Tuesday, March 25, 2014

Notes on Education and Literacy in Bilingual Settings

Baker (2006) distinguishes between three major approaches to bilingual education:
  • Monolingual forms of education for bilinguals
  •  Weak forms of bilingual education for bilinguals
  •  Strong forms of bilingual education for bilingualism and biliteracy
Hamers and Blanc (2000: 321) define bilingual education to “any system of school education in which, at a given moment of time, and for a varying amount of time, simultaneously or consecutively, instruction is planned and given in at least two languages.”

Monolingual and weak forms of bilingual education are described as subtractive because the second language is developed at the expense of the first or native language, whereas additive approaches foster the development of both languages (Brisk 2006: 46).

Monolingual Forms of Bilingual Education
  •  They may receive a period of intensive majority-language tuition before entering the mainstream classes.
  • They may initially receive intensive majority-language tuition in specific classes separate from mainstream classes.
  • They may spend half a day, or some other proportion of the day, in a majority-language class and the remainder of the day in mainstream classes.
  • They may be placed in a mainstream class but with additional assistance form specialist majority-language teacher working in the classroom.
  • They may be withdrawn from the class for specific periods of time for majority-language classes.
  • They may be placed in a mainstream class with no additional support other than the general classroom teacher.
Weak form of Bilingual Education

Bilingual transitional education is another term for weak forms of bilingual education. In this model, children begin school with some language support in their first language. This generally takes the forms of school content, such as mathematics, literacy, and other subjects, which are taught in the first language, with gradual and increasing move to teaching in the second language.

Bilingual transitional programs vary in both the amount of L1 and the length of time over which it is provided. Thomas and Collier’s (2002) extensive study, which reports on outcomes in terms of academic achievement by minority-language learners in the US, includes two models of transitional education.
  • Dual-language and immersion programs achieved the best results in L1 and L2, and had the lowest dropout rate
  • Minority-language learners who initially attended remedial, segregated programs did not close the achievement gap after integration into the mainstream, and the gap was either maintained or widened in later years
  • Short-term programs of one to three years for students with no proficiency in English were inadequate; the length of time taken to reach grade-level performance was a minimum of four years
  • The strongest predictor of L2 achievement of language-minority student achievement was the amount of formal schooling in L1
  • Children who had bilingual schooling outperformed monolingually schooled students in all subjects after four to seven years of dual-language schooling
  • While minority-language students who exited into mainstream English programs initially outperformed those in bilingual programs when tested in English,  by the high school years the bilingually-schooled children outperformed the monolingually-schooled children
Strong Bilingual Education Programs

Immersion programs for majority children where children coming from the same first-language background are educated in a second majority language

Dual-language programs in which mixed groups of majority-language and minority-language children are educated in the two languages

Programs designed to maintain or revitalize indigenous languages of bilingual children where children are educated in both the majority language and the minority language with the specific aim of maintaining, or in some cases revitalizing, the minority language

What is Biliteracy?

Hornberger (2004: 156) defined biliteracy as “any and all instances in which communication occurs in two (or more) languages in or around writing.” Reiterating the same idea in a more elaborate way, Pérex and Torres-Guzmán (1996: 54) described biliteracy as the “acquisition and learning of the decoding and encoding of and around print using two linguistic and cultural systems in order to convey messages in a variety of contexts.”

Biliteracy is affected by four variables:
  • Context s of biliteracy
  • The development of biliteracy
  • Content of biliteracy
  • Media of biliteracy
Hornberger refers to this as the continua of biliteracy. In this model, bilingual’s biliteracy is described in terms of their experience along the proposed continua rather than their proficiency in a set of skills.

The Interdependence Hypothesis

Cummin’s (1979) Interdependence hypothesis proposed the CALP (Cognitive Academic Language Proficiency) is transferred from one language to another.

The Prerequisites for Reading and Writing

Bialystok et al. (2005) note three specific prerequisite critical to literacy which may have developed as a result of the different backgrounds of bilinguals. They are:
  • Oral proficiency
  • Metalinguistic awareness
  • General cognitive development
Oral proficiency has been widely reported to be associated with children’s acquisition of literacy and is the most common test used to diagnose literary development (Geva 2000). Oral vocabulary has been correlated with oral proficiency (cf. Adams 1990; Garcia 2003).

Several studies found bilinguals to be metalinguistically more aware in certain tasks. That is, bilinguals were able to analyze some levels of linguistic structure more intensely. More pertinently, some researchers have reported a temporary bilingual advantage in phonological awareness for five-year-olds that disappears by age six. Phonological awareness has been widely acknowledged as a good predictor of alphabetic reading ability.

The third skill concerns the way orthographies or script type may influence the working memory by either constraining or facilitating the transfer of information.

Learning Two Different Scripts

Kenner et al. (2004) undertook a study whether bilingual pupils might become confused when dealing with scripts. In a role-playing game, they were able to explain to their classmates differences between the English writing system and those of their languages – Mandarin, Arabic, and Spanish. 

Furthermore, Kenner (2204) reports that five-year-old bilingual children in her study looked for connections between the two scripts and exploited the similarities as well as differences.

The Effects of Different Scripts

The process of learning how to read involves many steps; the key step is breaking the code of the orthographic system. There is great variation in how languages represent the spoken word in the written form, which we will refer to as script.

The way language has conventionalized the sound-letter representation or orthography also varies. A transparent orthography has a simple one-to-one correspondence between the written representation and sound, and in an alphabetic writing system transparency is enhanced by one-to-one correspondence between letter and sound.

At the extreme end of the opaque continuum, we have languages like Mandarin, which uses a different writing system—logographs. Logographs are not based on phonological representations, and in Chinese, each character represents a unit of meaning.
There are two hypotheses about the effect of orthography on reading. One hypothesis is script-dependent, claiming that reading varies across languages, and reading development should vary with the transparency of a particular orthography. Conversely, with less transparent orthographies, children will take a longer time to master the learning of the system.

In contrast, the central processing hypothesis proposes that reading is an independent skill which is not constrained by the type of orthography. Instead, literacy acquisition is dependent on common underlying cognitive processes such as “working memory, verbal ability, naming, and phonological skills,” and deficiencies in these processes are more likely to affect literacy acquisition (Veii and Everatt 2005: 240).

Though central processing hypothesis and the Interdependence hypothesis stress a common underlying source, the central processing hypothesis is different from the Interdependence hypothesis framework, as the central processing hypothesis refers specifically to the processing of script and reading, were as the Interdependence hypothesis refers to the broader notion of general communicative skills.

Notes on Language Attrition

Language attrition refers to the process whereby an individual’s ability to speak and understand a language is reduced. The term used of loss of language at a community level is language shift.

Language Attrition and Language Shift
While language attrition (Hansen 2001) is for the most part a psycholinguistic process which takes place at an individual level, it is strongly influenced by a number of social variables. Language shift, however, occurs at the societal level and is usually a result of language contact. Across the world, large numbers of languages have been lost as a result of contact between two or more languages, particularly where one language is dominant and is considered to be a prestige language.

Types of Language Attrition

Van Els (1986) identified four types of attrition, determined by two dimensions—firstly, what is lost, and secondly, the environment in which it is lost.

Where it was lost
What is lost
First language
Second language
First-language environment
Loss of the first language as a result of ageing and/or some pathological conditions
Loss of a foreign or second language upon return to the first-language environment or through lack of contact with the second language owing to end of schooling, moving, etc.
Second-language environment
Loss of the first language as a result of emigrating to a country in which a different language is spoken; especially likely to apply to children who emigrate
Language loss late in life after emigrating to a country in which a different language is spoken (may also be related to pathological conditions)

The most common types of language attrition are the loss of a second language in a first-language environment and the loss of a first language in a second-language environment. In these situations, attrition occurs naturalistically in environments in which another language or languages are dominant (Olshtain 1989).

Hypotheses about Language Attrition

Jokobson (1941) proposed the Regression Theory. He proposed, in the context of aphasia, language attrition was the reverse process of language acquisition—what was learned, in terms of language, would be first lost in the process of attrition. Although Jakobson’s hypothesis was designed to explain language attrition in aphasics, aphasic language attrition is not progressive and the hypothesis has not been empirically supported.

The Activation Threshold hypothesis (1997) takes into account the frequency with which a linguistic term is used. As Gürel points out, this theory predicts that individual’s ability to access a particular linguistic feature is use-dependent, such that if an item is not used frequently, it will be more difficult to activate (that is, it will have a higher activation threshold).

Second-Language Loss in a First-Language Environment

There are numerous cognitive and social factors which may impact on the degree of attrition of an individual’s language. When it comes to language attrition, age appears to be clear advantage.

Olshtain (1981) examined the English attrition of younger (five to eight-year-olds) and older (eight to fourteen-year-olds) children who returned to Israel after living for a minimum of two years in an English-speaking environment. She found attrition to be more rapid for the younger group than for the older group.

Cohen (1989) investigated his own children’s loss of their language, Portuguese, when they returned to Israel at the age of nine and thirteen (their two other languages are Hebrew and English). His findings correlated with those of Olshtain in that he found the younger child’s language loss was more rapid than that of the older child.

Cohen was interested in the attrition of productive lexicon, pointing out that there are four aspects of vocabulary that can contribute toward the forgetting of a particular lexical item (Cohen 1986). These are the form of the word, the position of the word, the function of the word, and the meaning of the word.

Cohen tested his hypothesis in the study of his own children, testing them at the end of the first, third month, and ninth month after their return to Israel. He found that although they were unable to produce the same number words at nine months that had been present in one and three months, they were able to identify these words at eleven months.
Cohen concluded that there are differences in the ways in which receptive vocabulary and productive vocabulary are lost, with receptive vocabulary appearing less vulnerable to attrition that productive vocabulary.  Cohen argued that the issue is thus more related to lexical access than to actual memory loss.

Reetz-Kurashige (1999) used storytelling and retelling form a picture book and examined English verb usage. She elicited data from three groups of nine-year-old children:

  • 18 Japanese returnees who had spent an average of 2.4 years living in the US
  • A baseline Japanese group who were living in Honolulu
  • A group of native English-speaking children resident in the US
The language of the returnee children was evaluated two or three times over a 12 to 19-month period following their return to Japan. Her findings revealed attrition in the verbal system, where irregular verbs were regularized and phrasal verbs were replaced with single verbs.  She also found that more proficient speakers had lower levels of attrition. Reetz-Kurashige argues, this supports the inverse hypothesis, which proposes that the higher the subject’s proficiency, the lower the degree of attrition (1991:41).

Yoshitomo (1999) used a variety of data collection techniques in her longitudinal study of four girls returning to live in Japan after three to four years in the US. Yoshitomo had three main findings:

  • She found less attrition that she had expected.
  • She found that the girls’ abilities to combine their language sub skills reduced over time.
  • She found that language attrition appeared to take place to a greater extent when the children had few opportunities for interaction with native speakers.

Yoshitomo argued that the testing of the girls’ language made them aware of the fact that they were suffering from some language attrition, and that this acted as a motivational factor which encourages them to work on maintaining their language.

Gardner et al. (1987) examined the second-language attrition of a group of students learning French as a second language, and found little relationship between the rate of retention of language and motivation, but found that level of proficiency was important.

First-Language Loss in a Second-Language Environment

Isurin (200) documented the loss of Russian by a nine-year-old child who was adopted and moved to the US. This detailed longitudinal study focused on vocabulary loss by the child once she no longer had access to her native language while having extensive exposure to English. 

Three different types of picture-naming tasks were used. In one task, the child had to name the picture in Russian and then in English; in another, the language used for picture naming was not specified; and in third task, picture naming was in Russian, but the stimuli were labeled in English. Reaction times were also measured.

The study found that nouns tended to be lost and acquired more rapidly than verbs, which underwent a slower transition, and that, in particular, high-frequency words, cognates, and semantically convergent pairs were the most vulnerable group.

Anderson (2001) followed two L1 Spanish-speaking siblings over a two-year period, beginning approximately two years after they had moved with their parents to the United States.

At the time of the move, the children were 1;6 and 3;6; they had lived for several years in the United States at the time of the study and were both attending elementary school. They were recorded interacting with their mother, in Spanish, for about 30 minutes every one to two months, for a period of 22 months. The focus of the analysis was on the verbal inflectional morphology—Spanish verbs inflect for person, number, mood, tense, and aspect, usually as suffixes.

Differences were found in the attrition rates of the two children, with the younger child’s languages appearing to be more susceptible. The younger child also tended to use more English in the recorded sessions, and appeared reluctant to use Spanish while other child expressed a desire to develop her Spanish skills.

It is important to note that although children may lose productive skills in their first language, the maintenance of a passive knowledge of the language even if they do not speak it, does appear to correlate positively with first-language recovery (Uribe de Kellet 2002).

Notes on Bilingualism and Cognitive Ability

A common assumption is that speaking too many languages leads to speakers confusing one language with another, leading to the inability to learn any one language successfully. Another assumption is that it is impossible to be good at two languages, and that one is likely to suffer.

As Grosjean (1982) pointed out, while we never doubt that the study of mathematics or the pursuit of music is good for our general development, the learning or use of an additional language seems to attract closer scrutiny.

There is no research to indicate that bilingualism will adversely affect stuttering or that a second language will be detrimental to those who are hearing impaired. Furthermore, there is no evidence to suggest that bilingualism is negatively associated with language impairment.

The Early Years

It was in the mid-nineteenth century that possible harmful effects of speaking a second language were formally expressed. Humboldt (1767-1835) argued that it’s only thought monolingualism that we can preserve the essence of each individual language.

A whole series of studies confirmed the inferiority of bilinguals in both verbal and non-verbal tasks. In some studies, they were found to perform poorly only on tests of non-verbal intelligence. These studies spanned four continents and involved 26 different cultural groups. Though more than 50 percent of the studies indicated weak levels of language competence in English for the bilingual population, English was the medium used for testing.

Despite methodological flaws, these studies had considerable influence, and by the middle of the twentieth century the opinion that bilingualism is detrimental to cognitive functioning was firmly established.

The studies between 1922 and 1943 used an assortment of intelligence quotient (IQ) tests as standard tools to measure the cognitive abilities of bilinguals. Studies which factored the language component indicated that bilinguals were not inferior to monolinguals in performance.

Another issue of concern is whether the various IQ tests used were valid measures of intelligence. It is widely recognized that such tests do not measure innate abilities; instead, they contain cultural bias which may place children at a distinct advantage.

The most severe criticism of studies from this early period concerns the inadequate definition of bilingualism and the confounding of SES and ethnicity variables.  In a detailed analysis of these studies, McCarthy (1930) reported that more than half of the bilingual schoolchildren came from working-class backgrounds, while the monolinguals were from educated middle-class homes. Furthermore, how bilingual proficiency was measured was problematic.

Bilingualism Enhances Cognitive Functioning

In the first documented case study of a bilingual, Leopold (1949a) cited his bilingual daughter Hildegarde’s metalinguistic awareness as evidence of the enhancing effects of bilingualism. He noticed that she was precociously aware of rhymes and would deliberately destroy rhymes in word play. Leopold argued that bilinguals were able to detach sound from meaning because of the constant early exposure to two languages. So, from a very young age a bilingual child is constantly aware of two competing forms for one meaning.

Peal and Lambert (1962) strictly controlled the SES and language background of the 364 bilingual and monolingual participants in their Canadian study. They screened their sample with great care, matching the children on SES, sex, age, language, intelligence, and attitude. Their study found that once SES and language competence variables were controlled, bilinguals outperformed monolinguals in IQ tests. Moreover, bilinguals were also found to have more positive attitudes towards French-speaking communities than their (English or French) monolingual counterparts had.

Another interesting question arose as bilinguals in Peal and Lambert’s study were found to be not categorically better on all measures of nonverbal tasks. Such task can be divided into two subgroups, one requiring spatial and perceptual processes and the other requiring symbolic manipulation. In spatial and perceptual processes, the emphasis is on spatial acuity and perceptual speed, whereas symbol manipulation tasks require understanding of abstract relations, concepts, and factual information.

In Peal and Lambert’s study, bilinguals were found to be better in the symbolic manipulation types of non-verbal tasks but performed the same as monolinguals in the non-verbal tasks requiring spatial and perceptual processes. Pearl and Lambert called this ability “mental or cognitive flexibility” and proposed that bilinguals’ early awareness of two different codes, and their ability to associate tow words with one object, may have enhanced the development of an increased cognitive flexibility.

Bilingualism and Intelligence

Pearl and Lambert’s own exploratory analysis seemed to indicate that some level of intelligence is essential for bilingualism. However, this does not rule out the fact that bilingualism may in some ways be positive influence on non-verbal abilities.

Hakuta and Diaz (1985) conducted longitudinal study which tracked the relationship between cognition and bilingual proficiency. By assessing both the language proficiency and the cognitive level of bilinguals over time, Hakuta and Diaz presented findings which conclusively supported the hypothesis that it is bilingual proficiency that exerts an influence on cognitive functioning and not the other way round. Hakuta (1987) reanalyzed the longitudinal data and argued that, while the degree of bilingualism was a better predictor of cognitive ability, the reverse was also true, although the effect was not quite as strong. He also reported another set of results, which indicated that there was no relationship between increased bilingual proficiency and the metalinguistic awareness for the bilingual sample, who were drawn mainly from lower socio-economic backgrounds. In this discussion, Hakuta argued that the context in which bilingual function must be taken into account.

Current Views on the Effects of Bilingualism

Since 1965, a stream of papers have highlighted the positive effects of bilingualism and marked a change in research focus. Instead of making a general search for IQ superiority, the new generation of researchers has been far more specific in their enquiry.

Cognitive Flexibility

Loosely defined, cognitive flexibility is used to mean creativity or ability use divergent thinking, such as the ability to generate multiple associations from one concept, or the ability to mentally reorganize the elements of a problem or situation.

Using the an Embedded Figure Test—a specific task for measuring cognitive flexibility which involve detecting simple objects embedded in larger, more elaborate figures such as the search for multiple hidden objects in the Where’s Waldo? picture book—bilingual children were found to perform significantly better than monolingual children (Balkan 1970; Cummins and Gulutsan 1973).

Metalinguistic Awareness

Broadly defined, metalinguistic awareness is the ability to focus on different levels of linguistic structures such as words, phonemes, and syntax. This ability to analyze language more intensely has been the subject of many studies. Metalinguistic awareness and bilingualism is burgeoning areas of study. Generally, these studies have targeted the following aspects of linguistic structure: word awareness, phonoligical awareness, sentence awareness, and semantic awareness

Word Awareness

There are two important interpretations of word awareness: the ability to recognize that the speech stream is composed of discrete units called words, and the awareness that the relationship between words and their meaning is arbitrary.

The Awareness of Word as a Discrete Unit

In a series of studies, Bialystok (1986a, 1986b, 1987a, 1987b, 1988) compared the performance of Grade 1 English monolingual and French-English bilingual children on tasks such as sentence segmentation (word count) and word judgments. The children were read sentences which were either intact or scrambled and asked to count the number of words in each utterance. The number of syllables and morphemes were also manipulated to further examine the child’s concept of a word. In these studies, bilinguals consistently outperformed monolinguals except in the cases of sentences consisting entirely of monosyllabic words.

The Awareness of the Arbitrariness of Language

The very basis of metalinguistic awareness requires the understanding of language is essentially symbolic and that the relationship between form and function is completely arbitrary. In studies Ianco-Worrall (1972) and Ben-Zeev (1977), children were asked to play a game involving symbolic manipulation. For example, they were asked if it was all right to switch names for ‘turtle’ and ‘plane’. They were then asked to play a game where they had to say turtle when they meant plane.

In other tasks, the substitution involved violating grammatical rules of the language. For example, children were told to replace ‘the boy’ with ‘pencils’ when given the sentence ‘The boy sings loudly’.

Here, bilingual children were first pre-tested  with Berko’s (1958) Wug Test (The Wug Test is a morphological test which tries to find out if children are able to generalize simple morphological rules such  as rules forming the past tense, using novel words). Only children who scored 70 percent and above for the Wug Test were selected to ensure that the violation of grammatical competence in the test language were voluntary and not due to the children’s lack of grammatical competence.

Bilinguals are expected to do better in such tasks as they are more aware of the arbitrariness of language. Hence, they are more likely to reject the notion that one object can only have one name. However, results varied from one study to another.

In a slightly different task with Irish-English bilinguals, Cummins (1978) asked children drawn from Grades 3 and 6 of middle-class Dublin schools to justify their responses to questions. The children were asked, “Suppose you were making up names for things, could you then call the sun ‘the moon’ and the moon ‘the sun’?” ad were then required to justify their response. The children’s justification fell into three categories:

Empirical justifications
For example, ‘the names could be interchanged because both the sun and moon shine.’

Rigid conventional justifications
For example, ‘they are their right names so you couldn’t change them.’

Arbitrary assignment responses
For example, ‘you could change the names because it doesn’t matter what things are called.’

The findings indicate that bilinguals were significantly more likely to respond with an appeal to the arbitrariness of language and that this tendency was even more pronounced among the children in Grade 6.

Phonological awareness

Phonological awareness is the ability to recognize that speech is composed of distinct unites of sound. In phonological awareness tasks the children are required to isolate relevant phonological segments as the basis of their analysis. For example, they may e asked to identify the ‘odd one out’, be asked to provide minimal pairs, or supply rhyming  words.

Using an assortment of tasks, Davine et al. (1971) found that bilingual children in elementary school in Canada performed better than monolingual children in distinguishing phonological differences. In a study made by Bruck and Genesee (1995), French-English bilinguals were given a battery of phonological awareness tests in both kindergarten and first grade. The findings indicated that bilingual children were better in certain phonological tasks such as syllable counting , while monolinguals were better at phoneme counting.

Sentence Awareness

Sentence awareness is the ability to recognize utterances which are grammatically acceptable within the language. In sentence awareness tasks, the children are often asked to detect, correct, and explain errors. Studies on monolingual children indicate that they have difficulty noting and correcting errors of this kind before the age of 5 to 6. However, Galambos and Goldin-Meadow ((1983) reported that Spanish-English bilingual children who proficient in both languages picked out such errors easily as young as 4 years of age. While young monolingual children focused on the message conveyed, bilingual children readily focused on the structure. In subsequent studies, the findings revealed that bilinguals were better than monolingual at noticing sentence errors but showed no advantages when they had to explain the errors.

Semantic Awareness

The Semantic-Phonetic Preference tasks were used in several studies, and bilingual and monolingual children’s ability to focus on semanitc similarity rather than phonetic similarity was compared. The bilinguals in Ianco-Wrrall (1972) and Ben-Zeev’s (1977) studies were found to give more semantic responses than phonetic responses. Ianco-Worall argued that this preference shows that the child is more able to focus on the meaning and is not tied to the form of the words.

Other studies of semantic awareness look at the children’s ability to form semantic hierarchy and organize objects into superordinate or subordinate categories. There is some evidence (Cummins 1978) to show that bilinguals may have some advantage in this kind of mental organization.

Threshold Hypothesis

Cummins attempted to resolve inconsistencies in this area by proposing that lower levels of proficiency attained could explain in the lack of advantage found for some bilingual populations. In his hypothesis, 

Cummins proposed two thresholds of language competence. He argued that to avoid negative effects form bilingualism, the lower threshold must be attained. The cognitive growth of children who fail to reach the first threshold will be adversely affected. Those who attain the second threshold will enjoy the enhancing effects of bilingualism. 

The main problem with the Threshold theory is the difficulty in establishing these thresholds in concrete terms. Given the opaqueness of the concept, the Threshold theory cannot be clearly defined or empirically tested.

Analysis and Control Hypothesis

The central thesis in Bialystok’s (2001a) proposal is that the enhanced metalinguisitc awareness effect operates differently for different linguistic structures. She argues that bilingualism does not have a general effect on a domain of knowledge such as metalinguistic awareness.  Rather, the effect is on the underlying cognitive processes that are activated in different tasks.

By focusing o the different task demands, Bialystok (1998) proposed a model whereby bilinguals , irrespective of their degree of bilingualism, may be more advantaged at tasks which place great demands on the control of linguistic processing. However, bilinguals who had attained high levels of proficiency in both languages are also advantaged at tasks that require more analyzed linguistic knowledge. Bialystok identified tow cognitive process, control of attention and analysis of representational structure, which she argues are able to explain the demands in the task. In tasks requiring higher levels of analysis, participants use their knowledge to work out relationships between concepts and ideas, while tasks requiring higher levels of control, participants are required to attend to some features while ignoring or inhibiting their responses to other distracting features.