第二语言阅读困难是当前儿童阅读学习及二语学习中值得研究的问题。二语阅读困难具有高度异质性，准确分析二语阅读困难不同亚型，有助于理解二语阅读及个体差异的认知机制，并为诊断和干预二语学习困难提供重要依据。二语阅读学习受到多种因素的影响，二语阅读相关认知技能、母语阅读水平、一般认知能力以及家庭环境都可能影响二语阅读学习。以往研究很少系统探索二语阅读学习困难的亚类型，且有关二语阅读学习的研究主要针对母语和二语均为字母语言的二语学习者。本论文以汉语为母语的小学4-6年级英语学习者为研究对象，分为五个研究，从语音意识、快速命名、语音记忆、正字法技能等主要的语言认知技能入手，重点探讨汉语儿童英语阅读困难者在英语语音和正字法加工技能缺陷类型的分布，系统对比具有不同技能缺陷类型的英语阅读困难儿童在英语单词认读和解码、英语和汉语语言技能上的特征，并考察母语阅读水平、语音加工技能干预、一般认知能力和母亲受教育水平等因素对于二语阅读困难不同亚型的可能影响。

首先，本论文基于语音加工、正字法加工缺陷等阅读困难主要理论假设，对母语为汉语的英语阅读困难儿童进行认知技能缺陷亚型分组，以系统分析特定认知技能缺陷的发生率、英语语音和正字法加工技能缺陷各亚型分布，并通过各亚型与正常组、以及各亚型组间对比，考查英语阅读学习困难不同亚型儿童在英语阅读与语言认知技能上的特征。研究1从1402名4-6年级的英语学习儿童中筛选出376名阅读困难儿童和490名阅读正常儿童，按照低于正常组均值1个标准差定义技能缺陷，对英语阅读学习困难的认知技能缺陷亚型分布及组间差异进行比较。结果显示，语音意识和快速命名双重技能缺陷达到22%，单纯语音意识和快速命名的技能缺陷发生率均超过10%，单纯语音记忆缺陷发生率仅为6%，且语音记忆缺陷并不显著增加语音意识和快速命名单一或双缺陷儿童的阅读困难程度。可见，语音意识和快速命名缺陷是阅读困难的主要亚型，语音记忆对阅读困难影响有限。研究2从986名4-6年级儿童中筛选出215名阅读困难儿童和187名阅读正常儿童，同样按照低于正常组均值1个标准差定义技能缺陷，进一步考察了语音意识、快速命名和正字法意识三项语言认知技能缺陷的类型分布及不同亚型的特点。结果显示，英语正字法意识缺陷比例达到68%。研究结果验证了英语语音意识、字母快速命名缺陷发生率均达到40%以上，而且字母快速命名缺陷组的英语阅读能力显著低于其他单一技能缺陷组。正字法意识缺陷虽然发生比例较大，但该缺陷未显著增加语音意识和快速命名双缺陷组的阅读困难程度。上述两项研究显示，语音意识、快速命名、正字法缺陷是汉语儿童英语阅读学习的语言认知技能主要缺陷，可以对接近90%的英语阅读学习困难儿童进行分类。两项研究还一致性地显示，存在两项和三项技能缺陷的英语阅读困难儿童比例（分别为31%和48%）高于任何单项技能缺陷儿童比例（6%到26%），且语言认知技能缺陷数量越多，英语阅读困难程度加重。因此，多重语言认知技能缺陷假设比语音单一核心技能缺陷假设更能解释英语阅读学习的个体差异。

其次，二语阅读往往建立在母语阅读基础上。本论文进一步探讨了英语阅读学习困难不同亚型儿童在母语阅读能力及语言认知技能上的特点。研究3a和研究3b的研究对象分别对应研究１和研究２，分别考查376名和215名英语阅读困难儿童在汉语阅读能力和汉语语音加工和正字法加工技能上的特点。结果显示：（1）英语阅读困难儿童在汉语阅读、汉语语音意识、汉语语音记忆和数字快速命名技能上均显著低于英语阅读正常儿童，F(3, 861) = 39.75, p < .001, η²_p = .122，效应量为0.56-1.43，英语阅读困难组和正常组在汉语正字法意识上不存在显著差异。（2）英语阅读困难者汉语语音意识缺陷的发生率为34%，数字快速命名的缺陷率为31%，汉语语音记忆缺陷率为12%，汉语正字法意识缺陷为9%。（3）英语阅读困难者中英汉语音记忆缺陷和英汉快速命名缺陷关联显著，但是英汉语音意识缺陷相关不显著；虽然英汉正字法意识缺陷同样存在关联，但是汉语正字法意识缺陷率（9%）显著低于英语正字法意识缺陷率（68%）。（4）英语阅读困难不同类型儿童在汉语语音意识上不存在显著差异；英语语音记忆缺陷组在汉语语音记忆上显著低于其他单一缺陷组；字母快速命名缺陷组在汉语数字快速命名上显著低于其他单一缺陷组；英语正字法意识缺陷组在汉语正字法意识缺陷上显著低于其他单一缺陷组。（5）随着汉语语言认知技能缺陷数量增加，不同技能缺陷亚型的英语阅读困难程度加剧，与正常组在英语单词阅读上的效应量从1.29增加到2.78，但是汉语阅读能力高低和英语语言认知技能缺陷数量并没有显著关联。（6）潜变量分析验证了英语阅读困难者存在英语语音意识、快速命名和正字法意识缺陷，有部分英语阅读困难者同时存在汉语阅读困难，他们在英汉语音意识、语音记忆和快速命名上存在缺陷但是正字法意识正常。上述结果提示，英语阅读困难在认知技能缺陷上的亚型不是汉语阅读能力或汉语认知技能的简单映射，汉语阅读能力和英语阅读困难认知技能缺陷亚型没有明显关联，英汉阅读困难存在共发与分离。

第三，研究普遍认为教育经验不足可能影响到英语阅读学习困难，那么英语阅读困难不同亚型是否因教育经验而改变呢？研究1中，阅读困难者随着年级增加，语音意识缺陷比例有所下降，语音记忆和快速命名缺陷的比例在4-6年级间相对稳定。为进一步考察教育经验的作用，研究4通过为72名英语阅读困难儿童提供语音意识和字母声音知识为核心技能的高强度干预，探讨英语阅读困难不同亚型的教育可变性。研究采用每组6-8人的小组干预形式，每次60分钟，每周2-3次，共计600分钟。结果显示有44%的英语阅读困难儿童从语音技能干预中显著获益，为干预应答者，其语音技能和解码能力达到正常组水平，另有56%的英语阅读困难儿童在语音技能和解码能力上没有达到正常组水平，为干预无应答组。对干预无应答的英语阅读困难儿童中，英语语音意识、字母快速命名、汉语语音记忆和数字快速命名的缺陷率显著高于干预有应答组。进一步对比分析认知技能缺陷不同亚型英语阅读困难儿童在干预后的变化，结果显示，英语正字法缺陷亚型和英语语音意识缺陷亚型干预应答较好，有应答的比例超过50%；有字母快速命名缺陷的亚组应答比例仅为20%。随着英语语言认知缺陷种类的增加，对语音技能干预的应答越差，存在三种以上语言认知技能缺陷者的应答比例仅为18%。不过，一年后英语阅读困难组平均水平仍然落后于正常组均值1个标准差以上，且语音意识、快速命名和正字法意识技能缺陷亚型分组保持稳定。可见，短期高强度语音技能干预可以修复部分阅读困难者即时的单词解码能力，但是长期影响有限，特别是对阅读困难的技能缺陷亚型分类影响较小。

综合以上结果，本研究表明，（1）汉语为母语的英语阅读学习困难儿童存在和英语母语阅读困难儿童相似的语音加工技能和正字法加工技能缺陷各亚型，其中正字法意识缺陷比例最高，快速命名技能缺陷的影响最严重，英语语音意识和快速命名缺陷更强的制约了单词解码能力，正字法意识缺陷主要影响单词阅读。（2）任何单一技能缺陷都不足以覆盖英语阅读困难儿童的整体，且各类技能缺陷存在较高比例共发。因此，汉语儿童英语阅读困难存在多种语言认知技能缺陷，支持阅读困难的多重技能缺陷假设。（3）英汉阅读困难存在共发与分离，英语阅读困难不同亚型儿童的汉语阅读能力均显著低于阅读正常组，英汉语音记忆和快速命名缺陷的发生存在显著关联，而英汉语音意识缺陷和英汉正字法意识缺陷的发生相对独立。（4）随着年级增长，语音意识缺陷比例逐渐减少，而其他语言认知技能缺陷的比例相对稳定。在给予充分的语音技能干预后，语音意识缺陷显著下降，但是长期看各项技能缺陷亚型比例没有显著改变。因此，英语阅读困难亚型具有一定的稳定性。

Second language reading difficulty is worthwhile studied in children reading and second language learning. As second language reading difficulty is heterogeneous, it is important to identify subgroups of second language reading difficulties for understanding the cognitive mechanism of second language reading and individual differences, and providing evidence for assessment and intervention. Multiple factors, such as second language reading skills, first language reading, general cognitive ability and family environment will influence second language reading. The previous studies seldom studied the subtypes of second language reading difficulty while most examined second language learners whose first and second languages were both alphabetic languages. The present dissertation consisted of five studies, which selected Chinese children at the fourth to sixth grades as the participants and examined their phonological awareness, rapid naming, phonological memory and orthographic skills with the focus on the distribution of English phonological and orthographic processing skills deficits. It compared the differences of subtypes of second language reading difficulties in English word reading and decoding, English and Chinese reading skills, and further examined the influence of first language reading, phonological skill instruction, general cognitive ability and mother’s educational level on the subtypes of second language reading difficulties.

First, on the basis of phonological processing deficit hypothesis and orthographic processing deficit hypothesis, this dissertation classified the subtypes of Chinese children with English reading difficulties and described the distribution of the subtypes of reading skill deficits and compared their reading and reading skills. The Study 1 screened 376 children with second reading difficulty and 490 typically developing children from 1402 students at 4^th to 6^th grades and examined the distribution of phonological processing skill deficits using one standard deviation below the mean of typically developing readers to define reading skill deficits. The results showed that the double deficits of English phonological awareness and rapid letter naming accounted for 22 percent, the single deficit of phonological awareness or rapid letter naming was more than 10 percent while the rate of phonological memory deficit was only 6 percent and the extra deficit of phonological memory would not make the single deficit or double deficits of phonological awareness and rapid naming reading worse. Therefore, phonological awareness and rapid naming are the two main subtypes of phonological processing skill deficits as the deficit of phonological memory was minor. The Study 2 further screened 215 participants from 986 students at 4^th to 6^th grades and examined the distribution of subtypes of deficits of phonological awareness, rapid letter naming and orthographic awareness. The results showed that orthographic awareness deficit accounted for 68 percent. It also confirmed the deficits of phonological awareness and rapid letter naming reached 40 percent and the subtype of rapid letter naming deficit was significantly lower than other subtypes in English word reading. Although orthographic awareness deficit was the most frequent deficit, the additional deficit of orthographic awareness did not worsen the reading difficulty of double deficits of phonological awareness and rapid naming as well. To sum up, the two studies indicated the deficits of phonological awareness, rapid naming and orthographic awareness were the most frequent deficits in second language reading difficulties, in which almost 90 percent students can be categorized into theses subtypes. Both studies indicated that the rates of double deficits and the triple deficits of phonological awareness, phonological memory, rapid letter naming and orthographic awareness (31 percent and 48 percent respectively in Study 1 and Study 2) were higher than other single deficits (ranging from 6 percent to 26 percent).  Furthermore, the larger the number of deficits, the worse English word reading would be. Thus, the multiple deficits hypothesis is more accountable than the single deficit in the prediction of second language reading difficulties.

Second, the acquisition of second language reading depends on first language reading. This dissertation further examined Chinese reading and skills of the subtypes of second reading difficulties. The Study 3a and Study 3b using the same participants in the Study 1 and Study 2 examined their Chinese phonological processing and orthographic processing skills respectively. The results showed that (1) All the English reading difficulty subtypes were significantly worse than the typically developing children on Chinese word reading, Chinese phonological awareness, Chinese phonological memory and rapid digit naming, F(3, 861) = 39.75, p < .001, η²_p = .122，ds=0.56-1.43, but there was no significant difference on Chinese orthographic awareness. (2) The rate of Chinese phonological awareness deficit was 34 percent, Chinese rapid digit naming deficit was 31 percent, Chinese phonological memory deficit was 12 percent and Chinese orthographic awareness deficit was 9 percent. (3) The deficits of phonological memory and rapid naming in English and Chinese were significantly correlated, but the deficits of phonological awareness in English and Chinese were not significantly correlated. Although English and Chinese orthographic awareness deficits were significantly correlated, the rate of Chinese orthographic awareness (9 percent) was significantly lower than English counterpart (68 percent). (4) There was no significant difference on Chinese phonological awareness among the subtypes. The rapid letter naming deficit group was significantly worse than other subtypes on rapid digit naming；the English orthographic awareness deficit group was significantly worse than other subtypes on Chinese orthographic awareness. Besides, the English phonological memory deficit group was significantly lower than other groups on Chinese phonological memory. (5) The more deficits of Chinese reading skills, the more difficult English reading would be and the effect sizes with the typically developing group increased from 1.29 to 2.78; but there was no significant correlation between the number of English skill deficits and Chinese word reading. (6) The latent profile analysis revealed that English reading difficulties had deficits in English phonological awareness, rapid naming and orthographic awareness. The readers with both English and Chinese reading difficulties had deficits in phonological processing skills but intact in orthographic awareness. The results suggested that the subtypes of English skill deficits were not the simple reflection of Chinese reading and reading skills. There were overlap and separation in English and Chinese reading difficulties.

Third, it was claimed that inadequate instruction might lead to reading difficulty. In the Study 1, the deficit of phonological awareness reduced from the 4^th to 6^th grade, but the rates of phonological memory and rapid letter naming deficits were stable. In order to examine the role of instruction, the Study 4 provided an intensive phonological skill instruction to 72 at risk readers to examine the change of subtypes. The instruction was conducted in 6-8 groups twice or three times every week, which last sixty minutes every time with totally 600 minutes instruction for every child. After the intervention, 44 percent of struggling readers (responders) caught up with the typically developing readers both in phonological processing skills and word decoding, but 56 percent of struggling readers (non-responders) still lagged behind after the intervention. Non-responders and responders were significantly different on English reading skills, Wilks’ lambda = .286, F (4, 61) = 38.08, p < .001, η²_p = .714. The non –responders had more deficits than responders in English phonological awareness, rapid letter naming, Chinese phonological memory, and rapid digit naming. It further compared the change of reading and reading skills after intervention among the subtypes. The results revealed that more than 50 percent at risk readers with only phonological awareness deficit or orthographic awareness deficit made response but only 20 percent at risk readers with rapid letter naming deficit made response to intervention. With more deficits in English reading skills, the response rate became lower and the rate of response for the triple deficits was only 18 percent. The one year follow-up also showed second language reading difficulty group was still 1 standard deviation below the mean of typical readers in word reading and the subtypes of reading skill deficits maintained stable. Therefore, effective instruction could reduce the incidence of second reading difficulties to some extent but the long term effect was limited.

In conclusion, the research suggests that (1) second language reading difficulties can be classified into subgroups with phonological processing and orthographic processing skill deficits as native English reading difficulties, in which the rate of English orthographic awareness deficit was the highest and the reading difficulty of rapid letter naming deficit was more serious than other skill deficits; phonological awareness and rapid letter naming influenced word decoding skills while orthographic awareness affected word reading. (2) The single deficit hypothesis cannot account for all the reading difficulties and the skill deficits occurred simultaneously at high rates. Thus, the distribution of subtypes of multiple deficits in Chinese poor English readers supported the multiple deficits hypothesis. (3) Children with second language reading difficulty were also inferior to typically developing readers in Chinese reading but there was no significant difference among the subtypes. English and Chinese phonological memory deficits and rapid naming deficits were correlated but English and Chinese phonological awareness deficits and orthographic awareness deficits were language independent. There was overlap and separation in English and Chinese reading difficulties. (4) Phonological awareness deficit would reduce with age but other subtypes remained stable. After the intensive instruction, the rate of phonological awareness deficits decreased but the subtypes of reading difficulties didn’t change significantly in the long term. Thus, the classification of skill deficits subtypes was comparatively stable.