, the different typologies and orthographies of the ELL groups’ L1s did not   36 appear to negatively influence their overall performance on the MAST; not all of their L1s were alphabetic.  When the MAST conditions were considered separately, some group patterns were observed. Both groups scored the lowest on the moderate semantic and low semantic conditions, likely due to the task difficulty as well as the age and reading levels of the participants. Cumulatively, the two language groups seemed to be better at making judgements regarding the sentences when there is either an obvious semantic relationship (high semantic, semantic only) or no semantic relationship (form only) – so for the more obvious cases. The EL1 group was the most accurate on the form only condition (e.g., “When you turn, you are a turnip”), the semantic only condition (e.g., “If you fight you argue”), and the high semantic condition (e.g., “If you sing, you are a singer”). The ELL group was the most accurate on the high semantic condition compared to the EL1 group, although the difference was not statistically significant. It is likely that the EL1 group performed better on the form only and semantic only conditions compared to the ELL group, because of their larger receptive vocabulary size in English. It appeared that the children did not benefit from the morphological relationships between the words to help them with their meaning interpretation, but had to be familiar with the meaning of each word.  Evident in the significant findings when comparing the five different MAST conditions for the combined sample (EL1 and ELL groups), it appears that participants perform better on the English morphological representation task when there is a more obvious semantic relationship within the sentence items. Performance on the moderate semantic condition was statistically significantly lower than the form only, high semantic, and semantic only conditions. Similarly, for low semantic condition, participants performed   37 statistically significantly lower compared to the form only and high semantic conditions. It appears that when more analysis into the semantics of the target and prime as well as within the semantics of the whole sentence are required, making a relatedness judgement is more difficult.  On the other hand, the ELL group tended to perform somewhat better than the EL1 group on the high semantic condition, which could be explained with the degree of morphological markedness of some of the L1s represented in the sample. Although this difference was not statistically significant, a discrepancy of 11.28% (the condition with the largest discrepancy between language groups) may be accounted for by amount of experience with English, which although morphologically rich, is considered to be opaque or fusional (Marinova-Todd, Siegel & Mazabel, 2013). In a morphologically opaque language, an affix may represent more than one meaning, corresponding to derivational changes in the stem of the word (Saiegh-Haddad & Geva, 2008). While some of the ELL participants in the current study come from language backgrounds that are more analytic (minimally morphologically marked), such as Chinese/Cantonese (n=4) more of the ELLs (n=9) came from L1 backgrounds that are morphologically transparent, or agglutinative (e.g., Tamil and Bengali). Such languages typically assign one meaning to each individual affix, adding them to the root word without changing the root’s underlying structures, through connecting multiple affixes to generate different meanings (Haspelmath & Sims, 2010). Considering the varying language backgrounds and the different ways in which words are constructed may explain some of the group differences on the five MAST conditions, although definitive conclusions regarding the interaction of the ELL students’ L1 and English morphology may not be drawn due to limited sample size.    38 Morphological knowledge has typically been probed at the word level (Berninger et al., 2010, Hong et al., 2017) with morphological representations used as a concept in priming studies investigating this skill (Gonnerman, Seidenberg & Andersen, 2007; Taft & Nguyen-Hoan, 2010; Hong et al., 2017; Quémart, Gonnerman, Downing & Deacon, 2017). The morphological representation measure used in the current study went beyond the morphological knowledge of single words, to the sentence level, accounting for the multiple levels of metalinguistic awareness that are accessed through morphological representations.  To illustrate the differences in cognitive load between varying morphological awareness tasks compared to the MAST, a few of the tasks used in previous studies probing morphological knowledge will be discussed. One example of a word level morphological awareness task consists of making a “Yes” or “No” decision whether paired compound words, presented orally, have affixes with the same meaning for example, “yè wǎn” meaning ‘evening’ compared to “yè jiān”, meaning “night”, to which the answer would be “yes” (Hong et al., 2017). Other examples include tasks with carrier phrases where a participant may have to respond to a question such as “Does corn come from corner?” (Berninger et al.,2010). Although there are instances where the participants must choose a correct nonword form to fill in a simple sentence such as “The _____ really enjoy sharing his sport with others” out of the options “one-wibbled, wibbling, wibbler, wibbly” (Berninger et al., 2010), such tasks do not have the additional demands of relating two probe words to the meaning of the various novel sentences, such as with the MAST. For example, the MAST task item “A person who will drag is a dragon” requires comparing “drag” and “dragon”. To respond, one must first compare the phonological information and morphological relatedness between the prime and target word, and further compare the meaning of those words to what one knows   39 about how sentences are constructed as well as consider the semantics of the entire sentence. This supports the argument that this morphological representation task is likely more demanding than more usual morphological awareness tasks. The morphological representation measure used in this study provided novel and varied sentence structures for each pair of target words. The children were expected to use the meaning of the sentence rather than relying on word meaning alone to determine how “silly” a sentence sounded or whether it “made sense. Therefore, it is possible that the participants were too young and not experienced enough with reading in either language group, making the sentences within the MAST difficult to process. In particular, the low semantic and moderate semantic conditions which had word and sentence relationships that were less obvious than the other three conditions, and were statistically significantly lower. It appears that when conditions required more analysis and are not as easily judged as “silly” or “makes sense”, the children tended to have a harder time correctly judging the semantic and morphological relationships.  Exposure to and interactions with a language is likely a mediator of morphological representation knowledge. The novel and complex sentences, such as the ones used in the MAST, may not have directly reflected a child’s abilities in reading comprehension because they have not had enough exposure to various forms of the English language, either spoken or written. Considering the cognitive demands of morphological processing tasks, Ke and Xiao (2015) evaluated eight studies that assessed language and literacy transfer between English and Chinese of children from Kindergarten to grade six (with one study investigating grades seven to nine). They found that L2 language exposure, as well as task demands, impacted the transfer of linguistic knowledge contributing to bilingual literacy skills. The authors discuss “less language-specific facets” of morphology, which they describe as   40 understanding how to segment words into morphemes, and “more language-specific facets”, such as the different meanings of affixes and bases when understanding a word, with the latter requiring greater demands on one’s cognitive system (Ke & Xiao, 2015). The MAST incorporates both linguistic processes. In order to make an accurate judgement regarding the “silliness” of a particular item, one must be able to identify the morphemes in the target words of the test item, identify the meaning of the affix, and further, judge whether the semantic relationship of the affix fits with the sentence. This process of integrating “less language-specific facets” with “more language-specific facets” and sentence level meaning imposes a high level of cognitive demands.  4.2 English Morphological Representation in relation to English Reading Comprehension:  In addressing the second research question, the MAST was not found to be a significant predictor of reading comprehension for either ELLs or EL1s in the current study. As mentioned above, it is likely that interaction and experience with a language and its written form may mediate this relationship.  Regarding the other metalinguistic measures in relation to reading comprehension and bilingualism, the results of the current study provides further support of previous research. Phonological knowledge has been found to be associated with word level linguistic skills in the ELL group. Even though vocabulary and word reading are word level tasks that incorporate phonological awareness, knowledge of phonology on its own has not been found to predict reading comprehension or other linguistic skills beyond the word level (Deacon & Kirby, 2004; McBride-Chang et al., 2008; Pan et al., 2016; Tong, Tong & McBride, 2016). The finding that phonological awareness correlated with word reading for the ELL group,   41 while word reading was associated with reading comprehension in the EL1 group suggests that the ELL group are relying more on phonological knowledge, a skill that is essential for the decoding individual words. Further, the EL1 group appears to use skills beyond the phonological level for reading comprehension, as they appeared to rely more on word level skills, such as word reading and vocabulary. Even though there was a significant difference between the vocabulary levels of the two language groups, there were no statistically significant differences between morphological representation or reading comprehension. These results suggest that ELLs can integrate multiple levels of linguistic knowledge in a L2 as well as achieve similar English reading proficiency as native speakers, by relying on different skills.  The current study also supports previous research regarding bilingualism and vocabulary size through the significant differences observed between ELLs and EL1s on the English vocabulary measure (PPVT-IV). Previous research has established that monolingual English speaking children have larger receptive vocabularies than various groups of bilingual children as measured by the PPVT (Bialystok et al., 2010). Although there was a significant difference between the vocabulary knowledge of the two language groups, this did not appear to affect overall performance on measures of morphological representation or reading comprehension. Because the ELL group had a significantly lower vocabulary score than the EL1 group without statistically significant differences between groups on reading comprehension, these results indicate that ELLs may be using different linguistic skills than the EL1s to achieve proficiency in English literacy activities, possibly supported by their native language.    42 Previous research has also highlighted general language and literacy exposure in the home and its importance for the development of reading comprehension in English. Children who have had limited literacy experience prior to entering the school system are at a higher risk of reading difficulties (Kamhi & Catts, 2012). In the case of the current population, the EL1 group had more English books at home and a higher level of maternal education. On average, the EL1 and ELL groups were reported to read English books a similarly frequently, however, the ELL group had more books in another language. It is possible that these factors explain the EL1s advantage in English vocabulary. Regardless of differences in literacy activities at home, both groups performed similarly on reading comprehension in English. Therefore, experience with a language and its’ printed form in both school and home contexts are paramount in the development of understanding written language.   4.3 Limitations One shortcoming of the current study is the limited sample size for each group, and thus, the results of this study must be interpreted with caution. A larger sample could help to elucidate whether some non-significant trends would become significant with greater power. Further, the ELL group was quite varied. Because there was some missing information regarding the age at which the ELL group began speaking English as their second language, the amount of second language exposure was not available, thus a major limitation for the interpretation of the results. Moreover, the children in the ELL group came from varied language backgrounds, of which there were not enough participants to examine the relationship between L2 morphological representations compared to the morphological processes of the ELL participants’ L1. Although it is important to consider demographic representations of different language backgrounds, it may have been more helpful to have   43 varying ELL participant subgroups, comparing separate ELL groups organized by L1. Further, because it is common for most children to be exposed to multiple languages at an early age bilingualism is becoming the norm (Bernhardt, 2003), making controlling for various forms of language exposure difficult to account for.  Another limitation of the present study is the high task demands of the MAST, used to probe morphological representations, derivation in particular. Morphological knowledge has been shown to have its fastest growth in the primary grades, namely between grade one and three, and continues to develop in subsequent grades (Berninger et al., 2010). Knowledge regarding compound and derivational morphology tends to predict English reading comprehension, more so than measures of vocabulary and grammar (Zhang & Koda, 2013). Zhang and Koda (2013) argued that in the process of reading comprehension children utilize knowledge of affixes to infer word meanings, and parse apart compound words. Moreover, Berninger et al. (2010) found that derivational morphology continues to develop past fourth grade potentially due to its higher cognitive demands, meaning that our population of students in grade two may not have had enough experience with morphology to have developed complex morphological representations. Potentially, testing morphological representation with older elementary grades – grade four and beyond – may be more valid and sensitive to detecting their true knowledge of the complex nature of morphology.  4.4 Directions for Future Research  Future investigations of the role of morphological representation in relation to reading comprehension should focus on older populations, to account for increased language and literacy experience. Second language groups should be compared, assessing larger sample sizes of each L1 individually. Although it might be helpful to consider all languages   44 represented in a given school system cumulatively, as in this study, it is important to discover how different prevalent languages in that school system interact with English language and literacy development. Task demands are an important consideration for the area of linguistic research with children, particularly those who are ELLs. Finally, further investigation is needed regarding the experimental measure, MAST, to gain a better understanding of how children of different languages respond to the graded nature of semantics in relation to phonology, morphology, and syntax. 4.5 Significance of Findings  It is important to continue to gain an understanding of bilingual language and literacy development, particularly in the school years, due to the high prevalence of ELL students within Canadian school systems. Examining the different ways ELL children are acquiring metalinguistic and print knowledge in their language of academic instruction is important to inform how to further support their literacy development as well as provide insight into the complex cognitive processing involved in reading comprehension. Future implications for research in the area of metalinguistic awareness, reading comprehension, and cross-language transfer will provide insight into the linguistic and cognitive outcomes of multilingualism as well as provide options for assessment and monitoring for children in their language of instruction. Understanding how bilingual children process and decode novel structures, in both reading and spoken language, will help inform what would be the best approach to support them for academic success.     45 Bibliography  Bernhardt, E. (2003) Challenges to reading research from a multilingual world. Reading Research Quarterly, 38(1), 112-117. Berninger, V. W., Abbott, R. D., Nagy, W. & Carlisle, J. (2010) Growth in phonological, orthographic, and morphological awareness in grades 1 to 6. Journal of Psycholinguistic Research, 39, 141-163. DOI: 10.1007/s10936-009-9130-6  Bialystok, E., Luk, G., Peets, K. F. & Yang, S. (2010) Receptive vocabulary differences in monolingual and bilingual children. Bilingualism, 13(4), 525-531. DOI: 10.1017/S1366728909990423  British Columbia Ministry of Education (1997; 2004 Revised) Language Education Policy. Retrieved from https://www2.gov.bc.ca/gov/content/education-training/administration/legislation-policy/public-schools/language-education-policy British Columbia Ministry of Education (2018) Student Statistics. Retrieved from http://www.bced.gov.bc.ca/reporting/province.php Carlisle, J. F., McBride-Chang, C., Nagy, W. & Nunes, T. (2010) Effects of instruction in morphological awareness on literacy achievement: an integrative review. Reading Research Quarterly, 45(4), 464-487. DOI: 10.1598/RRQ.45.4.5  Casalis, S. D. & Luis-Alexandre, M-F. (2000) Morphological analysis, phonological analysis and learning to read French: A longitudinal study. Reading and Writing: An Interdisciplinary Journal, 12, 303-335.  Cho, J-R., Chiu, M. M. & McBride-Chang, C. (2011) Morphological awareness, phonological awareness, and literacy development in Korean and English: A 2-year   46 longitudinal study. Scientific Studies of Reading, 15(5), 383-408. DOI: 10.1080/10888438.2010.487143  Deacon, S. H. & Kirby, J. R. (2004) Morphological awareness: Just “more morphological”? The roles of morphological and phonological awareness in reading development. Applied Psycholinguistics, 25, 223-238. DOI: 10.1017.S0124716404001117 Deacon, S. H., Wade-Wooley, L. & Kirby, J. (2007) Crossover: The role of morphological awareness in French immersion children’s reading. Developmental Psychology, 43(3), 732-746. DOI: 10.1037/0012-1649.43.3.732  Dunn, L. M. & Dunn, D. M. (2007) Peabody Picture Vocabulary Test–Fourth Edition [Assessment Instrument]. Minneapolis, MN; Pearson Assessments. Frost, R., Detusch, A. & Forster, K. I. (1997) What can we learn from the morphology of Hebrew? A masked-priming investigation of morphological representation. Journal of Experimental Psychology, 23(4), 829-856. Gonnerman, L. M., Seidenberg, M. S. & Andersen, E. S. (2007) Graded semantic and phonological similarity effects in priming: Evidence for a distributed connectionist approach to morphology. Journal of Experimental Psychology: General, 136(2), 323-345. DOI: 10.1037/0096-3445.136.2.323  Haspelmath, M. & Sims, A. D. (2010) Understanding Morphology 2nd Edition. London, United Kingdom: Hodder Education.  Henderson, L., Wallis, J., & Knight, K. (1984). Morphemic structure and lexical access. In H. Bouma & D. Bouwhuis (Eds.), Attention and performance X: Control of language processes (pp. 211-226). Hillsdale, NJ: Erlbaum.    47 Hong, L. Dronjic, J., Chen, X., Li, Y., Cheng, Y. & Wu, X. (2017) Morphological awareness as a function of semantics, phonology, and orthography as a predictor of reading comprehension in Chinese. Journal of Child Language, 44(5), 1218-1247. DOI: 10.1017/S0305000916000477 Kamhi, A. G. & Catts, H. W. (2012) Language and Reading Disabilities 3rd edition. Upper Saddle River, NJ: Pearson.  Ke, S. & Xiao, F. (2015) Cross-linguistic transfer of morphological awareness between Chinese and English. Language Awareness, 24(4), 355-380. DOI: 10.1080/09658416.2015.1114624 Kuo, L-J. & Anderson, R. C. (2010) Morphological awareness and learning to read: A cross-language perspective. Educational Psychologist, 41(3), 161-180. DOI: 10.1207/s15326985ep4103_3 MacGinitie, W. H., MacGinitie, R. K., Maria, K., Dreyer, L. G. & Hughes, K. E. (2000) Gates-MacGinitie Reading Test, Fourth Edition [Assessment Instrument]. Boston, MA; Houghton Mifflin Harcourt.  Manelis, L. & Tharp, D. A. (1977) The processing of affixed words. Memory & Cognition, 5(6), 690-695. DOI: 10.3758/BF03197417  Marinova-Todd, S. H., Siegel, L. S. & Mazabel, S. (2013) The association between morphological awareness and literacy in English language learners from different language backgrounds. Topics in Language Disorders, 33(1), 93-107. DOI: 10.1079/TLD.0b013e318280f5d5 McBride-Chang, C., Tardiff, T., Cho, J-R., Shu, H., Fletcher, Stokes, S. F., Wong, A. & Leung, K. (2008) What’s in a word? Morphological awareness and vocabulary   48 knowledge in three languages. Applied Psycholinguistics, 29, 437-462. DOI: 10.1017/S014271640808020X McBride-Chang, C., Wagner, R. K., Muse, A., Chow, B. W-Y. & Shu, H. (2005) The role of morphological awareness in children’s vocabulary acquisition in English. Applied Psycholinguistics, 26, 415-435. DOI: 10.1017.S014271640505023X Pan, J., Song, S., Su, M., McBride, C., Liu, L., Zhang, Y., Li, H. & Shu, H. (2016) On the relationship between phonological awareness, morphological awareness and Chinese literacy skills: Evidence from an 8-year longitudinal study. Developmental Science, 19(6), 982-991. DOI: 10.1111/desc.12356 Pasquarella, A., Chen, X., Lam, K., Luo, Y. C. & Ramirez, G. (2011) Cross-language transfer of morphological awareness in Chinese-English bilinguals. Journal of Research in Reading, 34(1), 23-42. DOI: 10.1111/j.1467-9817.2010.01484.x Quémart, P., Gonnerman, L. M., Downing, J. & Deacon, H. (2017) The development of morphological representations in young readers: A cross-modal priming study. Developmental Science. DOI: 10.1111/desc.12607 Saiegh-Haddad, E. & Geva, E. (2008) Morphological awareness, phonological awareness, and reading in English-Arabic bilingual children. Reading and Writing, 21, 481-504. DOI: 10.1007/s11145-007-9074-x Schiff, R., Raveh, M. & Fighel, A. (2011) The development of the Hebrew mental lexicon: When morphological representations become devoid of their meaning. Scientific Studies of Reading, 16(5), 383-403. DOI: 10.1080/10888438.2011.571327   49 Taft, M. & Nguyen-Hoan, M. (2017) A sticky stick? The locus of morphological representation in the lexicon. Language and Cognitive Processes, 25(2), 277-296. DOI: 10.1080/01690960903043261 Tong, X., Tong, X. & McBride, C. (2016) Unpacking the relation between morphological awareness and Chinese word reading: levels of morphological awareness and vocabulary. Contemporary Educational Psychology, in press.  Vacca, J. L.,  Vacca, R. T., Gove, MK, Burkey, L. C., Lenhart, L. A. & McKeon, C. A. (2009) Reading and Learning to Read 7th edition. Boston, MA: Pearson.  Wagner, R. K., Torgesen, J. K., Rashotte, C. A. & Pearson, N. A. (2013) Comprehensive Test of Phonological Processing–Second Edition [Assessment Instrument]. Austin, TX; Pro-Ed Inc.  Wang, M., Cheng, C. & Chen, S-W. (2006) Contribution of morphological awareness to Chinese-English biliteracy acquisition. Journal of Educational Psychology, 98(3), 542-553. DOI: 10.1037/0022-0663.98.3.542  Woodcock, R., Mather, N. & McGrew, K. (2001, 2007c) Woodcock-Johnson Test of Cognitive Abilities–Third Edition [Assessment Instrument]. Rolling Meadows, IL; Riverside.  Yeh, Y-F., Joshi, R. M. & Ji, X. R. (2015) The development of morphological awareness in Chinese ESL students. Contemporary Educational Psychology, 43, 51-60. DOI: 10.1016/j.cedpsych.2015.10.002  Zhang, D. & Koda, K. (2013) Morphological awareness and reading comprehension in a foreign language: A study of young Chinese EFL learners. System, 41, 901-913. DOI: 10.1016/j.system.2013.09.009   50 Appendix  1.1 MAST Protocol   51    52    53    54    55    56    57    58  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0365823/manifest

Comment

Related Items