Language and Cognitive Aspects of Child Bilingualism

Research Observations and Classroom Applications

by Maria Andreou (Author)
Monographs 192 Pages

Table Of Content

  • Cover
  • Title
  • Copyright
  • About the author
  • About the book
  • This eBook can be cited
  • Preface
  • Vorwort
  • Acknowledgments
  • Contents
  • 1 Introduction
  • 2 Theoretical background
  • 2.1 Some Background Information on Child Bilingualism
  • 2.2 External Factors
  • 2.2.1 Biliteracy
  • 2.2.2 Educational Setting
  • 2.3 Different Types of Bilingualism: Age of Onset
  • 2.4 Socioeconomic Status
  • 2.5 Different Types of Bilingualism: Input
  • 2.6 Language Dominance in the Bilingual Child
  • 2.7 Bilingualism and Vocabulary Skills
  • 2.8 Bilingualism and Cognition
  • 2.9 Bilingualism and Narrative Skills
  • 3 Participants’ Profile (Age, Vocabulary Capacity and Background Information)
  • 3.1 Preliminary Remarks
  • 3.2 Educational Setting
  • 3.3 Questionnaire Factors
  • 3.4 Profile of Language Abilities
  • 3.4.1 Nonverbal Intelligence
  • 3.4.2 Vocabulary Scores
  • 3.4.3 Methods for Assessing Dominance through Language Proficiency Scores
  • 3.4.4 Sentence Repetition Task
  • 3.4.5 Lexical Decision Task
  • 4 Cognition and Dominance: Educational Setting and Biliteracy Effects
  • 4.1 Cognitive Tasks
  • 4.2 Results: Balance with 6 Different Ways
  • 4.3 Results: Balance Language Proficiency (Between Group Comparisons)
  • 4.4 The Variables for a Balanced Bilingual
  • 4.5 The Biliteracy Variable
  • 5 Narration Task: Description and Coding
  • 5.1 Creation of the Four Scenarios
  • 5.2 Procedure
  • 5.2.1 Instructions
  • 5.2.2 Oral Retelling
  • 6 Narratives: Macrostructure vs. Microstructure Measures
  • 6.1 Objectives
  • 6.2 Macrostructure Measures
  • 6.3 Microstructure Measures
  • 6.4 Results
  • 6.4.1 Microstructure: Age Group 8–10 Yrs Old
  • 6.4.1 Macrostructure: Age Group 8–10 Yrs Old
  • 6.4.2 Microstructure: Age Group 10–12 Yrs Old
  • 6.4.3 Macrostructure: Age Group 10–12 Yrs Old
  • 6.4.4 The Age Effects: Micro- and Macrostructure Differences in 8–10 vs. 10–12
  • 6.4.5 Summary
  • 6.4.6 Oral Retellings, Language Abilities Tasks, Cognitive Tasks and Background Measures
  • 6.4.7 Microstructure Telling: Age Group 8–10 Yrs Old
  • 6.4.8 Macrostructure Telling: Age Group 8–10 Yrs Old
  • 6.4.9 The Mode Effect (Retelling vs. Telling): Microstructure Differences
  • 6.4.10 Oral Telling, Language Abilities Tasks, Cognitive Tasks and Background Measures
  • 6.4.11 Microstructure: Age Group 10–12 Yrs Old: Written Retellings
  • 6.4.12 Macrostructure: Age Group 10–12 Yrs Old Written Retelling
  • 6.4.13 The Effect of Mode (Retelling vs. Written Retelling): Microstructure Differences
  • 6.4.14 Written Retellings, Language Abilities Tasks, Cognitive Tasks and Background Measures
  • 6.5 Conclusions
  • 7 Practical Applications of Narratives in Language Instruction
  • 8 Discussion
  • References

←13 | 15→

1 Introduction

The aim of this research is to investigate the effects of bilingualism on verbal and nonverbal cognition, specifically how bilingualism affects language, cognitive and narrative abilities in bilingual children. Child bilingualism refers to children who are equally fluent speakers in two languages, and is a phenomenon that today applies to most countries of the world and crosscuts all social classes. Bilingual children, long considered a special group of language users as monolinguals were assumed to be the norm, have in recent years become the focus of research in cognitive psychology, linguistics and cognitive neuroscience.

A wealth of studies has investigated how the experience of being bilingual shapes our language and cognitive abilities and the way we produce narratives. In terms of their language abilities, bilingual children seem to have smaller vocabularies compared to their monolingual peers (see, for instance, Bialystok 2010), whereas their grammatical abilities may differ from those of monolingual children depending on the grammatical structure tested (Marinis and Chondrogianni 2010). In terms of their cognitive abilities, there is conflicting evidence about whether or not bilingualism leads to advantages in Executive Functions, i.e., the cognitive processes responsible for goal-oriented behavior, the capacity to think ahead, suppress impulses, and temporarily hold information (Bialystok 2011; Morales, Calvo and Bialystok 2013; Paap, Johnson and Sawi 2015; Valian 2015). Many studies have shown that systematic use of the two languages leads to a bilingual advantage in cognitive control (Adesope, Lavin, Thompson and Ungerleider 2010), but not all studies have found this bilingual advantage (see, for instance, Namazi and Thordardottir 2010). In terms of narrative production by bilinguals, a growing body of research has shown that narrative development is a lengthy process which continues well into school years and is closely related to discourse pragmatic development (Berman 2004). It seems that an important point in the use of narratives and their analyses are less biased against bilingual children than normed-referenced grammar assessment tools (Paradis, Genesee and Crago 2011).

←15 | 16→

One common denominator for studies on bilinguals is the individual variability in the participants’ performance on language, cognitive abilities, and narrative production. A key source of individual variability relates to literacy and to educational setting (biliterate vs. monoliterate) which comprise the external factors on child bilingualism (Paradis 2008). Some but not all bilingual children are also biliterate,1 and this may affect their performance in tasks measuring language, cognitive abilities, and narrative production. Also, a key source of individual variability relates to Age of Onset (AoO) which comprises an internal factor. Therefore, it is necessary to identify the contribution of external factors (i.e., biliteracy, socioeconomic conditions, etc.) and internal factors (such as Age at the Time of Testing (ATT), etc.) on language, cognitive abilities, and narrative production.

Research on biliteracy is limited compared to research on bilingualism. A small body of studies has revealed that biliterate bilinguals outperform monoliterate bilinguals and monolinguals in phonological awareness tasks and reading fluency measures (e.g., Leikin, Swartz and Share 2010; Swartz, Leikin and Share 2005). This is unsurprising because these tasks have been shown to be good predictors of reading abilities in monolingual and bilingual children (Ehri, Nunes, Willows, Schuster, Yaghoub and Shanahan 2001). However, it is unclear whether biliteracy or a balanced in terms of literacy educational system affect higher level language abilities, like grammatical abilities, and whether biliterate bilingual children have better vocabulary and grammatical abilities than monoliterate bilingual children in the two languages that they speak. With respect to socioeconomic factors, several studies have shown that the wellbeing of the family (SES) is a learning environment that affects the quantity and quality of the input which mainly has an impact on the pace of early acquisition and skill in later language use. However, other studies report that it is not the socioeconomic factor per se but a stimulating home environment rich in parental communicative interactions and learning experiences (i.e., mother-child interaction) which leads to a creation of a better input ( Hoff and Tian 2005).

←16 | 17→

In terms of their cognitive abilities and their relation to the external factors, previous studies have shown that biliterate bilingual children have higher fluid intelligence than monoliterate bilingual and monolingual children (Leikin, Swartz and Share 2010). However, it is unclear whether biliteracy also affects executive functions and cognitive control. The only study so far that has provided some evidence that biliteracy may lead to a cognitive advantage in executive functions has investigated the effects of bilingual education in executive functions (see Andreou, Marinis, Bongartz and Tsimpli under review). This study investigates executive functions in bilingual children attending bilingual educational contexts compared to bilingual children attending monolingual educational contexts and concludes that the former group has higher levels of updating skills. Updating is the capacity to monitor information entering working memory and replacing the memory representations no longer needed with those relevant to the task (Morris and Jones 1990). This provides evidence that biliteracy skills may confer an advantage in executive functions involving updating. In terms of their cognitive abilities and their relation with the internal factors, a large body of studies (e.g., Bradley and Corwyn 2002) sheds light on the role of the SES in cognitive skills, as their results with regard to the fluid intelligence present a difference of approximately one standard deviation between high and low SES groups. A further objective of our study is to see if external and internal factors can predict the cognitive abilities of bilingual children.

←17 | 18→

The narrative production consists of connected speech, which involves the entire range of linguistic as well as cognitive functions. More specifically, narrative production has been acknowledged as an effective technique to tap into structural grammatical aspects of children’s language performance, as well as into extra-linguistic processes which draw more broadly on children’s cognitive skills, world knowledge and social cognition. Children’s narrations consist of two distinct areas of discourse: microstructure and macrostructure (Liles et al. 1995). The microstructure of a narrative can be defined as linguistic structure at the lexical and syntactic level and it is used to evaluate the productivity and complexity of children’s language by calculating form and content of linguistic devices both sententially and intersententially (Halliday and Hasan 1976; Hughes, McGillivray and Schmidek 1997). On the other hand, the macrostructure of a narrative refers to its global hierarchical organization and coherence that transcends the level of the individual utterance. More specifically, macrostructure consists of story-plot elements (setting, problem, resolution) in which events are depicted (e.g., Liles et al. 1995; McCabe and Peterson 1984; McCabe and Rollins 1994). An additional objective of this research is, thus, to investigate if the external and the internal factors can predict narrative production of bilingual children along with monolingual controls. Also, a comparison between micro- and macrostructure will enable us to see if the two different domains work united or independently of one another. To understand the processes involved in speaking vs. writing of narrative production, our study will include both a speaking and a writing task.

A further significant point is to investigate the complex interaction between bilingualism and language abilities, cognitive abilities, and narrative production and to address the contribution of each of these factors to the others as outcome measures. This is a new approach, since up to this day there are very few studies which attempt to correlate the three domains (see, for instance, Marian and Kaushanskaya 2004).

To summarize, the overarching aim of this research is to investigate the effects of bilingualism on language, cognition and narrative production. The project has five objectives: 1) to observe if external and internal factors can predict the vocabulary and grammatical abilities of bilingual children and to see if there are differences between bilingual and monolingual children on vocabulary and grammatical abilities; 2) to see if external and internal factors can predict the cognitive abilities of bilingual children; 3) to investigate if the external or internal factors can predict narrative production of bilingual children and to see if bilingual and monolingual children demonstrate differences with respect to their narrative production; 4) to see which of the two ways measuring dominance (i.e., the distance between the two vocabularies or the distance between the external factors in the two languages) better explains language, cognitive and narrative performance; and 5) to investigate the complex interaction between bilingualism and language abilities, cognitive abilities, and narrative production, respectively.

←18 |

While previous research on child bilingualism has investigated aspects of grammar, vocabulary, cognitive abilities, and narratives, no studies have investigated the interaction and the complex relationship between these domains as yet. In addressing this gap, the present research adopts a new approach combining linguistic measures of vocabulary, grammar, cognitive, and narrative measures and aspires to provide critical new data on the role of bilingualism in language, cognition, and narrative production. This way, we may have a different perspective in our understanding of potential differences of the cognitive controls in bilingual compared to monolingual children, perhaps shedding light to the effects of bilingualism on the narrative production. The educational setting seems to play a crucial role in the differences of the cognitive abilities of bilinguals. In our sample there are bilinguals who attend a monoliterate educational setting and others who attend a biliterare educational setting. However, even within the latter educational setting there are internal differentiations, since some children are taught in equal amounts of time in both languages, whereas others are taught mainly in one language with fewer hours per week in the other. In other words, the degree of biliteracy differs depending on the school that the children attend.

Background information on child bilingualism is also important to be taken into consideration and to discuss multiple aspects that influence and contribute to bilingual performance, such as external factors, age of onset, socioeconomic status, linguistic input, language dominance, vocabulary, cognition and its relation to bilingualism as well as narratives and their relation to bilingualism. The current research was conducted using questionnaires and specific tasks in order to build the profile of our bilinguals that belong in different educational settings. On that basis, their categorization into different groups was made. The relationship between bilingualism and cognitive control (i.e., executive functions etc.) is also being investigated though a variety of cognitive tasks given when grouped according to the dominance resulting from external factors (i.e., input, output and educational setting) as well as their language proficiency abilities in the two languages. The extent to which balance in a higher or an intermediate level of proficiency plays a role in distinguishing the cognitive results of bilinguals is also discussed.

←19 |

Four scenarios were created for the current research that concerns the procedures of the oral retelling, oral telling and written retelling. The comparability of micro- and macrostructure measures in the four languages involved, i.e., Greek, German, English and Albanian is also discussed, as well as the procedure during the examination, along with the instructions that were given to the participants in the three narrative tasks, i.e., oral retelling, oral telling and written retelling. In addition, a comparison between oral retelling, oral telling and written retelling is conducted in order to check if there is a mode which favors the narrative abilities skills of our participants. Our research also discusses the practical applications of narratives in language instruction.

1 Biliteracy refers to the ability of bilinguals to read and write in the two languages involved.

←20 | 21→

2 Theoretical background

2.1 Some Background Information on Child Bilingualism

It has been argued in many studies (e.g., Baetens-Beardsmore 1982; McCardle and Hoff 2006a) that the term bilingualism is not well described or understood, as its definition is loosely presented. Meisel (2006) claims that a child can be considered bilingual when s/he grows with two or more languages from birth or soon afterwards. Bloomfield (1933: 55), on the other hand, chooses a narrower perspective in defining bilingualism as the ability to achieve “native like control of two languages.” Lim, Liow, Lincoln, Chan and Onslow (2008: 389) consider bilingual anyone who can interact in two languages using the oral and written form, regardless of whether performance reaches the native level or not.

However, in order for a child to be considered bilingual, there have been a number of alternative possibilities suggested, such as (see Bialystok and Barac 2012): 1) the child must use two languages throughout (childhood, adolescence, adulthood) at home and at school; 2) one language exclusively at home and to begin the second one at school age; 3) two languages at home, one at school; 4) one (or more) language at home and a different one at school; 5) one language at home and school and an additional one in adulthood.

←21 | 22→

For similar reasons, many researchers who have studied monolingual and bilingual children claim that examining bilinguals is a complicating procedure, since in bilinguals there is an individual variability over and above the individual variability of monolinguals. For instance, a typical factor that can create more confusion is the inclusion or exclusion criteria that researchers choose when they try to define who can be characterized as bilingual or not. Other variables are the different type of exposure (i.e., classroom vs. immersion) that the bilinguals might receive or the language dominance of the bilinguals which might change over time, etc. As a result of this large variation in bilinguals, it is common to get mixed results on certain cognitive and linguistic measures. This is due to a number of reasons, for instance, the “noise” that may characterize the data or other indeterminate factors that require further investigation, e.g., motivation, socioeconomic status, etc.

Studies so far claim that the bilingual children are different from their monolingual peers with respect to their linguistic and cognitive development. With regard to the linguistic aspects and more specifically in the domain of vocabulary the studies up to now present controversial results, since in many of them a bilingual disadvantage is observed (Bialystok, Luk, Peets and Yang 2010; Oller, Pearson and Cobo-Lewis 2007) but others find a similar performance for both bilinguals and monolinguals. With regard to the former view, Bialystok, Luk, Peets and Yang (2010) have conducted a study of over 1700 children between the age of 3 and 10, examining their receptive vocabulary scores in English. The results of the study indicate that monolingual children at every age exhibit higher performance than bilingual children. They report that their bilingual group of children was fluent in English. However, a difference in the frequency of the various types of vocabulary used has been detected, namely that monolingual and bilingual children had a comparable performance on words which were associated with schooling, but a lower performance was obtained for bilinguals for words associated with home. This fact leads the researchers to the conclusion that the smaller vocabulary of bilinguals of each language in relation to their monolingual peers is “somewhat complex.” In contrast, Pearson, Fernadez and Oller (1993) conducted a study based on a sample of 25 simultaneous English-Spanish bilinguals and 35 monolinguals, measuring Total Vocabulary and Conceptual Vocabulary in both languages. Their results indicate that the bilingual children develop early vocabulary on equal amounts as the monolingual peers. However, this study has some mishaps as well, one of them being the small number of participants, whereas another one has to do with the way of scoring, for instance the way of counting cognates, especially in cases that the pronunciation is not clear.

←22 | 23→

With regard to the cognitive development of bilinguals, again the results appear to be controversial, since many studies have shown that the systematic use of the two languages leads to a bilingual advantage in cognitive control (Adesope, Lavin, Thompson and Ungerleider 2010), but others find no such bilingual advantage (see, for instance, Namazi and Thordardottir 2010). In line with the former view, a large number of recent studies have revealed that bilingualism confers cognitive advantages at the level of executive functions (i.e., working memory, cf. Morales, Calvo and Bialystok 2013), inhibition (Bialystok 2011), as well as at the level of metacognitive skills such as metalinguistic awareness (Bialystok and Viswanathan 2009) and inferential skills in oral narrative comprehension (Tsimpli, Peristeri and Andreou 2016).

However, some of these findings are not always replicated (Colzato, Bajo, van den Wildenberg, Paolieri, Nieuwenhuis, La Heij and Hommel 2008; Costa, Hernández, Costa Faidella and Sebastián-Gallés 2009). According to these studies, it is possible for other factors, such as the degree of proficiency in each language or socioeconomic factors, to play an important role in supporting or altering this cognitive advantage. However, both approaches have their own limitations. More specifically, Treccani, Argyri, Sorace and Della Sala (2009) conducted a study in which they detect a bilingual advantage in inhibitory control. They used four different tasks for detecting inhibitory control (i.e., spatial negative-priming task, Simon, color negative priming and spatial cuing), but the bilingual advantage was detected only in the first task. Although the other tasks were administered to the same participants, they did not reveal any differences compared to their monolingual peers (de Bruin, Treccani and Della Sala 2014). In a similar vein, Paap and Sawi (2014) studied 58 bilingual and 62 monolingual university students using four tasks that examined executive functions controls (i.e., antisaccade, attentional network test, Simon and color shape switching) but they could not detect any bilingual advantage. They conclude that the low levels of convergent validity reveal that the majority of measures reflect “task-specific mechanisms rather than the efficacy of general functions” (p. 962). Also the studies which do not observe a bilingual advantage have similar limitations. For example, many of them do not take into consideration the factors that contribute to the “noise” of the data (e.g., environmental factors) or do not consider the possibility that in cases where there are no differences between bilingual and monolingual participants we may have to do with ceiling effects.

←23 |

2.2 External Factors

2.2.1 Biliteracy

Biliteracy refers to the ability of bilinguals to read and write in the two languages involved. To date no studies have investigated effects of biliteracy in the development of grammar, vocabulary, EFs and language processing and the complex relationship between these domains. The studies so far concern only phonological awareness tasks and reading fluency measures (Leikin, Swartz and Share 2010). The results from these studies prove that biliterate bilinguals outperform monoliterate bilinguals and monolinguals.

However, there is rich literature on literacy. Focusing on children, studies reveal that a child needs to understand the spoken language and further understand how this spoken language is represented in written form in order to be able to read and write fluently (Juel, Grifflth and Gough 1986; Perfetti 1985). Thus, it can be said that listening comprehension and decoding are necessary operations of literacy. Moving on to other studies concerning the way monolingual children develop literacy skills in different languages, it is shown that the transparency of the grapheme-phoneme conversion system has a positive effect on literacy acquisition (e.g., German children develop in literacy earlier than English children). In monolingual children, literacy development depends on oral language abilities in being associated with phonological working memory and phonological awareness.

In the area of literacy, an important subcategory is early literacy activities. As reported by studies (Sulzby 1985, Sulzby and Teale 1991), literacy practically begins from birth with activities that take place at home, e.g., book reading by mother. It has also been shown that exposure to early literacy activities in preschool age is vital for the subsequent development of the child, since it helps him in his vocabulary and language development and in demonstrating better academic performance (cf. Deckner, Adamson and Bakeman 2006).

←24 | 25→

In general, literacy is measured through decoding and reading comprehension, as well as through tasks of metalinguistic awareness (i.e., formal definitions for words or acceptability judgments for sentences). More specifically, the metalinguistic aspect of literacy includes decontextualization of language, the absence of gesture and facial expressions and the “distancing” of oneself from immediate human interaction. Literacy also consists of functional and syntactic awareness. By the term functional awareness, we refer to the knowing of the functions of print, which is related to letter discrimination ability and phonological awareness (Lomax and McGee 1987). On the other hand, syntactic awareness is crucial in understanding spoken language. Also, syntactic awareness, as measured by morphological knowledge, predicts spelling performance (Murter and Snowling 1997).

In literacy development in L2 language studies, it has been shown that phonological awareness skills are transferable across languages. Also, it has been shown that the decoding procedure is language-specific if the basis of the L1 writing system is different from the L2 writing system (e.g., alphabetic to syllabary, see Bialystok, Luk and Kwan 2005). However, we should note that L2 literacy is different from biliteracy. Another variable to be considered is that L2 literacy can be literacy in the weak language of the child (which happens to be the majority language) or in a foreign language (which is not the majority language and is introduced at school).

With regard to the relationship between L1 proficiency and L2 literacy, studies have shown that L1 use and early literacy skills in L1 can have positive effects in L2 literacy development. Another crucial point here is whether an L1 different from the language of instruction used at home plays a restrictive role (e.g., it does in standardized measures of reading for Asians but not for Hispanics).

It has been suggested that there is an interaction between cognition and reading development in L2 language. Koda (1990) claims that L2 adult readers overwhelmingly transfer reading strategies across languages. On the basis of this claim two main hypotheses are made. The first hypothesis suggests that general cognitive and linguistic development underlies literacy development in both languages of the child, whereas with the second hypothesis it is argued that language-specific knowledge (oral and orthographic) underlies literacy development in each language.

←25 | 26→

Prior to reading development and literacy, children develop knowledge about the properties of print. The transformation of knowledge of a formal system is based on the translation of visual features into a computational system of symbols. Initially, children may recognize letters and words but not necessarily as symbolic of meaning but only as visual images. Exposure to two writing systems (even if both alphabetic) may enhance the move from visual image to symbolic system. A study by Bialystok, Luk and Kwan (2005) with preliterate 4- and 5-year-old bilingual children (French-English, Chinese-English, Hebrew-English) reveals that all children could recognize letters and the sounds which corresponded to them but could not read as yet.

With respect to the literacy development in bilingualism, studies suggest that the effects of phonological awareness and functional awareness (concepts of print) can transfer from L1 to L2 (and vice versa), showing an advantage of bilingualism. However, the degree of difference in the writing system as well as the level of proficiency of the child in each language also affects transferability. From this, it is concluded that literacy levels interact with cognitive performance. This means that bilingualism and literacy interact in various ways both linguistically and cognitively.

This observation of transfer skills lies on Cummins’ (1976, 1979) interdependence theory and the concept of common underlying proficiency, suggesting that the home language and the second language are interdependent. On the basis of this study, it is assumed that once bilinguals have acquired basic literacy skills in their home language and at the same time, they have developed communicative abilities in the majority language they are capable of transferring the literacy skills acquired in the first language into the new language.

The literacy competency in two languages has been less examined in middle to late elementary school grades (e.g., Leider, Proctor, Silverman and Harring 2013; Rauch, Naumann and Jude 2011; cf. further Bialystok 1988; Bialystok, McBride-Chang and Luk 2005). Leider, Proctor, Silverman and Harring (2013) have examined language and reading comprehension as a function of biliteracy, studying Spanish-English bilingual children (grades 3–5). The results show that biliterates were better than monoliterates on English word recognition, but the two groups did not differ in English reading comprehension. The authors conclude that in this case we had a crosslinguistic transfer.

←26 |

However, when examining biliteracy, many other variables interfere such as majority language, minority language, mother-tongue language, and others, which deserve a brief discussion. Hamers and Blanc (2000) define as majority language the language that is used by a socially and culturally dominant group, and as minority language the one used by a subordinate group in a social and cultural context. There are several studies which highlight the importance of the minority language, reporting that minority L1s (home languages) are lost within a period of 2–3 years when children are immersed in L2 environment. This can mean that their receptive ability remains but expressive ability is gradually reduced to floor level. In addition, reports from developing countries suggest that 221 million children are educated in a language they do not speak at home, a fact that according to Cummins (2009) leads them to low education quality, high drop-out rates and low literacy outcomes.

The relationship between the bilingual and his/ her minority language has been described by different terms in the literature, such as “mother tongue”, “home language”, “heritage language”, etc. (cf. Skutnabb-Kangas 2000). We consider as a more effective term the use of “mother-tongue,” and, as pointed out by Olivier (2011), it is important to keep in mind that the language used by bilinguals is actually the person’s mother tongue.

Mother-tongue literacy for children attending schooling in L2 has revealed benefits in different domains such as (i) the strength of the minority language in its mental (conceptual and processing) competition with the majority language, (ii) working memory, and (iii) efficient transfer of basic and higher level literacy skills (Baker 2000; Skutnabb-Kangas 2000). Danesi (1990: 65), adding to the above advantages, reports that the literacy development in the mother tongue contributes the primary condition for the development of global language proficiency and the formation of the appropriate cognitive schemas needed in order to classify and organize experience. There are many studies pointing to the importance of supporting literacy in L1 from an early age (De Houwer 2007; Nesteruk 2010; Schwartz 2008, Tsimpli, Andreou, Kaltsa and Kapia, 2015) or other studies that show that attending Heritage Language classes or preschool bilingual education exerts a significant effect on L1 proficiency (Bylund and Dìaz 2012; Schwartz, Moin and Leikin 2012).

2.2.2 Educational Setting

←27 | 28→

Bilingualism has been investigated and analyzed from an educational perspective, and different types of bilingual education have emerged according to the purpose of the specific bilingual education provided. Several studies are conducted with regard to bilingualism in Greece. Some of them have placed emphasis on the educational perspective of bilingualism (e.g., Drettakis 2001; Maligkoudi 2009), while others have emphasized the intercultural issues involved (e.g., Paleologou and Evangelou 2003, etc.). However, to date there have been no studies on the effects of different educational frames on bilingualism and its interaction with linguistic, cognitive performance and literacy development on bilinguals.

The educational frame that our bilinguals are tapped into is the following:

Generally, when bilingual children grow older and enter educational contexts it is a common phenomenon to favor one of the two languages in question as the dominant language of instruction (Kohnert, Bates and Hernandez 1999; Kohnert and Bates 2002).

Up to now, it remains an open question whether different higher levels of biliteracy in different educational contexts or monoliterate versus biliterate educational contexts affect cognitive control and language abilities. The studies so far are very limited on this topic (e.g., Andreou et al. in press).

←28 | 29→

With respect to the educational system and its relation to cognitive controls, the bilingual context has been shown to confer cognitive advantages at the level of metacognitive ability (Cummins 1979, 2000; Bialystok 2005) but also at the level of cognitive control such as working memory and updating (Morales, Calvo and Bialystok 2013; Andreou and Tsimpli 2020; Andreou et al. 2020; 2021). In the last of these studies, the authors examined executive functions in bilingual children attending bilingual educational contexts compared to bilingual children attending monolingual educational contexts and found that the former group had higher levels of updating skills. In line with the above study, Tsimpli, Andreou, Kaltsa and Kapia (2015) examined 25 Greek-Albanian biliterate children and 25 Greek-Albanian monoliterate children measured in a task involving cognitive and morphosyntax abilities and found that biliterate children exhibit higher scores in cognitive control tasks. These findings lead to the conclusion that bilingual education for bilingual children which supports both languages (even if not to the same degree) is a necessary prerequisite for good performance on memory and cognition, which are instrumental for educational attainment and effective schooling. Elaborating on this statement and understanding the importance of the presence of a bilingual context several studies (e.g., Bialystok, Peets and Moreno 2012; Hermanto, Moreno and Bialystok 2012) have shown that even studying a second language in an immersion school program is enough to lead someone to similar benefits found for bilingual children but in a somewhat reduced form. Studies in other countries (e.g., Cummins 2000) focusing on the phenomenon of the early exit of bilingual education observe a disadvantage on children’s cognitive development.

2.3 Different Types of Bilingualism: Age of Onset

←29 | 30→

In linguistic research, types of bilingualism have been associated with the age of first exposure to each language. The age of onset of exposure to the second language gives rise to the basic distinction between simultaneous vs. successive bilinguals. The former have been exposed to the two languages from birth, while the latter were exposed to the second language some time after birth. In more recent research (Meisel 2009), however, successive bilinguals are further distinguished between early (age of onset of exposure to L2: 1–4 years) and late (age of onset of exposure to L2: 4 years and later) successive bilinguals. In this research, age 4 is considered to mark the end of the critical period for the native-like acquisition of morphosyntax. Along these lines, Paradis and Genesee (1996) conducting research in English-French bilingual children observe that the children differentiate between the two languages from early on in several domains, for instance the finite verb forms in French develop earlier than in English. Another category in which they differed was the subject pronouns in French which were used only with finite forms (in English with both finite and non-finite forms). The last category in which they differed was in the negation which was placed correctly (ne mange pas vs. do not eat).

The age of onset is an important variable, since a vast body of studies indicate that there are age effects in child bilingualism, with children exposed to their L2 around the age of 6 producing errors that are different in type from children exposed to L2 before the age of 3 (e.g., Meisel 2009). However, as Unsworth (2013) reports, differences in type of bilingualism (simultaneous, early and late successive bilingualism) have not been investigated in detail as potentially differentiating variables affecting linguistic and cognitive abilities within the group of bilingual individuals.

Rothweiler (2006) claimed a narrowest perspective, arguing that a child after the school age (i.e., age 6) is considered as L2 learner. In line with this view are the results of Granfeldt, Schlyter and Kihlsted (2007), who compared monolinguals, 2L1 and L22 children on the acquisition of various aspects of French morphology. The results of their study show that the 2L1 children show a similar performance with their monolingual peers, whereas the L2 children have a similar performance and produce similar errors with those of L2 adults. On the other hand, Meisel (2009) suggests that some windows of opportunity for certain areas of language may “close” earlier than for others. For instance, studies have revealed that vocabulary development and academic skills show better performance in child L2 learners who are not very young (i.e., over 5 years of age). Morphosyntax, on the other hand, shows an early advantage (i.e., AoO under 4 years). Other L2 learners show faster development in morphosyntax than younger ones, but younger learners were more likely to master the morphemes within five years (Jia and Fuse 2007).

←30 | 31→

Volterra and Taeschner (1978) suggest that the two languages were not differentiated from the start and, according to them, with respect to the AoO in bilingualism there are three stages of development. Stage one consists of one lexicon with L1 and L2 words, stage 2 consists of two lexicons and one grammar (L1 and L2), and stage three consists of two lexicons and two grammars.

2.4 Socioeconomic Status

It is acknowledged that socioeconomic factors, irrespectively of bilingualism, add to the social inequality affecting educational quality and learning outcomes (Pappu and Vasanta 2010).

Focusing on bilingualism and its relation with the SES, the results can be subgrouped in two domains, i.e., linguistic and cognitive skills. With respect to the SES and the relationship with the cognitive control, the results remain controversial. There are studies in favor of the view that a low SES background leads to a lower executive control (Morton and Harper 2007; Noble, Norman and Farah 2005), but there are other studies which claim that there is a bilingual advantage of cognitive control, regardless of the SES background (Calvo and Bialystok 2014; Engel de Abreu, Cruz-Santos, Tourinho, Martin and Bialystok 2012, for tasks involving inhibition and attention). Carlson and Meltzorff (2008) argue that the SES factor is not the determining indicator in the performance of bilinguals. They examined low SES Spanish-English children with high SES English controls on inhibitory control tasks and found that after controlling for the SES the bilingual children appear to have higher performance than their monolingual peers.

However, Engel de Abreu, Cruz-Santos, Tourinho, Martin and Bialystok (2012) conducted a study trying to see if there is a bilingual advantage in low SES, examining bilingual and monolingual 9-year old children on nonverbal working memory skills. Their results do not point to a bilingual advantage. However, a mishap of this study is that the parameters of country and school were not comparable, since the bilinguals and the monolinguals attended different types of schools and lived in different countries. See also Morton and Harper (2007) who claim that the bilingual advantage that they found in their study on executive control was driven by the socioeconomic differences.

←31 | 32→

In line with the above study, Morales, Calvo and Bialystok (2013) claim that perhaps bilingualism with the combination of higher levels of SES leads to cognitive advantage or that SES only compromises ability for certain levels of language experience, such as monolingual children (p. 278). In a similar approach, Calvo and Bialystok (2014) conducted a research on 25 6-year old children, who were assigned to one of four groups that differed in socioeconomic status (SES and working memory) and language background (monolingual or bilingual). The results of this study indicate that middle-class children present better performance relative to the working-class children in measures which concern executive functions, receptive vocabulary and attention based on picture naming tasks. Thus, they conclude that SES and bilingualism contribute significantly and independently to children’s performance.

Focusing on SES and its relation with language abilities (i.e., narratives), the results seem controversial as they depend on the measurements one examines. More specifically, several studies on narratives (Aksu-Koç 2005; Aksu-Koç and Küntay 2002) comparing monolingual preschoolers, primary school children and adults with low SES and high SES results reveal that the different SES does not create a different performance in terms of linguistic complexity (e.g., relative proportion of subordinate to total clauses). However, in the same studies SES was found to play a crucial role at the level of plot organization in favor of those with high SES. The results of Hart and Risley (1995) partially agree with the above studies, since they observe that oral language skills are often limited in low-income preschoolers.

In the majority of studies the SES of the children is measured by maternal education (cf. Greenberg, Lanza, and Blair 2011; Hoff, Laursen and Tardif 2002), and this seems to be the most appropriate way in contrast to other techniques that measure SES by means of financial factors.

2.5 Different Types of Bilingualism: Input

Bilingual children are exposed to half as much input as L1 children (e.g., Paradis and Genesee 1996; Unsworth 2013). However, as noticed by Unsworth (to appear), “the exact nature of the relationship between input quantity and language acquisition in a dual language setting remains largely unclear.” The variable of the home language input has been proved to affect early language development in bilingual acquisition but not always directly (cf. Paradis 2008; Thordadottir 2008).

←32 | 33→

The input or the output that the bilinguals receive varies, since there are quantity-oriented (amount of exposure, home/ school, length of exposure, etc.) and quality-oriented factors (richness of input, e.g., TV, friends, parents’ proficiency, variety of speakers, etc.). Quantity and quality of input may be hard to distinguish in some cases. More specifically, quantity measures are the length of exposure to each language at home (i.e., how many hours per language), the length of exposure to L1 or to L2 at school, weekends and holidays, the number of different speakers interacting with the child, older siblings, and literacy related activities (from preschool age). In child bilingualism, it is important to determine input quantity and type of input (home/ educational contexts) in the two languages so that a complete profile of the bilingual child is established.

For the quantity-oriented variables studies have observed an effect on the rate of acquisition for vocabulary (e.g., Golberg, Paradis and Crago 2008) and for morphosyntax, in particular for verbal morphology (e.g., for verbal morphology: Blom 2010; Paradis 2010; for wh-questions, passives and definite or indefinite articles: Chondrogianni and Marinis 2011; for grammatical gender: Montrul and Potowski 2007;). An important observation has been made according to which the quantity-oriented variables were found to interact also with other factors such as the chronological age (e.g., Goldberg, Paradis and Crago 2008), socioeconomic status (SES) (cf. Hoff 2006b for a review), and language proficiency (e.g., Chondrogianni and Marinis 2011, Paradis 2011).

With regard to the quality-oriented variables, it has been observed that there is an effect on the rate of acquisition with respect to the richness (i.e., the exposure to a variety of sources like TV, friends etc., see for instance Jia and Fuse 2007) for literacy and literacy-related activities (e.g., Scheele et al. 2010), for the variety of speakers providing input (e.g., Place and Hoff 2011), for the number of speakers exclusively speaking one language to the child (e.g., Place and Hoff 2011). The quality-oriented variables were also found to interact with other factors, such as SES (e.g., Hoff 2006b, Oller and Eilers 2002), and with school versus home variable (e.g., Cummins 1984).

←33 | 34→

2.6 Language Dominance in the Bilingual Child

For many researchers language dominance is being considered as an autonomous field which is not necessarily associated to other phenomena of early bilingualism.3 Hamers and Blanc (2000) claim that language dominance constitutes a reflection of the relationship between the competencies in the two languages of the bilinguals. According to them, being a balanced bilingual does not automatically imply a high level of competence in both languages. Montrul (2007) argues that language dominance does not capture only the notion of children’s relative proficiency in the two languages but it captures other domains as well, such as the input, for example.

Variation among bilinguals is associated with language dominance. Dominance refers to the fact that one of the two languages is more and better used by the bilingual individual. The dominant language is that in which bilinguals are informally considered to be most proficient (see Genesee, Nicoladis and Paradis 1995). Bernardini and Schlyter (2004), in describing the “other” language, introduce the distinction between the “weak” and the “strong” language in bilinguals. In general, dominance is additionally influenced by the educational frame the bilingual child is exposed to. Accordingly, bilingual speakers exposed to monolingual as opposed to bilingual educational setting show different levels of literacy development.

With regard to bilingual acquisition, it is widely accepted that very few children grow up as balanced bilinguals (i.e., they acquire simultaneously the two languages to the same extent and at the same rate, see Döpke 1992). As a result, bilinguals seem to create a dominant language which develops faster and with more solidly established intuitions than the other language (Yip and Mathews 2000). The factors that can lead to a creation of a dominant language mainly depend on the register (e.g., formal/ informal) and the communicative context (e.g., at home/ in school). Of course, dominance is a procedure that may change over time (Nicoladis and Genesee 1996).

←34 | 35→

So far studies have revealed that dominance in exposure affects morphosyntactic development (Bedore, Peña, Summers, Boerger, Resendiz, Greene, Bohman and Gillam 2012), complex syntax (Chondrogianni and Marinis 2011) and verb morphology (Paradis 2011). However, there are only a few studies that have investigated how dominance affects discourse-related domains.

A complex issue with regard to dominance is which experiential variable can best represent dominance effects in the available input (cf. Bedore, Peña, Summers, Boerger, Resendiz, Greene, Bohman and Gillam 2012 for an overview). To this effect there have been used several techniques with different measures. For instance, Bedore et al. (2012) claim that current language use of bilinguals is a more reliable predictor capable of predicting differences in morphosyntactic proficiency and semantics relative to the factor of the age of first exposure. Kohnert and Medina (2009) study the factor of dominance in school language input, investigating older bilingual children. The results of their study suggest that this factor is capable of affecting the lexical production skills, thus creating a shift point in the differential proficiency of the two languages from preschool to adolescent age.

There are many ways to detect dominance. The most common practice for measuring bilinguals’ language dominance is the creation of a balanced score based on the proficiency scores in the two languages involved. Following this technique, the vast body of studies creates the balance score by subtracting the values obtained for one-language tests from those of the other (cf. Romaine 1995 for an overview). The person whose differences are close to zero is considered balanced bilingual. Another measure which has been proposed for measuring language dominance among bilinguals is lexical richness (Malvern, Richards, Chipere and Duran 2004). The same technique is used by Treffers-Daller (2011) who investigates two groups of adult bilinguals with different language dominance profiles (25 Dutch-French from Brussels and 24 French-English from Paris). Her results show that lexical richness measures can be used to operationalize language dominance.

←35 | 36→

Following a similar approach Yip and Matthews (2006) detect language dominance in bilingual children on the basis of MLU (Mean Length of Utterances) scores in the two languages. However, as highlighted by Dopke (1998), measuring of dominance with the use of MLU has a number of limitations. More specifically, Dopke claims that MLU is not so useful for direct comparisons among languages, since there are languages which differ from one another in terms of morphological type.

There are several other ways which have been used in order to measure language dominance in bilingual children, like counting the longest utterance produced by a child in a total number of sentences, or the number of different words or verbs alone, or counting the mixing of the two lexicons and the total number of utterances in a sample (for overviews, see Cantone, Muller, Schmitz and Kupisch 2007; Yip Matthews 2006).

2.7 Bilingualism and Vocabulary Skills

It appears that bilingual children experience some disadvantages as some of their language abilities are underdeveloped, in particular their vocabulary knowledge (Bialystok 2001). To develop vocabulary, children typically need adequate exposure to sufficient amounts of language input. Reduced exposure may affect vocabulary acquisition in a negative way (Paradis and Genesee 1996). This fact is also confirmed by other studies (Mahon and Crutchley 2006; Bialystok and Feng 2011) which also detect smaller vocabulary sizes for the bilingual children as compared to monolingual age-matched controls in English. However, a serious limitation of these studies is that their results are based mainly on the “weaker” language (i.e., English) and then they compare these results to those of monolingual peers born and raised in the same country.

On the other hand, there are studies which detect no differences between bilinguals and monolinguals as far as vocabulary skills are concerned. More specifically, Lauchlan, Parisi and Fadda (2012) conducted a study of bilingual children in Scotland, revealing a significant advantage for the bilingual children relative to monolingual in vocabulary measure. Similar observations are presented by Vulchanova, Vulchanov, Sarzhanova and Eshuis (2012) who measure early lexicon development in bilingual and monolingual children of Russian and Kazakh in Kazakhstan with age ranging from 3 to 4 years old. Their results indicate that both bilingual groups perform better on receptive vocabulary tasks than age-matched monolinguals.

←36 | 37→

Bilingual children are exposed to less (at best half) language input in a language compared to monolingual children. Studies on Spanish-English bilingual children in Miami (Cobo-Lewis, Pearson, Eliers and Umbel 2002a, 2002b) and in other bilingual language contexts (Gathercole and Thomas 2009) have shown that children with more language input perform better on vocabulary measures irrespective of whether the language in question is a majority or minority language.

A cognitive function that has been found to contribute crucially to vocabulary development is the immediate phonological short-term memory and verbal working memory. In general, the lexicon of bilingual children is similar to that of monolinguals if both languages are considered (Pearson, Fernández and Oller 1993).

There are also other factors that may contribute to the creation of smaller vocabularies in bilinguals, for instance the factor of schooling, as most bilingual children attend monolingual schools (mostly in their L2). Another possible factor is the fact that bilingual children have lower literacy in one of the two languages and thus cannot reach monolingual norms.

2.8 Bilingualism and Cognition

Recent findings suggest that bilingual children exhibit higher performances on executive functions (Bialystok and Barac 2012), especially on tasks that demand strong inhibitory control (Bialystok and Viswanathan 2009). One interpretation of this is that bilinguals are trained to resist to interference, as they often need to solve linguistic competitions between the two languages and to inhibit one language while using the other (cf. Barrouillet, Bernardin and Camos 2004). This advantage also transfers to their short-term memory capacity and to working memory function (Morales, Calvo and Bialystok 2013).

←37 | 38→

Bilingual children need to learn phonemes, phonotactic rules and morphological properties of both languages as well as to process a variety of verbal information, boosting as a result their working memory capacity. Bilingual speakers need to constantly resolve the parallel activation of both languages (Colome 2001; Kroll, Bobb and Wodnieca 2006). Overtime, they develop an exquisite mechanism for cognitive control with consequences in more general domains (Bialystok, Craik and Luk 2012). The systematic use of both languages by bilingual children has been found to affect their cognitive functions. In particular, a bilingual advantage of both children and adults has been observed on nonlinguistic and cognitive processing tasks (Adesope, Lavin, Thompson and Ungerleider 2010; Bialystok, Barac, Blaye and Poulin-Dubois 2010; Bialystok, Craik and Luk 2008). In recent studies bilinguals appear to be better than monolinguals on tasks requiring updating of incoming information during task execution, as well as on tasks requiring temporary maintenance and processing of information in the mind during task performance (Bialystok 2010; Rosen and Engle 1997).

Here one question is raised, namely why bilinguals appear to be better in executive functions than monolinguals. One possible explanation is that bilinguals must process information deriving from two language systems concurrently. They have to temporarily maintain and process incoming linguistic codes in both languages until they master them. Also, even when they have mastered both languages, they are required to constantly change linguistic codes and switch from one language to the other in order to process the relevant information.

A vast body of studies also reveals that the two language systems of bilinguals are permanently activated (Francis 1999; Thierry and Wu 2007). Through this parallel activation heavy demands would pose loading cost during language processing, a fact that is translated to more practice of the executive functions operations relative to the monolinguals. It has also been suggested that to achieve this besides working memory (WM) bilinguals also depend on higher order executive skills, such as information updating (cf. Engel de Abreu and Gathercole 2012). Thus bilinguals are capable of monitoring, controlling and focusing attention on the language element needed for task execution (Morales, Calvo and Bialystok 2013), while inhibiting unwanted stimuli that come from the second language system, switching from one language code to the other as needed and concurrently updating working memory contents.

←38 | 39→

According to Morales, Calvo and Bialystok (2013), bilingual superiority derives from the continuous practice that bilingual children receive on monitoring, controlling, and focusing attention to the target language, while the non-target language is concurrently active. The activation of control processes, in order to use the target language every time, improves executive control efficiency as well as efficiency transfer to nonlinguistic and cognitive tasks (e.g., cognitive updating, cf. Bialystok 2011). At this point, according to Chrysochoou and Bablekou (2011), higher demands are necessary on young children who have not yet acquired automacity in language processing. However, Bialystok, Barac, Blaye and Poulin-Duboi (2010) evaluated children of 3 and 4.5 years either monolingual (English or French) or bilingual (English and another language) on three executive control tasks, finding the same bilingual superiority. This means that executive control superiority should not be solely attributed to the bilingualism variable, since it was also observed in very young children.

On the other hand, two contrasting accounts have been offered on the relationship between bilingualism and working memory. The first account is the need for concurrent processing in both languages which may call for increased working memory involvement. Thus, bilingualism may hinder information processing in working memory, because of the cognitive load and the memory processing demands required by two languages, instead of one (Lee, Plass and Homer 2006).

The second account refers to improved inhibitory control capacities in bilinguals which allow the inhibition of the non-target language while the target language is at use. This may result to improved working memory efficiency, exactly because capacity to inhibit irrelevant information is critical for the successful management of working memory resources (cf. Bialystok, Craik and Luk 2008; Rosen and Engle 1997).

However, until now both the underlying mechanisms via which working memory is involved in bilingual development and the relationship between working memory and bilingualism remain far from clear. In addition, findings in recent literature come from heterogeneous samples that have dissimilar educational background and socioeconomic status. Bilingualism has been shown to have beneficial effects on cognitive control (i.e., working memory and cognitive flexibility, allocation of attention sources and inhibition of inappropriate/ incorrect response biases) and on creativity4 (Leikin 2012, for children and Kharkhurin 2011, for adults).

←39 | 40→

2.9 Bilingualism and Narrative Skills

Narratives can be spoken or written and they are typically categorized in scripts, personal narratives and fictive narratives. According to Hudson and Shapiro (1991), each categorization differs on the demands of the narratives and the function that they serve. During the last twenty years, narratives have been recognized as a vital tool in children’s development. In particular, Nicolopoulou (1997: 247) says that in the case of children in preschool years mastering in narrating a story acts as a fundamental basis “for emergent literacy and long-term school success.”

Various studies show that children’s narratives are found to provide an index of their cognitive, semantic and social abilities (Liles 2003). The ability to tell stories requires an understanding of linguistic, cognitive, and social domains (Tager-Flusberg and Sullivan 1995). In addition to this and with respect to the different domains that constitute a narrative, Perkins (2007) claims that narrative analysis is part of discourse studies as well as of pragmatics, since inferencing and referencing are essential elements of narration.

Accordingly an effective narrator has to structure the story’s events in an intelligible and unambiguous way taking into consideration the listener’s needs for understanding the setting, characters and outcomes of the story (Rumpf, Kamp-Becker, Becker and Kauschke 2012). Focusing mainly on the listener’s perspective, Cummings (2009) states that the interrelation of world knowledge, language and cognitive skills with pragmatics is essential for the production of a “well-structured” narration. More specifically, he claims that “a narrative that fails to take account of listener knowledge by leaving certain information implicit and by presupposing other information will be inefficient” (2009: 23). An effective narrator is also required to consider the perspectives of the story characters in order to explain their motivations and reactions (Stein and Glenn 1979).

Narratives are instituted from two important domains, i.e., microstructure and macrostructure. The terms micro- and macrostructure were first introduced by Kintch and van Dijk (1978). In their model of text comprehension, macrostructure involves the characterization of the discourse as a whole, whereas microstructure was the local structure of individual propositions.

←40 | 41→

The term microstructure in narratives describes the sentence-level linguistic analysis (cf. Liles, Duffry, Merritt and Purcell 1995 for the frequency of grammatical utterances; Justice, Skibbe and Ezell 2006 for the use of syntactic structures). In other words, the microstructure of a narrative can be characterized as linguistic structure at the lexical and the syntactic level used to evaluate the productivity and complexity of children’s language by calculating form and content linguistic devices mainly on the sentence level rather than intersententially (Hughes, McGillivray and Schmidek 1997).

The measures that concern the linguistic forms involve analysis of children’s grammatical and syntactic abilities. To this end the majority of studies use the number of communication units and the different degrees of sentence complexity,5 such as grammatical forms, lexical forms and lexico-grammatical features (i.e., connectives, prepositional phrases, etc.). There are different ways of counting the Communication Unit, however the most common approach is to characterize as Communication Unit a main clause and its subordinate clause or clauses (Hughes, McGillivray and Schmidek 1997).

With regard to sentence complexity in the literature, one finds different ways of assessing this measure. The most prominent are by counting the number of subordinate or complex clauses (e.g., Bishop and Donlan 2005) and by calculating the average proportions of clauses in CUs (Reilly, Losh, Bellugi and Wulfeck 2004). A controversial measure of syntactic complexity is the mean length of Communication Units (MLCU). The criticism against this measure for counting syntactic complexity is based on the fact that syntactic complexity is not only the result of using longer utterances or CU length (see also Miller 1991).

←41 | 42→

According to the linguistic content a large body of studies has used measures such as type-token ratio and the number of different content and function words (e.g., Templin 1957, Miller and Klee 1995). Other measures reflecting the linguistic content both in a quantitative and a qualitative way are the number of different words and the number of different verbs. There has been a lot of debate over the last two measures, since many researchers (e.g., Schneider 1996; Muňoz, Gillam, Peña and Gulley 2003) take them as simple measures of fluency, others (Westerveld, Gillon and Miller 2004) as measures of lexical diversity (i.e., richness of a story), and still others as measures that combine both characteristics (Fey, Catts, Proctor-Williams, Tomblin and Zhang 2004). In any case, the linguistic content is a measure that detects children’s productive vocabulary skills. Thus, we may conclude that a good performance on the microstructure level rests on the participants’ skillful use of linguistic forms in functionally appropriate ways in order to achieve a coherent text expressed in a cohesive manner (Halliday and Hasan 1976). According to Berman and Slobin (1994), this level requires the knowledge of form-function relations for a participant to use linguistic means in order to connect and “package” the events in a syntactically cohesively way.

The development on the microstructure of the narratives of monolingual children is a gradually developing procedure which continues well after the age of ten (Blankenstijn and Scheper 2003). More specifically, Elbers and Van Loon-Vervoon (2000) examining the use of content words in children’s narrations observe that around the age of four children start to organize their vocabulary in semantic networks, leading to the creation of a higher variability in the use of content words. Moving on to children’s syntactic ability, a vast body of studies finds a developmental increase with age with respect to the number and proportion of complex propositions (Justice, Skibbe and Ezell 2006; Reilly, Bellugi and Wulfeck 2004). De Villiers (2003, 2005, 2009), adding to our knowledge with respect to children’s syntactic abilities, finds a connection between children’s production of complement clauses and their understanding of the mind of others. De Villiers suggests that children’s production choices with respect to that-clauses depend at least partially on their theory of mind ability.

Macrostructure is a universal knowledge about storytelling, consisting of the characters of a story, the story components and the sequencing of events. Macrostructure is commonly observed via the appearance of crucial story-plot elements (goal, attempt, outcome) and the sequential order (temporal, causal) in which events are depicted (McCabe and Rollins 1994). These components are elements of what is called “story grammar,” which constitutes the essential part of a story telling (Stein and Glenn 1979).

←42 | 43→

Tab. 1. Different Approaches of Macrostructural Analysis

Macrostructural Analyses
Horton-Ikard 2009 Cohesive adequacy
Shiro 2003 Evaluative language analysis
Ukrainetz and Gillam 2009 Expressive elaboration
Celinska 2004, McCabe, Bliss, Barra and Bennet 2008 High-point analysis

With respect to the story grammar, so far many different macrostructural analyses have been proposed, as shown in Tab. 1.

However, the most acceptable analyses contain the principles of the story grammar model (see, among others, Fiestas and Peňa 2004; Soodla and Kikas 2010). The story grammar model assumes that all stories have a setting system and an episode system. The setting system provides background information and introductory statements about the characters and the relevant context, while the episode system includes three main components that occur in all stories, namely (a) an initiating event (i.e., an external event that motivates the main characters to act), (b) internal plans (i.e., intended actions to reach a goal and solve the problem), and (c) consequences/ outcomes (i.e., success or failure in achieving a goal). According to Faulkner and Coates (2011: 15), the story grammar is an ideal way to detect “children’s developing ability to understand narrative structures.” An interesting observation is that in order to have a complete episode, the narrator must include all three of these key components. The more elaborate, sequential, thematic, and complex a story is, the better its perceived macrostructural quality is going to be (Applebee 1978; McCabe and Peterson 1984).

←43 | 44→

Furthermore, macrostructure is also characterized by coherence. Coherence exists only at the level of the macrostructure and deals with the global organization of the story in a meaningful way (Hickmann 2003; Hudson and Shapiro 1991). Up to date, the term coherence causes frustration, since many researchers tend to confuse it with the similar (in terms of functions) term cohesion. Cohesion occurs only at the microstructure level and according to Halliday and Hasan (1976: 4), “it occurs where the interpretation of some element in the discourse is dependent on that of another.” Perkins (2007: 132) tried to differentiate the two terms by characterizing coherence as a procedure that in addition demands memory and executive functions abilities (i.e., planning etc.), and by characterizing cohesion as a procedure that demands the use of explicit linguistic devices. The narrative cohesion connects the sentences by means of conjunctions, reference use, and through lexical cohesion.

Finally, the macrostructure consists of the referential cohesion, i.e., the introduction of characters, setting and time as well as the maintenance of character reference throughout the story. Many researchers believe that, due to the fact that linguistic devices are necessary in order to maintain the character reference in a story, this measure seems to belong to the domain of microstructure (e.g., Coelho 2007). Although this perspective is correct in some points, we adopt Hickmann’s (2004) approach who argues that character reference covers pragmatic knowledge as processing of verbal context and understanding listener’s perspective, and thus we place character reference in the macrostructure domain.

On the basis of the above observations, a good performance of a participant at the level of macrostructure requires a communication adequacy. In other words, a narration must be well structured and expressed by a participant in order to ensure its processability and comprehension by the audience (Johnston 2008). This level involves pragmatic capacity as well. In the same spirit, Astington and Baird (2005) add that the pragmatic knowledge also requires perspective-taking and meta-representational skills, covered by what has come to be known as “theory of mind,” as well as updating in working memory (Arnold and Griffin 2007; Johnston 2008).

The domain of macrostructure shows that the development of narrative skills is also related to the parameter of age. Berman’s (1988) research in Hebrew monolingual speaking preschoolers, school-age children and adults, using a story-generation (Frog, where are you?, cf. Mayer 1969) revealed that between the ages of 5 and 7 the participants still tended to use picture description rather than the global story sequence. A general description of the development of children’s ability to tell a story transferring from the actional-eventive stage to the episodic stage and to the global-thematic organization (following the example of Berman and Slobin 1994; Pearson and de Villiers 2006) is as follows:

We may conclude that the recent studies on bilingual children report that macrostructure is a domain that relies less on their language abilities. Illuz-Cohen and Walters (2011: 64) suggest that “story structure should be invariant across a bilingual child’s two languages and should lead readily to cross linguistic transfer, while lexical and morphosyntactic abilities should be more language-specific and less predisposed to transfer.”

←45 | 46→

The use of mental state terms is also a variable that plays a crucial role in children’s narration. Through the procedure of forming mental representations of the story’s characters the narrator has to take into consideration the listener’s perspective, in other words to involve theory of mind (Curenton and Justice 2004). Apart from the use of syntactic dependencies (e.g., that-clauses as we saw above and their connection with the theory of mind), it is possible that children may express their understanding of characters’ feelings in more indirect ways, for example through the use of internal state terms. The positive connection between theory of mind and bilingualism is established in a research conducted by Tsimpli, Peristeri and Andreou (2016). In this study, we compared typically developing (TD) monolingual and bilingual children with Specific Language Impairment (SLI) children. The results with respect to the use of emotion and mental state verbs (i.e., theory of mind related) revealed a bilingual advantage for both groups, although in the case of story presentation for the TD bilingual group the results are more salient. This finding is connected with recent studies (e.g., Bialystok 2009; Siegal, Iozzi and Surian 2009) which provide evidence for the advantage of bilingual children over monolinguals as far as their theory of mind and pragmatic abilities are concerned. Through the use of mental state terms, children may express their understanding of characters’ thoughts and feelings. Mental state verbs can be divided into different categories, as seen in Tab. 2.

Tab. 2. Categorization of Mental State Terms

Researchers Categorization
GREENHALGH and Strong 2001 Metaverbsa (i.e., metalinguistic and metacognitive verbs), perceptual (i.e., see, hear, feel, smell), physiological (i.e. thirsty, hungry, tired), consciousness (i.e., alive, asleep), emotion terms (happy, angry, worried), mental (want, think, know) and linguistic verbs (say, call, shout).
Westby 2005 Fuste-Hermann, Silliman, Bahr, Fasnacht and Federico 2006 Metalinguistic verbs (i.e., say, call etc.), metacognitive verbsb (i.e., know, think etc.), words of expressing emotions (e.g., sad, happy, angry)
Justice 2004 Emotional terms (i.e., happy, sad, feel etc.) and cognitive terms (think, remember, realize etc.)

a According to Wetsby (2005) the use of metaverbs in narrative production reflects the participants’ awareness of the goals and intentions of characters.

b Westby provides a further categorization on metaverbs into metalinguistic and metacognitive verbs. More specifically, according to this categorization the metalinguistic verbs are those which refer to acts of speaking, whereas the metacognitive verbs are those which refer to acts of thinking.

Up until now literacy has placed the mental state analysis at the level of macrostructure, since the mental state terms are connected with the goals and intentions of the characters, in other words with story grammar. However, recent studies (cf. Gagarina submitted) claim that mental state terms can refer to both the macro- and the microstructure levels of analysis. The results of the present work lead us to believe that the “linguistic verbs” (e.g., shout, say, cry, etc.) belong to the microstructure, whereas the rest of the internal state terms belong to the macrostructure analysis.

2 The age of onset for this group was between 3 and 6 years.

3 For instance, Gawlitzek-Maiwald and Tracy 1996 and Bernardini and Schlyter 2004 connect language dominance with the presence of crosslinguistic influence, whereas Nicoladis and Paradis 1995, Lenza 1992, 1997 connect language dominance with language mixing.

4 With respect to creativity Goff and Torrance (2012: 16) state that “the creative person is able to keep open and delay closure long enough to make the mental leap that makes possible original ideas.”


ISBN (Book)
Publication date
2021 (May)
Berlin, Bern, Bruxelles, New York, Oxford, Warszawa, Wien, 2021. 192 pp., 4 fig. col., 33 fig. b/w, 32 tables.

Biographical notes

Maria Andreou (Author)

<B>Maria Andreou </B>is a postdoctoral fellow at the University of Cologne, participating in the SFB Project 1252 "Prominence in Language". Her scientific interests include bilingualism and multilingualism, literacy, language and cognitive abilities in typical development and in (developmental and acquired) language disorders and narrative development.


Title: Language and Cognitive Aspects of Child Bilingualism