全民阅读功在当代，利在千秋，儿童阅读更是全民阅读的基础。儿童良好的阅读能力不仅是其积累知识、提升素养的重要途径，也是国民素质提高和社会文明进步的支撑和保障。然而，儿童的阅读能力并非与生俱来，而是后天习得的结果。其中，字词识别不仅是阅读过程的基础，也是儿童学习阅读并良好发展的必要条件。因此，如果想深入了解汉语儿童阅读能力的发展规律，有效提高儿童阅读素养，就需要明确儿童字词加工的过程和影响因素，并进一步明晰其字词认知发展模式。以往研究更多关注儿童认知能力对字词识别的影响或者字词识别对阅读能力的影响，较少关注儿童字词认知本身的加工过程和发展机制，以及字词特征在儿童字词认知中的作用，关于汉语儿童字词认知的发展过程以及影响因素尚不清楚。为了全面考察字词加工的影响因素，越来越多的语言建立了心理语言学数据库，采用大型研究的方法考察了字词特征在字词加工中的重要程度，并且大量研究报告了心理语言学特征在字词加工中也具有重要的作用。然而，以往的研究往往仅关注成年读者的字词加工。因此，字词特征和心理语言学特征在汉语儿童字词认知中的作用还需要进一步探讨。基于以上背景，本研究通过四个系列研究(共包括七个小研究)，明晰字词特征和心理语言学特征在儿童字词加工和阅读中的作用，进而阐明汉语儿童字词认知和阅读的发展过程和影响因素，并构建了汉语字词认知的系统框架。
研究一包含两个子研究，分别构建了汉字和汉语词汇的心理语言学数据库，并考察了心理语言学特征在汉字和词汇加工中的作用。研究1A基于《通用汉字规范表》等构建了包含3600个汉字的心理语言学数据库，通过441名被试评定了汉字的4个心理语言学特征，即获得年龄、熟悉度、可表象性和具体性，并且收集了这些汉字的频率、字形、字音、字义等12个客观特征。以大学生的汉字命名反应时和正确率为因变量，考察了汉字客观特征和心理语言学特征在汉字加工中的作用。基于大样本的回归分析表明，汉字的客观特征能够显著预测成人汉字命名的反应时和正确率。更重要的是，在控制了汉字客观特征后，汉字的心理语言学特征(获得年龄、熟悉度、可表象性、具体性)均能够显著增加额外的解释率。结果表明，汉字心理语言学特征在汉字加工中具有重要的作用。研究1B基于《义务教育常用词表》构建了包含10426个双字词的汉语词汇心理语言学数据库，通过1317名被试评定了这些词汇的4个心理语言学特征，即获得年龄、熟悉度、可表象性和具体性，并且收集了这些词汇的14个客观特征，以MELD-SCH中的词汇判断反应时和正确率为因变量，考察了词汇客观特征和心理语言学特征在词汇加工中的作用。基于大样本的回归分析表明，词汇的客观特征能够显著预测成人词汇判断的反应时和正确率。更重要的是，在控制了词汇客观特征后，词汇的获得年龄、熟悉度、可表象性均能够显著增加额外的解释率。结果表明，词汇的心理语言学特征在词汇加工中具有重要的作用。
研究二包含两个子研究，基于汉语心理语言学数据库，以三年级和五年级儿童为被试，分别考察了字词特征在儿童汉字和词汇加工中的作用。其中，研究2A基于汉字心理语言学数据库中的3600个字，以422名三年级和485名五年级儿童为被试，通过汉字命名实验考察了儿童的汉字命名正确率和反应时。结果表明，三年级和五年级儿童的汉字加工准确性和自动化程度均取得了较大的发展，五年级儿童的汉字加工已接近成人水平。回归分析表明，汉字的客观特征和心理语言学特征在三年级和五年级儿童的汉字加工中均具有重要的作用。相对重要性的分析表明心理语言学特征在儿童汉字加工中的重要程度与客观特征类似。线性混合效应模型的结果表明，汉字特征在三年级儿童的汉字加工中具有更加重要的作用。研究2B基于汉语词汇心理语言学数据库中的10426个双字词，以630名三年级和593名五年级儿童为被试，通过词汇判断实验考察了儿童的词汇判断正确率和反应时。结果表明，三年级和五年级儿童的词汇加工准确性和自动化程度均取得了较大的发展，但与成人读者仍有一定距离。回归分析表明，词汇的客观特征和心理语言学特征在三年级和五年级儿童的词汇加工中均具有重要的作用。相对重要性的分析表明所有词汇特征对于词汇加工的准确性更加重要。而心理语言学特征在词汇加工中也具有非常重要的作用。线性混合效应模型的结果表明，词汇特征在三年级儿童的词汇加工中具有更加重要的作用。
研究三包含两个子研究，以三年级和五年级儿童为被试，采用句子阅读的眼动实验，分别考察了字词客观特征和心理语言学特征在儿童自然阅读中的作用。研究3A考察了42名三年级儿童和26名五年级儿童句子阅读中的眼动过程。通过以汉字为兴趣区，提取了句子中所有汉字的特征以及儿童的眼动表现。线性混合效应模型的结果表明，与单个汉字识别一致，汉字的客观特征以及心理语言学特征在儿童的汉字阅读中也均具有重要的作用。不同的是，儿童的汉字阅读会受到句子特征的影响。此外，汉字可表象性在汉字命名和阅读中具有不同的作用。交互效应表明，汉字特征在三年级和五年级儿童汉字阅读中的作用类似。研究3B基于研究3A，通过以词汇为兴趣区，提取了句子中所有词汇的特征以及儿童的眼动表现。线性混合效应模型的结果表明，与单个词汇识别一致，词汇的客观特征以及心理语言学特征在儿童的词汇阅读中也均具有重要的作用。不同的是，儿童的词汇阅读会受到句子特征的影响。此外，词汇可表象性在词汇判断和阅读中具有不同的作用。交互效应表明，词汇特征在三年级和五年级儿童词汇阅读中的作用类似。
研究四结合前三个研究的数据集，通过机器学习算法综合考察了汉语心理语言学特征在字词认知中的相对重要性。首先，通过43名大学生补充了成人字词阅读的表现。结合前面六个子研究中的各个数据集，共包含了六个汉字加工和六个词汇加工的数据集。以成年人、三年级儿童和五年级儿童各个字词加工表现为因变量，字词客观特征、心理语言学特征、句子特征以及读者特征为自变量，采用随机森林算法共构建了48个模型，以评估所有特征在汉语认知中的相对重要性。结果表明，现有的心理语言学特征在汉语的字词认知中均具有非常重要的作用，并且重要程度与字词客观特征相当。不同年级儿童和成人研究筛选出的重要特征也基本一致，进一步表明汉语儿童和成人字词加工模式的一致性。但相较于孤立的字词识别，句子特征和读者特征在字词阅读中也具有重要作用。
综合以上结果，本研究构建了汉语字词认知发展的系统框架，基于心理语言学数据库，采用大型研究的方法，不仅从宏观层面发现了汉语心理语言学特征在儿童的字词认知中的作用，也从更加生态的角度探究了各种特征的重要程度。在汉语心理语言学特征中，字词的获得年龄和熟悉度具有较高的重要性；从具体作用来说，可表象性和具体性在不同字词加工任务中有不同的作用，对于深入探讨汉语儿童的阅读发展过程具有重要意义。另外，字词的客观特征在三年级和五年级儿童的字词加工中也具有重要作用，从字词的角度明晰了汉语字词难度的影响因素。而无论是字词客观特征还是心理语言学特征，在儿童和成人字词加工中作用具有一致性，结果表明汉语儿童的字词加工模式与成人类似。但关于字词识别与句子阅读的结果表明，字词特征在儿童自然阅读中的作用有所减弱，儿童句子阅读中的字词加工还会受到句子特征和读者特征的影响。本研究基于发展的视角，为了解汉语儿童字词加工和阅读发展的规律提供了科学证据，对未来汉语认知研究和阅读教学干预也具有重要指导意义。

Reading has immense value for present and future generations, and the foundation of national literacy is built on children’s reading skills. Children’s ability to read is crucial not only for the accumulation of knowledge and the improvement of education, but also as a support system for the improvement of national quality and social progress. However, reading ability is not innate and word recognition serves as a fundamental aspect of the acquisition process necessary for the optimal development of children’s reading skills. To gain insight into the developmental patterns of Chinese children’s reading skills and to improve their literacy, it is essential to understand the processes and factors that influence character and word processing. Although previous studies have investigated the effects of cognitive skills and word recognition on reading ability, few have examined the mechanisms underlying children’s word recognition or the role of lexical variables in their word processing. Recent mega-studies have provided psycholinguistic databases for many languages, indicating the importance of lexical and psycholinguistic variables in word processing. However, their role in Chinese children’s character and word processing remains unclear. Therefore, this study aims to clarify the roles of these variables through four series of studies (comprising seven sub-studies), to establish a systematic framework for Chinese reading, and to provide a better understanding of the developmental process of Chinese children’s character and word cognition.
Study 1 consisted of two sub-studies that constructed psycholinguistic databases for Chinese characters and compound words, respectively, and investigated the role of psycholinguistic variables in character and word processing. Study 1A created a database of 3,600 Chinese characters. Four psycholinguistic variables, namely age of acquisition, familiarity, imageability, and concreteness, were evaluated by 441 participants, while 12 lexical variables were also obtained. The latent response durations (RT) and accuracy (ACC) of college students’ character naming were used to assess the influence of lexical and psycholinguistic variables on character processing. Regression analyses showed that lexical variables significantly predicted both RT and ACC of character naming. Moreover, after controlling for lexical variables, psycholinguistic variables accounted for additional significant variance. These results suggest that psycholinguistic variables are essential for character processing. Study 1B developed a psycholinguistic database of 10,426 two-character compound words. A total of 1,317 participants rated these words on age of acquisition, familiarity, imageability and concreteness. In addition, 14 lexical variables were collected to examine the influence of lexical and psycholinguistic variables on word processing through RT and ACC of lexical decisions using the MELD-SCH. Regression analyses revealed that, after accounting for lexical variables, age of acquisition, familiarity, and imageability significantly increased the additional variance. These results show that psycholinguistic variables play an important role in word processing.
Study 2 consisted of two sub-studies that investigated the role of lexical and psycholinguistic variables in Chinese children’s character and word processing, respectively, utilizing the Chinese Psycholinguistic Database. Study 2A involved 422 Grade 3 and 485 Grade 5 children and examined their character naming ACC and RT through a character naming task. The results indicated that both Grade 3 and Grade 5 children showed improvements in accuracy and automation of character processing, with Grade 5 children approaching adult levels. Regression analyses highlighted the critical roles played by both lexical and psycholinguistic variables in character processing for both age groups. Relative importance analyses suggested that psycholinguistic variables were similar in importance to lexical variables in character processing, with all variables contributing more significantly to Grade 3 character processing according to the linear mixed effects model. Study 2B examined the ACC and RT of lexical processing in 630 Grade 3 and 593 Grade 5 children through the use of a lexical decision task based on 10,426 two-character words. The results indicated that both age groups showed increased accuracy and automation in word processing, although some gaps remained compared to adult readers. Regression analyses showed that both lexical and psycholinguistic variables played an important role in word processing for both age groups. Relative importance analyses indicated that all variables were more significant for accuracy, with psycholinguistic variables also playing a significant role in word processing. Overall, the linear mixed effects model suggested that all variables had a greater impact on the children’s word processing in Grade 3.
Study 3 included two sub-studies that employed an eye-movement sentence-reading experiment to investigate the role of lexical and psycholinguistic variables of Chinese characters and words in children’s natural reading. Study 3A assessed eye-movement processes during sentence reading in 42 Grade 3 and 26 Grade 5 children. The results showed that children’s character processing was more efficient during sentence reading, with both lexical and psycholinguistic variables of characters playing important roles in children’s character reading, consistent with isolated character recognition. However, children’s character reading was more influenced by sentence variables, and imageability played different roles in character naming and reading. Interaction effects indicated that lexical and psycholinguistic variables had similar effects on character reading. Study 3B, derived from Study 3A, extracted features of all words in sentences using two-character areas of interest and assessed children’s eye movement performance during word reading. The results showed that children’s word processing was also more efficient during sentence reading, with both lexical and psycholinguistic variables playing a crucial role in children’s word reading, consistent with isolated word recognition. In addition, children’s word reading was more influenced by sentence variables, and imageability played different roles in lexical decision and reading. Interaction effects suggested that lexical and psycholinguistic variables influenced word reading similarly for both age groups.
Study 4 integrated the datasets from the previous three studies and evaluated the relative importance of psycholinguistic variables in character and word processing by machine learning. First, 43 undergraduate students were administered the character and word reading performance. Twelve datasets for character and lexical were created by merging individual datasets. Using random forest, a total of 48 models were generated with the character and word processing performance of adults, Grade 3 and Grade 5 as dependent variables and lexical, psycholinguistic, sentence and reader variables as independent variables to determine the relative importance of all variables in character and word processing. The results indicate that psycholinguistic variables play a crucial role in both character and word processing, and their importance is comparable to that of lexical variables. The important variables identified by adults and children showed significant overlap, further confirming the consistency of character and word processing patterns among children and adults. However, compared to isolated character and word recognition, sentence and reader variables had a greater impact on character and word reading.
The present study constructed a systematic framework for Chinese reading with a series of mega-studies based on a psycholinguistic database to examine the role of psycholinguistic variables in children’s character and word processing at the macro level, as well as the importance of various variables from an ecological perspective. The results indicate that, among psycholinguistic variables, age of acquisition and familiarity with the character and word are extremely important. Imageability and concreteness play different roles in different character and word processing tasks and are essential for studying the reading development process of Chinese children in depth. In addition, lexical variables also play a significant role in character and word processing for Grade 3 and Grade 5, clarifying the factors that influence the difficulty of characters and words. While both lexical and psycholinguistic variables consistently affect both children’s and adults’ word processing, the results suggest that children’s character and word processing patterns are similar to those of adults. However, the results for character and word recognition and sentence reading suggest that the role of lexical variables in children’s natural reading has diminished and is influenced by sentence and reader variables. From a developmental perspective, this study provides scientific evidence for understanding the patterns of Chinese children’s character and word processing and reading development, with important implications for future interventions in Chinese literacy and reading instruction.