Linguistic Evidence for Script Relativity

Abstract

Using the universal grammar of reading and the system accommodation hypothesis (Perfetti, 2003) as theoretical frameworks, this chapter reviews a wide range of linguistic evidence that supports script relativity. Universality and specificity found according to script features are discussed with respect to the operating principle (alphabet vs. logography), psycholinguistic gran size (phoneme vs. syllable), graph configuration (linearity vs. block), symbolic representation (arbitrariness vs. iconic quality), graph complexity (traditional characters vs. simplified characters), and multi-script representation (phonogram Kana vs. Ideogram Kanji). Linguistic skills associated with reading in terms of orthography, phonology, morphology as well as cross-linguistic and cross-scriptal transfer are reviewed. Next, based on the reviewed literature, each criterion for causality from script to cognition through reading as a multifaceted cognitive activity is checked. Although the existing literature did not aim to directly test script relativity, research findings collectively suggest script effects on readers’ thought and cognition.

“The relation between thought and word is a living process; thought is born through words. A word devoid of thought is a dead thing, and a thought unembodied in words remains a shadow.”

- Lev S. Vygotsky (1986, p. 255).

“All writing systems represent spoken languages, a universal with consequences for reading processes.”

- Charles A. Perfetti (2003, p. 3)

Vygotsky (1986) stresses the interlocking relationship between language and thought and underscores the fact that thought is dependent on language, as in “thought is born through words” (p. 255). Perfetti (2003) furthermore notes the writing system dovetails with its spoken language. As a collection of Vygotsky’s and Perfetti’s assertions, the interdependent relationships among thought, spoken language, and written language are conceivable. The role of written language has become significant in our daily lives, especially in the digital era.

Given that reading skills are integral for academic success, reading research has been a mainstay in education, psychology, cognitive science, and applied linguistics for many decades. Research has identified the common precursors of fluent reading among emergent readers, including phonological awareness, working memory, phonological retrieval (e.g., rapid automatized naming, RAN), and an awareness of morphological structures within the word. The National Reading Panel (2000) also identified the five pillars of reading based on studies that used experimental or quasi-experimental designs and studies that met rigorous scientific standards, such as well-defined instructional procedures, verified causality from instruction to student outcomes, and large sample size for adequate statistical power and generalizability, as follows: phonemic awareness, phonics, vocabulary, fluency, and reading comprehension. Of these five pillars, the first two are related to phonological properties. As metalinguistic skills, phonological awareness is considered to be a significant predictor of efficient reading in alphabetic orthographies and even in nonalphabetic orthographies, such as Chinese and Japanese, although the level of phonological units involved in reading and the time of phonological activation (i.e., prelexical or post lexical) during reading vary across orthographies (Cho & McBride-Chang, 2005; Perfetti & Liu, 2005). In addition, beyond the three East-Asian scripts, Desrochers et al. (2018) investigated the early contribution of morphological awareness to reading skills among English-, French-, and Greek-speaking children and found that second graders’ morphological skills were a common predictor of reading comprehension and spelling in the three languages, whose magnitude of the contributions was not significantly different among the three languages. The general consensus is that sensitivity to both meta-linguistic phonological properties and morphological structures is script-universal, while orthographic awareness is script-specific (Cho, 2018; Perfetti & Liu, 2005; Wang, Park, & Lee; Yamashita, 2018).

Beyond reading being a critical catalyst for academic success, identifying the significant predictors of reading is important because reading is the complicated activity of information processing. Our habitual reading lays a foundation for our cognitive prototype. Aggregated research results point toward universality involved in reading all scripts as well as specificity found according to the linguistic characteristics of scripts being read. The universal grammar of reading postulates that all writing systems that encode spoken languages capture the universal aspect of reading across scripts (Perfetti, 2003). Relatedly, the linguistic constraints hypothesis addresses how reading involves the way in which spoken language is encoded within the writing system, and, as a result, reading engages in script-dependent reading processes to the extent that the graphic features and orthographic configurations of each script lead to script-specific processes in reading (Perfetti, 2003).

Figure 8.1 shows that commonalities lie in the three East-Asian scripts as well as differences in their own scripts. As explained in Chapter 5, the Chinese script adopts the writing system that has logographic, or morphosyllabic, characteristics, along with Pinyin (a Latin alphabet phonetic system that supplements the morphosyllabic script) as an additional alphabetic system. The Japanese use multi-scripts of logographic Kanji and phonographic Kana including Hiragana and Katakana. The Koreans now officially use alphabetic Hangul only, discouraging Hanja (traditional Chinese characters used in Korea), but a part of the population still learn and use Hanja.

Figure 8.1.

Commonalities and Differences among Chinese, Japanese, and Korean scripts

Note: The parenthesis used for Hanja at the center denotes unofficial use of Hanja in Korean. The size of the balloon indicates the degree of usage in the given writing system; that is, a supplementary use of Pinyin in China, the co-use of Kana in Japan, and the sole use of Hangul in Korea.

Based on the commonalities and differences of the three East-Asian scripts, this chapter reviews studies of word reading in the three East-Asian scripts of Chinese, Japanese, and Korean, in relation to English. Reading is a multifaceted cognitive skill, involving word decoding, vocabulary, and sentence comprehension at the stimulus level as well as attention, memory, retrieval, and inference at the cognitive level. Since sentence comprehension involves a wide range of skills, including word identification, vocabulary, working memory, prior knowledge, and intrinsic and extrinsic motivation, it is difficult to rule out intervening or spurious factors involved in reading. Hence, this review limits its scope to word reading in order to focus on script effects without uncontrollable variables involved. The theoretical frameworks are first discussed. The extant literature is next reviewed in light of scriptal dimensions, including the operating principle, grain size, graph configuration, symbolic representation, graph complexity, and multi-graph representation. Cross-linguistic influences on reading in a second language (L2), in comparison to English when necessary, are also discussed. Hill’s (1965) criteria for causality are examined between independent variables and dependent variables. Finally, based on the results of previous research, the main thesis of this book, script relativity, is revisited to substantiate it with empirical corroboration. It should be noted that previous research is reviewed in a way that each study’s findings are autonomously reviewed in Chapters 8 and 9 in order to provide the reader with more contexts for each study, compared to other chapters that integrate and synthesize all pertinent information for claims made for specific purposes in the given chapter.

1 Theoretical Considerations

Perfetti (2003) claims that “writing systems encode spoken language” (p. 3) such that languages place the constraints of speech on writing systems. The indispensable ties between spoken language and written language are indisputable, because “[t]here are no writing systems currently in use that bypass language to erect an independent system of signs” (Perfetti, 2003, p 5). This claim is consistent with DeFrancis’ criterion (1989) for a full-scale writing system. DeFrancis notes that all full systems of writing are grounded in speech and that, if written symbols do not correspond to the sound of any language, it cannot be considered a full-fledged writing system. This is why pictographs and semasiography (e.g., Blissymbols or Blissymbolics, which is an ideogram comprising several hundreds of pure symbols) do not qualify as a script (Sampson, 2015). Daniels’ (1996) view of the writing system is also in line with this. The pattern and the extent to which spoken language is related to written language are variable across languages and scripts, however. Perfetti (2003) uses the phrase “universal grammar” to dub the universal grammar of reading in order to highlight the universal aspect of reading processes as an overarching framework. He articulates three propositions under the universal grammar of reading such that reading is characterized by spoken language and written language. The first proposition is to define reading as an activity that is inextricably interconnected with a language and its corresponding writing system. The second proposition is to delineate the subcomponents of language that characterize spoken and written languages. This proposition implicates the schematic model of subcomponents as follows: Language comprises grammar, phonology, and pragmatics; grammar consists of syntax and morphology; morphology is composed of lexical roots and inflections; and lexical roots comprise syntactic categories and meaning. Among these universal subcomponents of language, phonology and grammar (particularly the morphological aspect of grammar) are most relevant to reading. The third proposition explains two levels of writing systems: (a) a higher level of the mapping principle at the graphic unit and language levels and (b) a lower level of spelling or orthographic constraints that provide mapping details. The higher level implicates language universal, while the lower level of orthographic constraints associates with language specific.

Under the universal grammar of reading, Perfetti (2003) proposes two universal principles with respect to reading: (1) the universal writing system constraint and (2) the universal phonological principle. The universal writing system constraint postulates that writing systems reflect the linguistic properties of spoken language, as all writing systems express spoken language in written signs (Perfetti, 2003; Perfetti & Liu, 2005; Sampson, 2015). The universal phonological principle notes that the phonology of the word being read is activated at the smallest unit for which its writing system allows. Even Chinese not only conforms to the principle that graphic units are tied to pronunciation, but also makes use of phonological mapping to access meaning in reading, although Chinese phonology is not activated at the phonemic level. The model of reading universality is augmented by accommodation that spoken language and written language provide. Specifically, Perfetti and Liu (2005) further propose the system accommodation hypothesis to explain that both reading processes and neuronal networks and structures¹ that are involved in reading accommodate the specific visual and structural features of the script being read.

Perfetti (2003) provides empirical substantiation for the system accommodation hypothesis by looking at the linguistic characteristics of Korean Hangul, given that the Korean writing system offers a unique opportunity for comparisons of Hangul and other alphabetic writing systems due to the characteristics of being an alphabet and, at the same time, of using non-Roman script and structural autonomy of syllabic blocks. As mentioned in the previous chapters, Chapter 5 in particular, Korean Hangul is an alphabetic script, but spoken language represents a syllable largely conforming to the three-sound system of initial sound, middle sound, and final sound, literally referring to the three-sound system, for the CVC syllable which accounts for about 70% of the Korean lexicon. Due to the syllabic requisite, the onset consonant cannot stand alone for all syllables. This syllabic characteristic of spoken language seems to exert an idiosyncratic segmenting or processing unit in naming Korean words and reading Hangul. Specifically, the onset-rime primacy (i.e., tendency to segment the CVC syllable into C-VC units; e.g., hat → h + at) that has been found in European alphabetic scripts is hardly found in processing Hangul, despite being an alphabetic script. For example, Yoon, Bolger, Kwon, and Perfetti (2002) examined Korean children’s sensitivity to phonological units using a grapheme substitution task. Their findings indicated that a different phonological unit was preferred by Korean children, showing higher sensitivity to syllable body (i.e., body-coda, CV-C, segmentation) than the onset-rime structure (i.e., onset-rime, C-VC, segmentation) that has been found in English (Treiman, Fowler, Gross, Berch, & Weatherston, 1995). This is consistent with the findings of Kim’s (2008) study that shows that Korean monolingual kindergartners and first graders prefer segmenting Korean syllables into the body-coda structure. She also shows that emergent Korean readers’ awareness of the body-coda structure is a salient predictor of word decoding and spelling in Korean. She further indicates that the salient subsyllabic body-coda structure involved in reading Korean is attributable to the phonotactic features of Korean (i.e., CV co-articulation). Earlier findings of Yi’s studies (1995, 1998) are also in line with this. These findings are important because this evidence notably indicates that different levels and units of phonological awareness are involved in reading according to the linguistic properties of the given language. Considering this line of evidence, Perfetti and Liu (2005) assert that “languages themselves can be the source of the variability in reading process” (p. 199), as language can impose constraints on the level of mapping between graphemes and phonemes. This provides a foundation for the extension of linguistic relativity to script relativity.

Notably, the segmentation of the word cat into the body-coda structure in Korean results in two solid syllables, 캐 /kæ/ and 트 /tə/, with an insertion of an epenthetic vowel “으” /ə/² to the coda unit in Korean. The use of vowel epenthesis results from coarticulation of a consonant and a vowel, which is the linguistic mandate of Korean to the extent that a consonant cannot stand alone without a vowel (i.e., the complementary and consummating nature of consonants and vowels within the syllable). This combinatory rule underscores the syllabic feature of the Korean language and Hangul. This is why consonant strings cannot occur in Hangul with an exception of 13 digraph codas in the orthography (i.e., CVCC). However, each two-consonant coda has only one consonant sound value with the other being a silent (i.e., CVC in the phonology). This accords with the three-sound system of the Korean spoken language. This is different from English which allows for consonant strings from two to five consonants in a row (e.g., within, strengths) in which each consonant has a solid sound value. An extreme example in terms of the discrepancy between Korean and English is the fact that the one-syllable word strike becomes five syllables when it is transcribed into Korean, su tu ra i ku, {스트라이크} with epenthetic vowels added to each consonant and the diphthong split into two sounds. This linguistic phenomenon is an example of the notion that the properties of spoken language constrain written language (i.e., the writing system constraint hypothesis, Perfetti, 2003). This is closely related to the system accommodation hypothesis that implicates that “[r]eading makes accommodation to the language” (Perfetti & Liu, 2005, p. 199) as well as “[r]eading accommodates the writing system” (Perfetti & Liu, 2005, p. 199). This hypothesis gives rise to an important implication for script relativity, which is discussed at the end of this chapter.

2 Universality and Specificity According to Script Features

Reading is a cognitive activity. As a sophisticated cognitive function, the magic of reading begins with the automaticity of reading. Although we are never born to read (Wolf, 2007), reading becomes automatic, once it is acquired, to the degree that it is difficult to suppress. The automaticity of reading is evidenced by the Stroop effect (Stoop, 1935). The Stroop effect refers to the magnitude of the interference found in response to the incongruent word and color, compared to that in the congruent word and color. For example, when the word “’ is printed in blue ink instead of ink (i.e., conflict words and colors), naming the color of the word takes longer and is more prone to produce errors than when the ink-color of the word matches the name of the color (i.e., the word “” is printed in ink). The implication of the Stroop effect is the difficulty of suppressing our tendency to read words because we are conditioned to read text instantly, automatically, and effortlessly once a reading skill reaches the threshold of fluent reading. One example of the difficult suppression of reading is found when we talk with someone on the phone while the caption on the television is turned on; it is hard to not read the caption because we read it involuntarily. The acquired automaticity of reading is to be linked to cognitive functions.

As indicated in earlier chapters, the Chinese, Japanese, and Korean writing systems serve as a practical means to identify and compare research findings with respect to universality and specificity involved in reading, which provides evidence for script diversity and cognitive diversity. In the following section, research findings with respect to the script features, including the operating principle, psycholinguistic grain size, graph configuration, symbolic representation, graph complexity, and multi-script representation, are reviewed, and then research on second language studies in terms of cross-linguistic influences is discussed. Table 8.1. summarizes the script dimensions which are not mutually exclusive as well as attributes within each dimension.

Table 8.1. A Binary Contrast of Script Dimensions

2.1 Operating Principle (Alphabet vs. Logography)

The most widely used scripts in the world are alphabets and logographies. Alphabets are composed of letters that represent sounds rather than meaningful components of words. In contrast, logographies comprise characters representing words or morphemes as a whole or the meaningful components of words as a part. English is a representative alphabet given the greatest number of users as the first language (L1) and as a second language (L2), while Chinese characters are a representative of logographies, along with Kanji which is a Chinese-derived script primarily used for content words in Japan. Korean Hangul shares the alphabetic characteristics with English due to conforming to the alphabetic principle. The alphabetic principle refers to operating rules that graphs represent sounds rather than morphemes to the extent that the minimal sound unit corresponds to the phoneme and that multiple phonemes are combined to construct a syllable. In this regard, Hangul is closer to English than to Chinese and Japanese scripts, but uses non-Roman written signs.

Research shows that the operating principle of the writing system affects how readers process written words. Since alphabets rely on phonology for representation, readers of alphabetic scripts are more likely to rely on phonology than other constituents. This explains why phonological awareness skills are the dominant predictor of successful reading in alphabetic orthographies (Brady, 1986; Goswami, 2002; Wagner & Torgesen, 1987). Along with phonological awareness, RAN, which demonstrates the skills of phonological code retrieval and the speed of lexical access as a part of phonological processing, has also been found to be a significant predictor of Korean reading. Pae Sevcik, and Morris (2004) found that one measure of phonological awareness skills (blending words) and verbal working memory (digit span) were significant predictors of L2 Korean reading skills. When phonological awareness and RAN were controlled in the different language order (i.e., L1 vs. L2) within regression models in a subsequent study (Pae, Sevcik, & Morris, 2010), phonological awareness in English as a dominant language (equivalent to L1) was a salient predictor of sequential Korean (equivalent to L2) reading performance. However, RAN became a more robust predictor of sequentially-learned Korean reading for more skilled readers than less skilled counterparts. A more important role of naming skills than phonological awareness in reading Korean was also highlighted in another study. Cho, McBride-Chang, and Park (2008) found that speeded naming was a salient predictor of both regular and irregular Korean word reading across Korean emergent readers.

Due to the use of a logographic orthography, however, it is assumed that Chinese readers make use of semantic information in word recognition more than phonology. Research shows that the awareness of morphological structure, morphemic meaning, and homophone awareness are predictive of Chinese word reading among Chinese native readers (Chen, Hao, Geva, Zhu, & Shu, 2009; Liu & McBride-Chang, 2010; Tong & McBride-Chang, 2010). The awareness of morphological structures is a stronger predictor of Chinese word reading than homophone awareness. Specifically, Liu et al. (2013) have found that both lexical compound awareness and homophone awareness are significant predictors of character reading among 9-year-old Hong Kong children. However, when character reading at the baseline time-point is controlled, only morphological structure awareness becomes a unique predictor.

Chinese characters comprise simple characters that cannot be divided into a smaller component and compound characters that can be divided into smaller components. Compound characters include phonetic radicals and/or semantic radicals. A rule governs the position of the phonetic radical and semantic radical. For example, radicals {亻} and {扌} appear only in the left of the character, while {刂} appears only to the right (Lin, Wang, & Singh, 2018). Research shows that both children and adults rely on radicals and their positions when reading compound characters (Shu, Anderson, & Wu 2000). Readers segment compound characters into the phonetic radical to use phonological information as well as the semantic radicals and the whole characters to make use of semantic information, when decoding characters. Liu, Shu, and Xuan (2002) asked children and adults to judge whether the cue and target characters were semantically related or not. Specifically, first presented was either the cue that had the same semantic radical as the target (e.g., {始} <begin>) but was a semantically unrelated character or the cue with an unrelated character to the target in terms of the radical and meaning (e.g., {收} <receive>), and then the target {姐} <sister> that had a shared semantic radical {女} <woman> with the cue that was presented before the target. The correct answer for both types of cues was “No” in the semantic judgment test because both types of the cues were semantically unrelated to the target at the whole word level despite the shared radical. Results showed that both children and adults took longer to reject the pair of words sharing the semantic radical due to an interference effect, although the prime {始} <begin> was not morphologically transparent (summarized from Lin, Wang, & Singh, 2018). This suggests that radicals play a significant role in reading Chinese (Ding, Peng, & Taft, 2004; Liu, Shu, & Xuan, 2002; Shu, Anderson, & Wu 2000).

Other than compound characters, compound words are prevalent in the Chinese lexicon. Compound words comprise more than 75% of Chinese (Koh, Chen, & Gottardo, 2018). For example, the word {冰山} /bīngshān/ <iceberg> is composed of a syllable meaning <ice> and a syllable meaning <mountain> to have an independent word <iceberg>. The word for {computer} {电脑} /diànnǎo/ is also a compound word with a syllable meaning <electric> and another syllable meaning <brain>. The converging empirical evidence on the sensitivity to meaning in reading Chinese comes from the nature of a logography as a script in which a graph represents a morpheme rather than a sound, wherein the character <山> /shān/ represents a mountain, not the sound of the character. It is natural that due to the morphology-derived writing system, morphological awareness plays a key role in reading Chinese, as opposed to alphabetic orthographies showing phonological awareness to be important in reading. McBride-Chang et al. (2005) have found that morphological awareness plays a greater role than phonological awareness in the development of reading in Chinese.

Relatedly, research shows that native Japanese readers tend to rely heavily on visual codes and far less on phonological codes in graph processing (Mizuno & Matsui, 2013). Mizuno and Matsui (2013) further explored to identify predominant information involved in the lexical access to Kanji words for native Japanese readers. They aimed to determine whether or not the lexical access to Kanji characters was related to Japanese implausible-word processing. Two lexical decision experiments were conducted in three nonword conditions, including orthographically confusing transposed-letter nonwords, phonologically confusing pseudohomophones, and standard nonwords. The results showed that the participants were disrupted by transposed-letter nonwords, but not by pseudohomophones. This finding confirms that native Japanese speakers are more likely to rely on visual information than phonological information in the lexical access to Kanji words.

The different results of studies of Korean, Chinese, and Japanese readers indicate that the difference resulted from the operating principle of the script as being an alphabet or a logography is the source of the differences found in reading. This is related to script relativity because reading is a cognitive process that adapts to the demand of the stimulus (i.e., the script being read) for optimal information processing.

2.2 Psycholinguistic Grain Size (Phoneme vs. Syllable)

The alphabetic principle not only denotes that letters represent sounds rather than meaning, but also signifies that the minimal graphic unit corresponds to the phoneme and that a syllable is formed by combining multiple phonemes. Due to the minimal grain size of phonemes, phonemic awareness has been found to be a robust predictor of skillful reading in alphabetic orthographies. Consequently, phonemic awareness is considered a language-universal metalinguistic skill across Roman alphabetic languages and scripts (Brady, 1986; Goswami, 2002; Wagner & Torgesen, 1987). A question has arisen as to whether or not the same level of phonological awareness (i.e., phonemes, sub-syllables, or syllables) is uniformly related to efficient reading across languages. This question is valid, given that each language has its own grain size as the smallest psycholinguistic processing unit (Ziegler & Goswami, 2005).

The characteristics of being an alphabetic script and showing syllabic autonomy in the orthography of Hangul have propelled research that examines the role of consonants, vowels, and syllables in reading Hangul as a psycholinguistic grain size. Cho (2009) examined Korean kindergarteners’ development of sensitivity to consonants, vowels, and open syllables (i.e., CV syllables) and their longitudinal contributions to reading. Results showed that Korean emergent readers identified CV syllables better than individual consonants and vowels. The sensitivity to CV syllables predicted Hangul word recognition cross-sectionally and longitudinally. However, the awareness of consonants and vowels per se did not account for a significant variance in Hangul word reading when syllable awareness was controlled. Moreover, CV syllable awareness facilitated longitudinally the knowledge of consonants and vowels as well as onset and coda awareness. Cho (2009) underscores “the salient roles of syllables in the early literacy development of Korean” (p. 938). Cho (2018) also examined the extent to which orthographic, phonological, and morphological skills in the first language (L1) contribute to Korean fifth graders’ reading in Hangul, Hanja, and English as a second language (L2). She found that orthographic awareness of Hangul predicted Hangul word reading. In contrast, phonological awareness and RAN predicted reading skills in English. This within-script prediction led her to conclude that the effect of orthographic awareness would be script-specific.

In a similar vein, the dominant syllabic structure of Korean seems to make a difference in processing words in L1 and L2. Kim (2011) found that L1 syllable awareness among Korean emergent readers was positively associated with their word reading and spelling skills in L1 after print-related and phonemic awareness were controlled. Cho and McBride-Chang (2005) have also demonstrated that syllable awareness accounted for a greater variance than phonemic awareness in children’s reading in Korean as L1. Specifically, Korean native second graders’ Korean (L1) syllable deletion skills played a significant role in predicting third-grade Korean reading, whereas phonemic awareness did not account for a significant variance in third-grade Korean word recognition. In contrast, only phonemic awareness predicted English (L2) word recognition one year later (Cho & McBride-Chang, 2005).

Since the unit of Chinese and Japanese writing represents the syllable, the level of phonology involved in reading is likely to be at the syllable level. McBride-Chang and Ho (2005) found that Chinese phonological-processing skills and orthographic knowledge accounted for unique variances in Chinese character recognition among Chinese kindergartners. Of different levels of phonological awareness, syllable awareness was the most salient predictor of reading in Chinese (McBride-Chang & Ho, 2005; McBride-Chang, Bialystok, Chong, & Li, 2004; Shu, Peng, & McBride-Chang, 2008). This finding can be attributable to the fact that the minimal unit of grain size is a morpheme that corresponds to a syllable in the Chinese language. Although these researchers (Cho, 2009, 2018; Kim, 2011; McBride-Chang & Ho, 2005) did not link the results to script relativity, this line of findings aligns well with the notion of linguistic and script relativity because reading is a cognitive function such that the linguistic and scriptal attributes yield different cognitive outcomes. In short, the involvement of different grain sizes depending on the script being read is another evidence of script relativity.

2.3 Graph Configuration (Linearity vs. Block)

The graph configuration has not been directly studied, although it has the potential to provide an insight into explaining a mechanism behind reading. Not all scripts have luxury to offer opportunities for addressing this matter. Hangul is a good fit for this inquiry because it is written in blocks as an alphabetic script that can be written in railroad sequence as in English, although it is not conventional. The block format makes Hangul syllables bear structural autonomy such that syllabic boundary is unambiguous (e.g., {한글} <Hangul> rather than {ㅎ ㅏ ㄴ ㄱ ㅡ ㄹ}). In relevance to the block format, graphs are arranged in left-to-right, left-right-down, or top-to-bottom orientation within each syllable, as opposed to the left-to-right linearity in English. This visual configuration makes Hangul look closer to Chinese and Japanese than English in appearance, although the scriptal properties of Hangul point to proximity to English.

Since the mind constitutes interwoven conscious and unconscious strands, perception and behavior in relation to reading that resulted from automatic and unconscious processes are likely to be anchored in thinking and cognition. Based on this idea, a subliminal intervention called priming has been used for experiments for several decades. The use of the priming paradigm provides a further understanding of the way in which unconscious processing influences reading behavior. In priming tests, participants do not notice or recognize primes because the exposure was too brief or because participants disregarded primes in order to better pay attention to the task at hand. Pae, Bae, and Yi (2019a) examined the role of consonants and vowels in both linear format and block format using the priming paradigm in a lexical decision task. As opposed to the consonant primacy found in Roman alphabetic script (i.e., consonants play a more significant role in word recognition than vowels; Bonatti, Pena, Nespor, & Mehler, 2005; Carreiras & Price, 2008; Carreiras, Duñabeitia, & Molinaro, 2009), Pae and colleagues did not find significant priming effects for consonants compared to vowels when primes were presented in a linear way (e.g., {ㅇㄴㅅㅁ – 인삼}; – 인삼}; Experiments 1 and 2). The linear way was first presented in order to rule out the format effect on rapid word recognition as a way of determining whether individual graphs themselves exerted an effect. Since the syllable block is the standard orthographic configuration for Hangul, Experiment 3 used the standard block format to present consonant and vowel primes (e.g., { – 침술} for consonant prime-target condition; { – 불법} for vowel prime-target condition) in order to capture the natural processes with the conventional format. The consonant effect was not significant compared to that of vowel effects in the blocked experiment, either. Note that, due to the use of CVC syllables, the consonant graphemes were twice as many as those of vowels in the prime. Similar results have also been found with a naming test in another study by Pae et al. (Pae, Kim, Mano, & Wang, 2019b). These results suggest that the consonant primacy found in Roman script may not extend to Korean Hangul. Because the stimuli used in these studies are fragments of the Korean syllable block, it is difficult to relate the findings to script relativity per se. What is important is that the findings suggest a possible difference exiting between European alphabets and the Korean alphabet. This is consistent with the finding that the rime primacy found in Roman script does not apply to Korean Hangul, given a more salient effect of the body-coda structure than the onset-rime structure (Kim, 2008; Yi, 1995, 1998; Yoon, Bolger, Kwon, & Perfetti, 2002), as reviewed earlier. Taken together, these results suggest that reading Hangul may require a different mechanism than that found in Roman alphabetic scripts. These findings have implications for script relativity because the format effect (i.e., linearity vs. block) may result in different modes of information processing in reading as a cognitive activity. More research is warranted in this area.

2.4 Symbolic Representation (Arbitrariness vs. Iconic Quality)

The alphabet is a system of arbitrary symbols whose meaning has been assigned by users. In other words, letters in European alphabetic orthographies do not bear obvious resemblance to the object or concept that is signified because they are arbitrarily named and assigned. In contrast, the graphs of Korean Hangul are rather iconic, indicating that they bear pictorial representations or graphically symbolic meanings. As discussed in Chapter 5, the core Hangul consonants (i.e., {▭}, {ㄱ}, {ㄴ}, {ㅅ}, and {ㅇ}) depict the place and manner of articulation, while the core vowels (i.e., { · } , {ㅡ}, and {ㅣ}) portray the trinity of the universe <heaven>, <earth>, and <human beings>, respectively. This feature makes Hangul retain iconic qualities, by and large, such that the shapes of consonants and vowels represent the physical attribute or the resemblance of articulation organs and the universe, respectively. Chinese characters and Japanese Kanji are also iconic symbols because the written signs are in principle related to the object, concept, or meaning that is supposed to represent.

There has been little research that particularly investigates the role of symbolic representations in word recognition or reading. It is worthwhile to investigate whether the linguistic property of conicity (i.e., symbols convey meaning about what they represent) plays a role in word recognition or reading across different writing systems, especially between European alphabets and the three East-Asian scripts. This has the potential to address script relativity as well like the other linguistic features covered in this section.

2.5 Graph Complexity (Traditional Characters vs. Simplified Characters or the Number of Strokes)

The effect of character complexity (i.e., the number of strokes) on reading has been investigated. Li et al. (2019) investigated the effect of character complexity on Chinese reading and visual search using eye movement. They found that visual complexity affected fixation durations and the skipping rates of both English speakers and Chinese native readers. This finding is consistent with the findings of Chang and Perfetti’s study (2018). Chang and Perfetti (2018) calculated the number of strokes and the number of radicals of each character in both traditional and simplified characters used in their study, in which the numbers of strokes and radicals showed different distributions. They indicated that reading more complex scripts would require stronger visuo-spatial skills. The findings of McBride-Chang et al.’s study (2011) were also in line with this, which found that Hong Kong children (traditional character readers) performed better on a visuo-spatial task than Spanish-reading counterparts. These findings suggest that reading more complex scripts facilitates visuo-spatial skills, which might have been strengthened by habitual reading over time.

The effects of visual complexities and frequency of Kanji words on Kanji word recognition were also examined. Tamaoka and Kiyama (2013) investigated the effects of visual complexity on Kanji processing using simple (2–6 strokes), medium (8–12 strokes), and complex (14–20 strokes) Kanji words with high and low frequencies. The results of a lexical decision task (Experiment 1) and a naming task (Experiment 2) showed that visual complexity negatively affected the processing of low-frequency Kanji words, but not high-frequency Kanji words. These findings suggest that reading experience has a significant impact on visual discrimination performance, which is related, if tangentially, to script relativity.

2.6 Multi-Script Representations (Phonogram Kana vs. Logogram Kanji)

The Japanese multi-scripts, including morphosyllabic Kanji and phonosyllabic Kana, offer a unique opportunity to examine script effects within the writing system. Although the Japanese writing system employs both Kanji and Kana scripts, Kanji are used for the majority of content words (Yamashita, 2018). Some researchers have investigated malleable factors within the sub-script (i.e., within Kanji or within Kana including Hiragana and Katakana), while others examined between-relationships in Kanji and Kana, as reviewed below.

A handful of studies have examined within-script reading of either Kanji or Kana. Both orthographic awareness and phonological awareness seem to be important skills for reading in Japanese. Sakuma, Sasanuma, Tatsumi, and Masaki (1998) examined orthography and phonology in Kanji word reading using a semantic decision task with homophones among native adult Japanese readers, given the spelling-sound correspondence was complex in Kanji. Based on the significant homophone effects, the researchers concluded that both orthography and phonology played an important role in the judgment of Kanji words. However, the effect of phonology disappeared when the item was presented only for a brief duration. This led Sakuma et al. (1998) to conclude that orthography was the primary source of meaning extraction of Kanji words. Since Kanji was a Chinese-derived script, a similar finding to that of Chinese was not surprising.

A study of the other phonographic script, Kana, would add another piece of evidence to the understanding of orthographic and phonological codes in Japanese. Besner and Hildebrandt (1987) examined reading efficiency of Japanese Kana. They found that words written in Katakana were named faster than nonwords printed in Katakana or words printed in Katakana that were typically written in Kanji. They concluded that lexical access of words written in Katakana could bypass phonological involvement.

Koyama et al. (2008) examined the role of phonological and orthographic skills in Kana reading and writing as well as Kanji reading and writing among Japanese children. The children’s Kana reading showed a ceiling effect. The different patterns of prediction were found between Kana and Kanji. Phonological awareness predicted Kana but not Kanji skills. However, orthographic awareness and short-term memory predicted both Kanji and Kana skills. The findings reflected the difference in the scriptal properties of Kana as a phonosyllabary and Kanji as a morphosylabary. Koyama et al. (2008) attributed the lack of phonological effects on Kanji recognition to the absence of tonal characteristics (which is a part of phonology) in the Japanese language unlike the Chinese language.

In light of a difference in color processing between Kanji and Kana, Feldman and Turvey (1980) have found that words written in Kana are named more quickly than the same two- to four-word-syllable color words written in Kanji, although colors are more frequently written in Kanji. Feldman and Turvey have explained the results in terms of the pathway differences that Kana has a closer relationship to phonology than Kanji due to being a phonogram. The sound-referencing phonographic quality that is related to the phonographic principle might have allowed for faster naming. Feldman and Turvey (1980) note that the particular properties of the writing system and the specification of phonology that are intrinsic to its orthographic form are likely to facilitate Kana naming.

2.7 Linguistic Components (Orthography, Phonology, and Morphology)

Three key components involved in reading include orthography, phonology, and semantics (morphology). Since reading is about converting written signs (i.e., words and sentences) into their corresponding sounds, the interplay between orthography and phonology is inevitable. Orthographic awareness refers to the ability to use visual orthographic information to identify or recognize words. Yamashita (2018) summarizes two types of orthographic processing based on Conrad, Harris, and Williams’ (2013) conception. The first is word-specific knowledge, which involves the word’s spelling and shape. The second is general orthographic knowledge, which involves the conventional patterns of letter combinations within the script. The latter is more abstract in nature than the former. It constitutes letter sequential dependencies (i.e., letter collocations in letter arrangements), structural redundancies (i.e., letter collocations that occur in different words), and letter position frequencies (i.e., the number of occurrences in the position of letters within the word; Conrad, Harris, & Williams, 2013; Yamashita, 2018).

Due to the intersecting relationship between orthography and phonology, both orthographic and phonological skills are crucial for reading. However, their involvement in reading varies depending on the script to be read. Research has shown that phonological awareness is less important and that visual or visuo-spatial skills are more important in reading logographies than alphabetic orthographies (Huang & Henley, 1995; McBride-Chang et al., 2005). Yamashita (2018) notes that the dominant role of orthographic, phonological, and morphological skills differs depending upon various variables at hand, such as reading tasks, the reader’s proficiency level, word frequency, and familiarity of words. Orthographic awareness appears to be a key predictor of fluent reading in Japanese due to its use of Kanji, which requires visual discrimination more than reading other orthographies.

The constituent skills of orthography, phonology, and morphology are significant predictors of skillful reading in Korean. However, the level of phonology involved in reading seems to be different from that in English. Since not only is Korean spoken language syllable-based, but Hangul is also based on the consonant-vowel complementarity in the syllabic unit, syllable awareness is found to be a more salient predictor of reading Korean than phonemic awareness. Overall, phonological awareness seems to play a greater role than vocabulary in reading Korean. In a study that examined the role of speech perception, phonological awareness, and receptive vocabulary in reading and spelling among Korean-speaking first graders, compared to English-speaking counterparts, Chiappe, Glaeser, and Ferko (2007) found that speech perception and phonological awareness explained a significant variance in early literacy skills beyond oral language skills for both groups.

Another line of research regarding phonology has shown a significant relationship between suprasegmental tone sensitivity and visual word reading (Arciuli, Monaghan, & Seva, 2010; Tong, Tong, & McBride). Tone is a suprasegmental feature that is found in Chinese. A large number of Chinese characters have the same onset-rime sound but have different tones (e.g., {峰} /fēng/ <peak or hill> - {逢} /féng/ <encounter>). Research shows that tone sensitivity, along with syllable awareness, is a significant predictor of Chinese character recognition among Chinese kindergarteners (McBride-Chang et al., 2008; Shu et al., 2008). Research also shows that awareness of different lexical tones explained a unique variance in Chinese character reading, after syllable and onset awareness and morphological awareness were controlled for 6-year-old children in Hong Kong (Tong et al., 2015).

A multi-national study is also available. McBride and colleagues (2005) examined the relationships among phonological awareness, morphological awareness, vocabulary, and word recognition of second graders from China (Beijing), Hong Kong, South Korea, and the U.S. They found significant relationships between phonological awareness and morphological awareness and between these two skills and vocabulary knowledge. However, the extent to which both phonological awareness and morphological awareness were significantly related to word recognition skills was different across scripts. Specifically, phonological awareness was more crucial for reading in English and Korean than in Chinese. However, morphological awareness was more important for reading Chinese and Korean than in English. The findings suggest that both phonological and morphological awareness skills are crucial in reading Korean probably due to the universal role of phonology and the dominant morphological feature of the Korean language. Cho (2018) also found that morphological awareness accounted for a significant variance in fifth graders’ writing skills in both Hangul and Hanja in South Korea. The salient role of morphology might have resulted from the significant portion accounted for in the Korean lexicon. Regarding Korean vocabulary, the Korean lexicon is composed of three types of words, according to Kim (1993), including (1) native-Korean words, which cannot be written in Chinese characters (24.4% of Korean vocabulary), (2) Sino-Korean words, which are Chinese-derived but have Korean pronunciation (69.3%), and (3) loan words, which are mostly borrowed from English (6.3%). While Sino-Korean Hanja have a consistent syllable-to-morpheme correspondence, native-Korean words can take more than one syllable to represent a morpheme. For example, the morpheme <love> takes two syllables in the native Korean word (i.e., {사랑} /sɑ rɑŋ/, in which the individual syllable {사} or {랑} alone does not refer to <love>). In contrast, the Sino-Korean word for <love> takes only one syllable referring to one-morpheme (i.e., {愛} /æ/).

Concerning the use of morphological information in word recognition, Bae and Yi (2016) examined the effects of Hangul and Hanja primes on lexical decisions among Korean undergraduate students who were matched with their proficiency levels of Hanja reading. Hanja primes facilitated the recognition of Hanja words printed in Hangul more than Hangul primes for proficient Hanja readers. However, less proficient Hanja readers did not benefit from either Hanja or Hangul primes. What was notable was that proficient Hanja readers recognized Hangul words much faster than their less proficient counterparts. This result suggests that readers who are proficient in Hanja are likely to decode Hangul more efficiently, probably because they are able to utilize extracted morphological information in the resolution of homographic and homophonic representations (e.g., the word {사과} means <apple> and <apology> with the same orthography and phonology in Hangul but is written differently in Hanja, {沙果}, {謝過}, respectively). Likewise, the absence of priming effects for less-proficient Hanja readers may be attributable to the lack of ability to resolve ambiguous Sino-Korean homophones represented in Hangul that do not provide much morphological information.

Table 8.2. summarizes the findings of studies reviewed so far across the scripts in terms of linguistic components. Both orthographic awareness and morphological awareness play important roles in reading all scripts. The results about phonology have implications for the role playing in different scripts. Although phonology is a dominant constituent involved in reading, it is evident that the level of phonology varies according to the properties of spoken language and written language.

Table 8.2. The Role of Orthographic, Phonological, and Morphological Awareness in Reading

In summary, due to the nature of mixed-scripts in the Japanese writing system, the role of each constituent of orthography, phonology, and morphology involved in reading seems to be variable depending on the script being read and the task to perform. Kanji reading requires orthographic and morphological awareness more than phonology. However, phonological awareness is a significant predictor of Kana reading. As for Chinese, despite its character complexity, readers of Chinese make use of both phonetic and semantic radicals for word identification. Due to being a morphosyllabary, the level of phonology involved in reading is at the syllabic level. Regarding Korean, the spoken Korean syllable requires co-articulation of a consonant and a vowel, which is reflected in Hangul to the extent that consonant-vowel complementarity governs the syllable in the writing system. The findings of studies reviewed above indicate that reading a particular script is sensitive to the nature and the property of a given script being read. This accords with Perfetti’s language constraint hypothesis (2003) in that the spoken unit places constraints on the writing system. This is directly associated with script relativity because the linguistic characteristic is reflected in the script and it, in turn, puts a constraint on a cognitive function, reading. In other words, it is apparent that readers develop different processing mechanisms in response to the demands of linguistic and scriptal properties. This suggests that our cognition is shaped by habitual reading, which refers to script relativity.

3 Cross-Scriptal Influences

The linguistic phenomenon of cross-linguistic influences (i.e., linguistic knowledge or skills of one language affect(s) the acquisition or use of another language) has been of interest “since antiquity and most likely ever since language evolved” (Jarvis & Pavlenko, 2007, p. 1). Recent decades have witnessed a surge of systematic L2 studies in the light of cross-linguistic influences or cross-language transfer. Although scholars have their own preference of terms to refer to the effect of one language skill on another, the terms cross-linguistic influences and cross-language transfer are used interchangeably in this chapter because of the particular interest in addressing the carryover effect from acquired linguistic skills to another or the effect of one skill on another. Cross-linguistic influences or transfer can occur in the direction from L1 to L2 forward transfer, from L1 to L2 reverse transfer, or from L2 to L3 lateral transfer (Jarvis & Pavlenko, 2007). However, only L1-to-L2 forward transfer is considered in this chapter for two reasons. First, we have stabilized linguistic abilities in L1, which attest to the invariance of linguistic dominance across speakers. Second, L2 skills are variable across learners depending on the age of acquisition, linguistic distance between L1 and L2, and variables related to individual differences such as personality attributes, motivation (both extrinsic and intrinsic), and the goals of L2 learning. It is assumed that the cognitive processes and modes of thought flow from a more dominant skill to a less dominant one, although bidirectional fluidity is possible. In the following section, studies on cross-linguistic influences from Chinese, Japanese, and Korean as L1 to English as L2 are reviewed with respect to orthography, phonology, and morphology, and then linguistic and scriptal transfer from L1 English to L2 East-Asian scripts is reviewed under the same themes.

3.1 From L1 Chinese, Japanese, and Korean to L2 English

As L2 studies have mushroomed in reading science in the last two decades, L1 influences on L2 reading in the direction from L1 Chinese, Japanese, and Korean to L2 English have been well documented in the literature (Akamatsu, 2003; Cho & McBride-Chang, 2005; Jiang & Pae, 2020; Pae, Kwon, & Lee, 2015; Wang, Koda & Perfetti, 2003b; Wang & Koda, 2005). Beyond the findings that phonological awareness is a critical precursor to proficient reading in L1, its importance has been expanded to L2, regardless of the linguistic features and writing systems. A myriad of studies in Korean-English bilinguals have also supported this line of findings (Cho & McBride-Chang, 2005; Pae, Sevcik, & Morris, 2004, 2010; Wang, Park, & Lee and more). Wang, Park, and Lee examined the role of phonology and orthography in the biliteracy development of Korean-English bilingual children with Korean as L1 and English as L2. Results show that phonological skills in L1 and L2 are strongly correlated to each other and that L1 Korean phonological skills account for a significant variance in L2 English pseudoword reading after English phonological and orthographic skills are controlled. However, L1 orthographic skills are not a salient predictor of L2 reading. Wang, Park, and Lee note that the phonological processes are language-universal, while orthographic skills are script-specific.

Akamatsu (1999) examined the effect of L1 orthographic features on L2 English word recognition among Chinese, Japanese, and Iranian participants using a naming test with the stimuli of case-alternated words (e.g., cAsE aLtErNaTiOn). Since case-alternated words preserve spelling patterns but lose word-shape cues, Akamatsu hypothesized that if readers were sensitive to subword coding, the effect of visual disruption at the subword unit would be smaller than others. Chinese and Japanese participants with nonalphabetic L1 scripts showed larger case-alternation effects than Iranian counterparts with an alphabetic Persian L1 script. He concluded that L1 orthographic features affected the word recognition mechanism in L2 English. Similarly, Pae, Kwon, and Lee (2015) reported a different pattern of resolving visual noise (i.e., deviations from the typical orthographic form) in print, using alternated (e.g., eNgLiSh), inverse (e.g., ), and typical fonts (e.g., English), across native speakers of Chinese, Korean, and English. The performance pattern on visually manipulated fonts in the native Korean speakers was more similar to that of English-speaking participants than that of Chinese-speaking counterparts. This suggests that the relatedness in the alphabetic property plays a role in word recognition.

Ben-Yehudah, Hirshorn, Simcox, Perfetti, and Fiez (2019) also reported cross-language transfer from L1 lexical coding to L2 English word recognition among Chinese-English and Korean-English bilinguals. They used regular fonts and inverse fonts in a word naming test. Korean native speakers named words faster and more accurately, especially inverse stimuli, than Chinese counterparts, even when English proficiency was controlled. Chinese speakers were more sensitive to word frequency, while Korean speakers were more sensitive to orthographic-phonological consistency in the word. These results serve as another piece of evidence for L1 script effects on L2 word recognition.

Chinese children seem to develop phonological awareness more slowly than their English-speaking peers. Research showed that Chinese-English bilingual children outperformed monolingual Chinese children on phonemic awareness and that L2 English learning facilitated L1 Chinese phonological awareness and Pinyin skills (Bialystok, McBride-Chang, & Luk, 2005; Chen, Xu, Nguyen, Hong, & Wang, 2010). This suggests fluid cross-language transfer occurs once a threshold of L2 proficiency is reached.

Although tone is specific to Chinese among the languages under consideration, sensitivity to tone appears to facilitate L2 English word reading as well (Tong, He, & Deacon; Wang, Perfetti, & Liu, 2005). Tong et al. (2017) found that the tone awareness of Cantonese Chinese measured in grade 2 predicted L2 word reading in grade 3. Wang, Yang, and Cheng also showed that L1 Chinese tone awareness of Chinese-English bilingual children predicted a unique variance in L2 English reading above and beyond phonological awareness and reading-related variables. They have interpreted that tone awareness is general auditory processing, as tone awareness is closely related to awareness of the prosodic features of language, such as stress and rhythm, which also contributes to efficient reading.

Wang, Koda, and Perfetti (2003b) examined alphabetic and nonalphabetic L1 effects on L2 English word recognition among native speakers of Chinese and Korean. The two L1 groups showed the different patterns of reliance on phonological and orthographic information in a semantic category judgment task in L2 English. They showed using a categorization judgment task that native Chinese and Korean speakers exhibit differences in making use of phonological or orthographic information for word recognition. They asked the two groups of students to judge whether a noun (e.g., rose or feet) represents an exemplar of a certain category (e.g., flower or body part). They used homophonic words (e.g., rows for rose) with varied orthographic similarity (e.g., rows was less similar to rose than feat to feet). Results showed that native Korean speakers tended to make more errors on homophonic targets than on control items, but Chinese speakers did not show such homophonic effects. Chinese speakers were affected by orthographic similarity (e.g., feet vs. feat; rose vs. rows). The findings suggest that Korean speakers whose L1 is an alphabetic script are likely to rely on phonological information, while Chinese whose L1 is a logographic script tend to rely on orthographic information.

Morphological skills seem to be not as straightforward as phonological awareness in terms of cross-language transfer. Wang, Cheng, and Chen showed reverse transfer in that English compound awareness was a significant predictor of Chinese character reading among Chinese-English bilingual children from grades 1 through 4 in the U.S. Chinese compound awareness did not explain a significant variance in English word reading, however. Similarly, Pasquarella, Chen, Lam, Luo, and Ramirez (2011) found using structural equation modeling that English compound awareness predicted Chinese reading comprehension in Chinese-English bilingual children of first to fourth graders in Canada. However, the same effect was not found on Chinese word reading, although a bidirectional relationship between English compound awareness and Chinese vocabulary was found. Pasquarella et al. (2011) explained that the structures of the languages and writing systems of the two languages were likely to determine the direction of transfer. However, Cheung et al. (2010) showed in a study of Hong Kong children in kindergarten, grade 2, and grade 4, that Chinese compound awareness accounted for a significant variance in English word reading and that English compound awareness did not account for a significant variance in Chinese character reading. The direction of transfer might also depend on the relative proficiencies of the languages involved and the language-learning context because Chinese is a more dominant language for Hong Kong children, while English is likely to be a more dominant language for Chinese-English bilinguals in the U.S. and Canada. These findings suggest that cross-language transfer occurs from a more dominant language to a weaker language in Chinese and English. A longitudinal study shows more complicated results. Luo et al. (2014) examined both cross-sectional and longitudinal cross-linguistic relationships between morphological awareness and word reading in both Chinese and English among Chinese-English bilinguals. They found that English compound awareness predicted Chinese word reading through a mediator of Chinese compound awareness. However, longitudinal cross-language contributions were significant neither from English morphological awareness to Chinese word reading nor from Chinese morphological awareness to English word reading.

Wang, Ko, and Choi also examined the contribution of morphological awareness to Korean and English reading skills after controlling for phonological awareness among Korean-English bilingual children. Morphological awareness was measured using a derivational morphology task in both Korean and English as parallel measures. Bilingual children completed a sentence based on a prompt provided. For example, when the prompt farm was provided, the child was to complete the sentence “My uncle is a _____” (farmer). An equivalent measure was constructed in Korean to assess the comparable skills. Results showed that the morphological awareness of derivational structures in L1 uniquely accounted for a significant variance in reading real words in L2. The significant forward transfer from L1 Korean to L2 English as well as backward transfer from L2 English to L1 Korean suggest that morphological awareness facilitates word reading across alphabetic writing systems even in different scripts.

A further study on the function of morphological awareness in Korean and English was carried out with adult Korean-English readers. Ko, Wang, and Kim (2011) investigated whether adult Korean-English bilingual readers would activate the constituents of compound words of L1 while processing L2 compound words by decomposing the constituents. The results of lexical decision tasks showed that the recognition of L2 English compound words was more accurate than that of compounds translated into Korean. They concluded that morphological decomposition and cross-language activation occurred concurrently in bilingual reading of compound words.

A study of cross-language priming effects on the processing of derived words was conducted with Korean-English bilingual adults (Kim, Wang, & Ko, 2011). Strong evidence for the cross-language activation of morphological structures was observed such that L1 morphological structures were elicited to process L2 morphological stimuli. Kim and Wang (2014) conducted a follow-up study to further examine the time course involved in the cross-language activation of constituent morphemes in Korean-English bilingual readers, using a similar masked priming experiment with three prime durations (36, 48, and 72ms). Results showed that, when derived real words of Korean were used as primes, participants’ response times were significantly faster on the corresponding English L2 translated word stems at all prime durations. However, derived Korean pseudoword primes (i.e., an illegal combination of a stem and a suffix) showed a significant priming effect on English L2 word stems only at longer prime durations (48 and 72 ms). These results suggest that the cross-language activation of constituent morphemes occurs very early in bilingual reading. The lexicality factor plays an important role in the time course of decomposing L1 morphologically complex words. While an analysis of supra-lexical information is involved in early morphological processing in bilingual readers, a sub-lexical analysis is involved in the later cross-language activation of morphemic information.

Overall, the unique features of the Chinese, Japanese, and Korean scripts exert unique effects on L2 English word recognition. Along with the cross-language transfer of script-universal skills in phonological and morphological processing, the transfer of script-dependent skills in orthographic awareness has also been observed. A review of linguistic transfer from L1 English to L2 is in order.

3.2 From L1 English to L2 Chinese, Japanese, or Korean

Studies of cross-language transfer from L1 Chinese, Japanese, or Korean to English as L2 have outnumbered those from L1 English to L2 Chinese, Japanese, or Korean. A handful of studies have generally shown similar results regardless of the direction of cross-language transfer. Chikamatsu (1996) examined, focusing on phonological and visual coding involved in word recognition, the effect of L1 orthography on L2 Japanese word recognition among American and Chinese learners of Japanese by controlling for visual familiarity and word length. Participants read words written in Japanese Kana including both Hiragana and Katakana. Results showed that native Chinese speakers relied on visual information for Kana reading more than did native English speakers and that English speakers relied on phonological information more than did Chinese counterparts. These findings suggest the effect of L1 scripts on L2 word recognition. Another study by Chikamatsu (2006) with American adult students who learned Japanese as L2 showed similar results but provided more information about the developmental aspect of word recognition strategies and orthographic interferences. Results showed that L2 word recognition strategies were developmental and were reconstructed as proficiency improved, as evidenced by the results of two proficiency groups of skilled and less skilled participants. These results suggest that the use of cognitive problem-solving strategies is dependent upon literacy skills.

Similar results were found in Mori’s (1998) and Matsumoto’s (2013) studies. Specifically, Matsumoto (2013) investigated learners’ Kanji recognition to examine whether learners of Japanese as L2 would use different recognition strategies according to their L1 writing systems and whether L2 exposure would affect L2 Kanji recognition, using a lexical judgment task with three types of Kanji characters including pseudo-homophones, pseudo-homographs, and real words. Three groups of learners participated in the study: (1) beginning-level learners of Japanese whose L1 was English, (2) intermediate-level learners of Japanese whose L1 was English, and (3) beginning-level learners of Japanese whose L1 was Chinese. Results demonstrated that learners with both proficiency levels whose L1 was English showed poorer performance on the lexical judgment on L2 Kanji words, possibly due to relatively lower L2 orthographic relatedness to their L1 than that of Chinese speakers whose L1 script was similar to L2 Kanji. Matsumoto (2013) concluded that different reading strategies tended to be used by learners of different L1 backgrounds and that L2 exposure did affect L2 reading.

Lin and Collins (2012) examined the effects of L1 as well as the effects of orthographic regularity and consistency on naming Chinese characters among speakers of English and Japanese speakers who learned Chinese as L2. Their participants read Chinese characters that varied in the dimensions of regularity, consistency, the number of strokes, and familiarity (i.e., frequency appeared in instructional texts). The two groups showed a difference in reading L2 Chinese according to lexical and sublexical features, such as regularity and consistency. Lin and Collins (2012) attributed the group differences to differences in L1 phonology and the writing system. L2 proficiency also influenced performance on naming Chinese characters. However, the two L1 groups seemed to follow the trajectory of developing orthography-to-phonology knowledge (i.e., exposure to orthography occurs first and then sound is learned afterward), which was consistent with the findings with native Chinese-speaking children (Shu & Wu 2006).

Pae, Sevcik, and Morris (2010) examined reading performance of children of Korean immigrants residing in the U.S. Consistent with the performance of heritage language learners, their English skills were more dominant than Korean skills. Although Korean was the language to which they were first exposed in the home setting, English was the language they first learned to read in mainstream schools and Korean was sequentially learned to read at heritage language schools on weekends. Therefore, Pae et al. dubbed English as a dominant language and Korean as a sequential language in reading that was weaker in skills. Results showed that RAN in Korean as a sequential language (equivalent to L2) played a more dominant role than PA in the Korean within-language relationship. However, in reading English as a dominant language (equivalent to L1), phonological awareness accounted for greater variance than RAN in the within-language relationship. With respect to the cross-language associations, phonological awareness was more important than RAN skills. Again, these findings suggest that cognitive mechanisms accommodate the reader to the script specificity as a result of reading experience.

Based on the review of the literature on cross-language influences in relevance to Chinese, Japanese, and Korean, Table 8.2 summarizes significant cross-language transfer from L1 and L2 in terms of the three constituents of orthography, phonology, and morphology. Although significant cross-language transfer has been found in phonological and morphological skills, the absence of orthographic-skill transfer has been consistently observed. This suggests that phonological awareness and morphological awareness are language universal as malleable skills to transfer, while orthographic awareness is script specific due to each script’s unique visual structure and configuration (Table 8.3).

Table 8.3. Cross-Language Transfer of Orthographic, Phonological, and Morphological Skills

4 Meeting Criteria for Causality

Since script relativity builds on the capacity of written language to influence the reader’s thinking and cognition, of interest is the causal relationship from habitual reading to cognitive change. As discussed in Chapter 1, causation runs from habitual reading to the reader’s cognition, not the other way around. This is evidenced by the fact that our thinking can change as a result of reading; that is, thinking can be restructured by habitual reading. However, language does not change as a result of thinking per se, although new words are coined when necessary to meet the needs of the outcomes of new discoveries, new social movements, popular culture, and new technology.

The linguistic dimensions discussed above can be viewed as independent variables and the cognitive function as dependent variables. These relationships can also be substantiated through the prism of Hill’s (1965) nine criteria for a causal relationship. The nine criteria include strength (effect size), consistency (reproducibility), specificity (no spurious variables involved), temporality (no delays), biological gradient (exposure-incidence relationship), plausibility, coherence, experimental evidence, and analogy (similarities between the observed relationship and any other relationships). Another criterion of conditionality (if the cause disappears, the effect should disappear) is added to the batch. For example, if reading is removed, the difference between literate and illiterate people’s cognitive functions would be removed, which is solid evidence for script relativity (Bramão, et al., 2007; Petersson, Reis, Askelö, Castro-Caldas, & Ingvar, 2000; Petersson, Reis, & Ingvar, 2001). Duñabeitia, Orihuela, and Carreiras (2014) showed how literacy could change a visual mechanism of flexible position coding, which is essential for efficient reading. Table 8.4 shows the capacity to meet each criterion based on empirical evidence reviewed in this chapter. As an extension, Figure 8.2 depicts the causal relationship that runs from the scriptal dimensions to cognitive response (i.e., reading). However, it is hard to prove the causal relationship that travel from cognitive response to scriptal dimensions.

Table 8.4. Criteria for Causality

Figure 8.2.

Relationships between Scriptal Dimensions and Reading as Cognitive Responses

5 Toward the Script Relativity Hypothesis

This chapter has reviewed the role of linguistic dimensions, including the operating principle, psycholinguistic grain size, graph configuration, symbolic representation, graph complexity, and multi-graph representation, in skillful reading in Chinese, Japanese, and Korean. Also reviewed are cross-linguistic influences of linguistic skills on reading in these three scripts as L1 and L2 in relation to English. Although minor inconsistencies exist in the findings of studies depending on scripts being reviewed, tasks and measures used in studies, participant characteristics, and the foci of studies, research findings generally and consistently come together on the notion that phonological and morphological processing is script-universal, while orthographic processing is script-specific.

Reading begins with word identification, and word identities are characterized by the three interlinking constituents of orthography, phonology, and morphology. The universal features of reading center around the role of these three constituents involved in word recognition, although the degree to which they involve varies according to the script being read because reading is built upon the language system. This view points to the language constraint on writing systems that claims that writing systems encode spoken language (Perfetti, 2003). One example of spoken language constraints on reading comes from Korean speakers’ tendency of the subsyllabic segmentation into body-coda units with the insertion of an epenthetic vowel to the coda unit, which is different from English speakers’ tendency to segment syllables into onset-rime structures. This illustrates how linguistic features affect reading. The example also demonstrates that “… word reading activates phonology at the lowest level of language allowed by the writing system”, which implicates the universal phonological principle (Perfetti & Liu, 2005, p. 194).

Reading is also dependent upon the visual orthographic configuration (i.e., script). Within the constraints of reading universality, differences occur as a function of graphic codes illustrated under the graphotactic rules within the writing system. Although Chinese, Japanese, and Korean scripts share visual similarities that are closer to one another than to English, research results show differences in reading among these three scripts especially for the orthographic processing of Korean Hangul. This suggests that the functionality of orthographic features is script specific. This also suggests that script can be the source of the differences found in reading processes and behaviors.

In essence, since writing systems reflect and fulfill the linguistic demands of spoken languages, reading accommodates the demand of both spoken language and written language. Evidence of psycholinguistic research reviewed in this chapter accords with the concepts of the universal grammar of reading, comprising the linguistic constraint hypothesis and the universal phonological principle, as well as the system accommodation hypothesis (Perfetti, 2003; Perfetti & Liu, 2005). All these models and hypotheses are closely related to script relativity in that the main findings of dominant precursors of fluent reading and cross-linguistic influences point toward both unity and variability of information processing depending on script differences.

In a similar vein, Wolf (2007) asserts that readers develop different cognitive structures and mechanisms in response to the script in which they read over time. As discussed in Chapter 1, Logan (2004) also claim that alphabetic reading and Chinese character reading result in significant differences in readers’ thought patterns, reasoning, values, and problem solving. Given the differences among the four writing systems, it can be surmised that readers of Chinese, Japanese, Korean, and English demonstrate different cognitive mechanisms, which, in turn, escalate the difference between the East and the West on the global level.

Relatedly, as discussed in Chapter 6, a series of studies in social psychology have shown differences in attention and reasoning styles between Asians and Americans. According to Masuda and Nisbett (2001), Asian students tend to describe an animated underwater scene in a relational way by focusing on the background environment of the scene, whereas American students are likely to zero in on a big fish (i.e., focal object) in the background of small fish swimming around. Similarly, Ji, Zhang, and Nisbett (2004) found using word triplets (e.g., panda, monkey, and banana) that Asian bilinguals are more apt to pay attention to and describe objects focusing on relations (i.e., monkeys eat bananas) rather than categories (i.e., monkeys and pandas are both animals). In short, Asians show a tendency of understanding objects or concepts based on thematic relations. However, Americans are likely to attend to categorization or classification of objects based on the similarity of attributes or taxonomic categories. These patterns have been consistently observed in many other studies (see Chapter 6 for detail).

Nisbett (2003) attributed the differences in attention and reasoning styles between Americans and Asians to broader factors, such as differing geographies, ecologies, social structures, philosophies, and educational systems. These factors, except for philosophies, do not belong to the microsystem that is directly connected to the individual in light of Bronfenbrenner’s ecosystem, which is discussed in Chapter 7. Instead, it makes more sense to attribute the differences in attentional focus and reasoning between Americans and Asians to the cognitive differences within the microsystem—that is, reading or written language. In other words, Asian students’ relationship-grounded attention might have resulted from the spatial relationships embedded within the character of the script. Specifically, most characters of Chinese Hanzi, Japanese Kanji, and Korean Hanja not only comprise multiple subcomponents, such as semantic radicals and phonetic radicals, but also are written in top-down orientation (e.g., {氣}, {炎}), left-to-right orientation (e.g., {明}, {信}), or a mixture of both orientations within the word. Due to the flexibility in the subcomponents of semantic and phonetic radicals as well as in the orientation of the character, Asians might have developed the sensitivity to the relationships among the multiple attributes of the subcomponents in terms of the spatial characteristics of structure and orientation. Korean Hangul, despite being an alphabet, is not far from having these intrinsic characteristics of the graph. As consequences of literacy, attention to relations might have been inscribed in Asian readers’ minds. It is reasonable to assume that this acquired sensitivity through habitual reading becomes the default or template for information processing in general. If the script or habitual reading does not affect our cognition, script-specific effects and cross-scriptal influences (Akamatsu, 1999, 2003; Ben-Yehudah et al., 2019; Chikamatsu, 1996, 2008; Pae, Kwon, & Lee, 2015) would hardly be observed because reading is a very cognitive function of the mind. When the empirical findings were viewed through the prism of the causality criteria, the causal relationships from the script to thinking as cognitive responses are tenable.

It should be noted that the studies that I have reviewed in this chapter have been conducted without aiming at specifically testing script relativity. However, it is apparent that the implications of these studies, especially reading Chinese characters, and cross-scriptal influences converge on the fact that our thinking and cognition can be restructured by habitual reading. This suggests that script influences go beyond the absorptions of linguistic influence.

From methodological perspective, as Lucy (1997) pointed out, research on script relativity inherently bears challenges, just like testing the linguistic relativity hypothesis, due mostly to the interlocking relationships among spoken language, written language, and culture. However, it is still possible to tease apart the script effects from intervening or spurious variables with advanced research techniques. Since script relativity is newly proposed to overarchingly interpret research findings, more research that specifically tests this hypothesis by other scientists is needed.

With affirmative research evidence for script relativity presened in this chapter, the functions of the brain upon reading the three East-Asian scripts in relevance to reading English are reviewed in the following chapter. The next chapter also expands on the system accommodation hypothesis that is used as one of theoretical frameworks in this chapter because it supports neural structures that accommodate specific visual and structural characteristics of a given script.