The Relative Contribution of L1 Reading Ability and L2 Proficiency to L2 Reading Ability.doc

The Relative Contribution of L1 Reading Ability and L2 Proficiency to L2 Reading AbilityAbstract：This study aims at exploring the relative contribution of L1 reading ability and L2 proficiency to L2 reading ability in terms of the linguistic threshold hypothesis (LTH) and the linguistic interdependence hypothesis (LIH). It addresses the following two research questions: 1. What is the relative contribution of L1 reading ability and L2 proficiency to L2 reading ability at the low L2 proficiency level ? 2. What is the relative contribution of L1 reading ability and L2 proficiency to L2 reading ability at the high L2 proficiency level ? In this study, L2 proficiency is operationalized as a comprehensive English proficiency test paper, which includes listening comprehension, vocabulary, grammar and writing test. Another two reading test papers are employed for the measurement of L1 reading ability and L2 reading ability respectively. Eighty English majors, who are either freshmen or juniors, participated into the research and each one of them completed these three tests. Then, based upon their scores on L2 proficiency test, the subjects are divided into two groups: the low L2 proficiency group and high L2 proficiency group. Multiple regression analyses of the two variables predicting L2 reading ability indicate that at the low L2 proficiency level, only L2 proficiency plays a role in L2 reading ability and that at the high L2 proficiency level, both L1 reading ability and L2 proficiency contribute significantly to L2 reading ability, with L2 proficiency remaining a better predictor. The results confirm LTH and provide partial support for LIH, thus suggesting important pedagogical implications for the second language instruction. If the students L2 proficiency is below the threshold level, teachers should attach much importance to L2 linguistic knowledge instruction, otherwise, teachers should include L1 reading strategies instruction to help their students to have a smooth transfer of L1 reading skills and strategies to L2 reading. In addition, teachers should choose appropriate reading materials for their students based on their L2 proficiency levels, so that they could not only understand the major content of the reading passages by employing L1 reading strategies and skills as well as L2 knowledge in the reading process, but acquire L2 linguistic knowledge from the reading materials to further improve their L2 proficiency level. Key words: L1 reading ability L2 proficiency L2 reading ability linguistic threshold hypothesis linguistic interdependence hypothesis母语阅读能力和外语语言水平对外语阅读能力的不同作用摘要：本文旨在研究母语阅读能力和外语语言水平对外语阅读能力的不同作用，并对两个理论假设（语言阈限假设和不同语言相互依赖性假设）进行检验论证。主要回答以下两个问题：1当外语语言水平较低时，母语阅读能力和外语语言水平对外语阅读能力的作用关系如何？2当外语语言水平较高时，母语阅读能力和外语语言水平对外语阅读能力的作用关系如何？在本研究中，外语语言水平由一份综合性语言水平测试卷测量，包括听力理解、词汇、语法和写作四个部分。另外两份阅读试卷分别用于测量受试的母语阅读能力和外语阅读能力。共80名英语专业一年级和三年级学生参加了实验，每人按要求完成了三份试卷。基于不同的外语语言水平，将受试分成两组：低水平组和高水平组。对两组数据的多元回归分析表明：当外语语言水平较低时，只有外语语言水平对外语阅读能力有显著的贡献作用；当外语语言水平较高时，母语阅读能力和外语语言水平都对外语阅读有贡献作用，但是外语语言水平的贡献作用要比母语阅读能力大得多。因此，研究结果基本证实了语言阈限假设，同时也部分支持了不同语言相互依赖性假设。本次研究结果告诉我们：针对不同外语语言水平的学生，教师应采取不同的教学方法。对于外语语言水平低于语言阈限水平的学生，教学重点应放在提高学生的外语语言技能上；对于外语语言水平越过语言阈限的学生，教师应注重母语阅读技能的训练，使学生在外语阅读中更好地运用母语阅读技能。此外，教师应当根据学生的外语语言水平有目的地选择难度适宜的阅读材料，使他们既能通过在阅读过程中运用母语阅读技能和外语知识恰如其分地领会文章的主要内容，又能有效地吸收、积累阅读材料中的词法、句法等，从而进一步提高其外语语言水平。关键词：母语阅读能力 外语语言水平 外语阅读能力 语言阈限假设 不同语言相互依赖性假设1. IntroductionThe past quarter century has witnessed much first and second language reading research, resulting in many new insights into second language (L2) reading instruction. During this period of time, language researchers have posed many seminal research questions concerning L2 reading, among which is whether second language reading is a language problem (linguistic threshold) or a reading problem (linguistic interdependence).The linguistic threshold hypothesis (LTH) posits that L1 reading ability is “short-circuited” if the reader does not have an adequate level of L2 knowledge whereas the linguistic interdependence hypothesis (LIH) supposes that L1 reading ability can be greatly transferred to L2 reading. 1.1 Need for the studyAlthough several experimental studies have been carried out both at home and abroad to verify these two hypotheses, almost all of them have exposed a major problemthe inappropriate operational definition of L2 proficiency. Some researchers either use the subjects grade level or their instructional level or their part of L2 language proficiency (vocabulary and grammar) to represent their overall L2 language proficiency. Whats more, the related experiments have rarely been done in the Chinese context. In contrast, most of them have been undertaken in the context of Indo-European languages. Whether the conclusions drawn from these studies are applicable in the Chinese context is still unknown yet. Consequently, further research into the relative contribution of L1 (Chinese) reading ability and L2 (English) proficiency to L2 (English) reading ability needs to be carried out.Except for theoretical needs, the research into this question in the Chinese context should be undertaken also because the corollary of it will have significant implications for teaching English reading in China. If the linguistic threshold hypothesis (LTH) is verified, the teachers should focus on improving students second language proficiency. If the linguistic interdependence hypothesis (LIH) is confirmed, then presumably the teaching of foreign language reading should include instruction in appropriate L1 reading strategies so that the students can transfer their L1 reading strategies to L2 reading. 1.2 Research purposeThe present study is intended to find out the relative contribution of L1 reading ability and L2 language proficiency to L2 reading ability. Specifically speaking, how much contribution Chinese (L1) reading ability and English (L2) proficiency make to English (L2) reading ability at different L2 proficiency levels will be examined in the current empirical study. Its purpose is to ascertain whether LTH or LIH can still be confirmed even if L2 proficiency varies.2. Literature reviewIn this chapter, previous literature is reviewed in three parts, first on two theoretical hypothesesthe linguistic threshold hypothesis (LTH) and the linguistic interdependence hypothesis (LIH), second on related empirical studies testing the two hypotheses and third on problems in the previous studies.2.1 Two theoretical hypotheses: LTH and LIHFor more than thirty years, the question of whether reading in a foreign language is a reading problem or a language problem has aroused much attention in the second language reading research community, beginning with Aldersons (1984, p. 4) testing of the following hypotheses:1. Poor reading in the L2 is due to poor ability in the L1; therefore, poor L1 readers will read poorly in the L2 and good L1 readers will read well in the L2. 2. Poor reading in the L2 is due to inadequate knowledge of the L2.Underlying these hypotheses are two other seemingly conflicting yet complementary ones: the linguistic threshold hypothesis (LTH) and the linguistic interdependence hypothesis (LIH).The first hypothesis, also known in its earliest iterations as the short circuit hypothesis, is originally stated as: In order to read in a second language, a level of second language linguistic ability must first be achieved (Bernhardt & Kamil, 1995). The second hypothesis, at times termed Common Underlying Proficiency (CUP), is originally stated as: L2 reading ability is largely shared with L1 reading ability. Concretely speaking, LIH assumes that language operations such as reading and writing are transferable and intertwined; that is, second language skills are only superficially distinct and in essence they are interdependent or the same. 2.2 Related empirical studiesSo far，a number of reading specialists have attempted to verify these two hypotheses empirically so as to identify the main factor to influence L2 reading comprehension.Firstly, many experiments have provided strong evidence for LTH (Cummins, 1979; Carrell, 1988b; Bossers, 1991; Davis & Bistodeau, 1993; Taillefer, 1996; Lee & Shallert, 1997; Schoonen, Hulstijn and Bossers, 1998). Cummins (1979, p. 229), based on studying the reading behavior of English-French bilingual children, proposed that the positive effects of balanced bilingualism were “unlikely to come into effect until the child has attained a certain minimum of threshold level of competence in a second language.” Studying the strategies of low-level ESL students who were either good or poor readers in their native Spanish and English, Clarke (1978, 1980) found that low L2 proficiency restricted a readers ability to interact with an L2 text. Bossers (1991) used 50 native speakers of Turkish in Dutch post-secondary education as subjects, who read 4 passages in their first and second languages and answered 16 multiple-choice questions per text. Bossers (1991) was able to account for 73 percent of the variance in their scores, 19 percent of which was attributable to first language reading. For the least skilled readers (N=35), only L2 knowledge was able to account for the variance in L2 reading ability; for the most skilled (N=15), L1 reading was the primary predictor, accounting for 34 percent of the variance. These findings, Bossers argues, are evidence for a threshold hypothesis: Direct transfer of L1 reading skills occurs only when a certain level of L2 proficiency has been attained. Bernhardt and Kamil (1995) further examined the relationship between L2 proficiency and L1 reading ability with L2 learners of Spanish at three levels of L2 proficiency. They reported a shared variance of 48% for L1 reading ability and L2 proficiency in L2 reading in their study. Using stepwise regression analysis, they further determined that, at the beginning level, L1 reading ability accounted for about 10% of the variance in L2 reading ability, while L2 proficiency accounted for about 38% of the variance. At the more advanced level, L2 proficiency shared only twice as much variance with L2 reading ability. Their finding is mainly consistent with Lee and Shallert (1997)s : L1 reading ability and L2 proficiency together accounted for 62% of the variance, with L2 proficiency sharing twice as much variance with L2 reading ability as L1 reading ability did. When correlating the scores for two different grade levels of Korean learners of English, Lee and Shallert (1997) observed that the correlation between L1 reading ability and L2 reading ability was significant and stabilized in an ascending pattern only beyond level 2. Like Lee and Shallert, Taillerfer (1996) and Schoonen et al. (1998) provided evidence for support of the LTH in their experimental studies as well.Secondly, in reference to LIH, there are far few experimental studies verifying it. As early as in 1977, Pat Rigg, who, using miscue analysis, found considerable similarities in reading miscues for EFL learners from a variety of different language backgrounds. If the reading process is the same or very similar in all languages, then one could expect reading ability to transfer across languages. That is, as Clarke (1978, p. 121) states, “If the reading process is basically the same in all languages, we would logically expect good native language readers to be good second language readers. Furthermore we would expect good readers to maintain their advantage over poor readers in the second language.” Shohamys study (1984) of different methods of measuring reading comprehension (multiple-choice or open-ended questions in the native or target language) provided compelling yet confounded (due to some target language response measures) evidence for LIH. She found that she could account for student performance across proficiency levels based on the method of testing in conjunction with the language used in the test: the more L2 proficient the students were, the less influence the L2 proficiency had on their reading performance.Lastly, other studies of second language reading provide support for both LTH and LIH. Carrell (1991) used two different sets of subjects: 45 first language speakers of Spanish at varying levels of L2 proficiency and 75 speakers of English at three levels of Spanish proficiency. It was observed that for both groups, L1 reading and L2 proficiency contributed significantly to L2 reading. Together, L1 reading and L2 proficiency accounted for 35% of the variance in the case of the Spanish L1 group and 53% in the case of the English-speaking group. As to the relative importance of each of these variables, L2 proficiency was a better predictor of reading performance for the English L1 group ( about four times more powerful), whereas L1 reading ability was better for the Spanish L1 group (about twice as powerful). Carrell (1991) attributed this difference to the fact that the Spanish L1 group was learning English in a second (not foreign) language context. Brisbois (1992) made the most extensive analysis of literacy interdependence and linguistic threshold. She tested 126 college-level learners of French using an array of predictor variables: two L1 reading measures, an L1 writing measure, a vocabulary measure, and a grammar measure. Brisbois found that L1 reading (measured through a standardized reading test for adults) accounted for 14 percent of the variance in second language reading scores (measured across two passages). Other variables (L1 writing ability, L2 vocabulary, and L2 grammar) accounted for a total of 59 percent of the variance in the comprehension scores. Both Carrells and Brisbois studies provide compelling evidence for both LTH and LIH: L1 reading ability and L2 proficiency together play a role in successful L2 reading. When the relative contribution of each variable is examined, L2 proficiency appears to play a greater role initially and L1 reading ability begins to play a greater role at more advanced levels of L2 proficiency, which constitutes evidence for a language threshold. Below this threshold, the transfer of L1 reading skills to L2 is presumably obstructed by limited L2 knowledge.2.3 Problems in the previous studiesIn a critical review of the design of the previous experimental studies, Alderson (1984, p. 21) suggested that studies pertaining to the relationships among L2 proficiency, L1 reading, and L2 reading be designed as follows: What is needed is at least information on reading ability in the first language, ideally where the sample of subjects contained both “good” and “poor” first-language readers; information of the reading ability in the foreign language; and information about the level and, importantly, the nature, of the foreign language proficiency of the same individual.However, there are still some defects in the designs of previous experi