1
094102309
Acquisition of English morpho-syntax and the development of reading by immigrant adults
By
Kate Moss
Abstract
Migrating to a different country is a reality faced by approximately 214 million people worldwide, many of whom are illiterate and learn the language of the host country. Despite this, the process by which illiterate adults become L2 speakers is largely ignored by SLA researchers. Research presented here concerns particularly the relationship between literacy and the development of L2 morph-syntax. Data was collected from four low-literate male L1 Kurdish speakers relating to their literacy level, phonemic awareness, and competence in English morpho-syntax. Literacy level was found to positively correlate with both phonemic awareness and morpho-syntactic competence, though the direction of causation remains unclear. However, this case-study provides qualitative analysis of the circumstances in which the participants learn both English and literacy, which offers further insight into the relationship between the seemingly parallel development of L2 morpho-syntax and literacy.
Acquisition of English morpho-syntax and the development of reading by immigrant adults
1. Introduction
Migrating to a different country is a reality faced by approximately 214 million people worldwide, for political, economic or social reasons (International Organisation for Migration 2011), and a substantial proportion of these migrants are known to have low levels of formal education and literacy (UNESCO 2004). Low-literate migrants often encounter problems with social and economic integration when settling in highly literate societies, which strains the social services and diminishes economic output (Cameron and Cameron 2005). However, learning to communicate with members of the host community can alleviate these problems considerably (Dustmann and Van Soest 2002), therefore learning literacy, as well as the language of the host community, is important for social and cultural integration (Dustmann and Van Soest 2002). Unfortunately, second language acquisition (SLA) research investigates predominantly the process by which highly literate people learn a second language (L2) (Comings et al. 2003), and the vast majority of literacy research concerns how children learn to read in their first language (L1) (Comings et al. 2003; Thompkins and Binder 2003). Illiterate adult L2 and literacy learning is largely ignored (Tarone and Bigelow 2005), though this bias in research has started to be addressed in some European countries, the USA and Sweden (Van de Craats et al. 2006). This study aims to investigate the effect literacy, as a general cognitive mechanism, has on the process of SLA, which as Bigelow et al. (2006) note is relatively understudied. Research so far indicates vast differences between literates and illiterates in terms of L1 and L2 language and linguistic capabilities and seems to be a key factor in SLA (Vainikka and Young-Scholten 2007).
Generative linguists, however, do not believe that general cognitive mechanisms influence SLA, rather SLA is a product of an innate language-learning module inside the brain. Chomsky (1957; 1981) proposed the idea that all children are born with an innate language learning mechanism, termed Universal Grammar (UG). It consists of a set of principles and parameters that guide first language (L1) acquisition (Chomsky 1986). An innate predisposition to learn a natural human language allows for rapid and complete learning of a language (Pinker 1994) and accounts for the asymmetrical development of language and other cognitive skills (Curtiss 1982; Curtiss and Yamada 1981). Children learning a first language go through similar stages of syntactic and morphemic acquisition, regardless of the language they are learning (Klima and Bellugi 1966; Slobin 1970). For instance, Brown (1973) investigated the acquisition of morpho-syntactic elements of L1 learning of English, and found a common route of acquisition, though the rate of acquisition varied. L1 acquisition requires little involvement of general cognitive mechanisms (Gombert 1992); the common order must be attributed to innate universal learning processes (Vainikka and Young-Scholten 2007). This evidence suggests that language-learning is a specialised and modular function of the brain that is not influenced by general cognitive mechanisms (Chomsky in Piatello-Palmirini 1979), and is the generative linguistic perspective.
Following up from Brown’s (1973) findings, Dulay and Burt (1973; 1974; 1975) investigated the route of acquisition of English morpho-syntax as a second language by Spanish and Chinese speaking children, and, again, found little difference in the order of morpho-syntactic acquisition despite different language backgrounds. Bailey, Madden and Krashen (1974) established further that there is a common order of morpho-syntax acquisition for adult L2 learners of English, again with no variations in order of acquisition despite differences in L1, age of exposure, and type of exposure. A common route of L2 morph-syntactic development indicates that there is some involvement of a specialised language module in SLA with little or no L1 transfer (Vainikka and Young-Scholten 2007).
This is the position of the theory of Organic Grammar (OG) (Vainikka and Young-Scholten 1994; Young-Scholten and Ijuin 2006). OG holds that, in the initial stages of acquisition, transfer from L1 alone guides production. Only non-finite verbs and their complements are used, and make a verb phrase (VP). The word order in the VP is that of the native language, not the target language. After the first stage, a ‘tree’ of syntactic structures of the target language is built through the interaction of UG and primary linguistic data (PLD). Each projection acquired includes lower projections, making the tree increasingly complex until it resembles native speakers’.
Alternatively, Bley-Vroman (1990) and Schmidt (1993) believe that UG has no involvement in SLA, rather there is a ‘critical period’ of access to UG that closes at the onset of puberty. It is highly unlikely that anyone beyond this age will learn a language to native-like fluency (Penfield and Roberts 1959; Lenneberg 1967). Additionally, observations of immigrant children and their parents show that although children achieve native-like fluency in the language of the host country, their parents fail to do so (Johnson and Newport 1989). There is, however, some contradictory evidence. For example, immigrant children receive more input than immigrant adults (VanPatten 1988), and this could account for the differences observed (Young-Scholten and Strom 2006). Additionally, Hawkins (2001) and White (2003) argue that the inter-language of immigrant adults is constrained by the principles and parameters of UG, suggesting that UG does influence adult inter-language.
Although generative linguists believe otherwise, Schwartz (1993) believes it would be incorrect to attribute the source of all L2 knowledge to UG because L2 knowledge can also arise as a product of general cognitive mechanisms. The effects of many general cognitive factors on adult SLA have been researched thoroughly, for example, working memory capacity (Baddeley and Hitch 1974; Juffs and Rodriguez 2008), gender (Sunderland 2000), attitude (Krashen 1982) and linguistic identity (Liebkind 1999). Also, learners who receive input in the form of instruction and correction can use general cognitive mechanisms to build knowledge of the L2 (Schwartz 1993). This process results in ‘Learned Linguistic Knowledge’ (LLK), a conscious, declarative knowledge of the target language (Schwartz 1993). It is difficult to determine the source of knowledge an L2 learner is drawing on when using the L2 (Vainikka and Young-Scholten 2007). Illiterate people, however, have limited access to L2 classroom instruction and therefore opportunities to develop LLK are minimal (Vainikka and Young-Scholten 2007). Studying SLA in this group there can offer new insights into the UG access debate in SLA theories.
The effect of literacy, as a general cognitive mechanism, on SLA, is relatively understudied (Bigelow et al. 2006). Regarding the relationship between literacy and morph-syntax specifically, the evidence presented here and elsewhere (Bigelow et al. 2006; Young-Scholten and Strom 2006) indicates that the development of literacy in the L2 positively correlates with the development of L2 morpho-syntax, though the direction of causation is still unclear. What follows is an investigation into how having low levels of L1 and L2 literacy may influence SLA, particularly the acquisition of L2 morpho-syntax. To this end, chapter 2 discusses the effect of becoming literate on phonological skills and chapter 3 the influence that this has on L1 language processing firstly children and then in adults. Chapter 4 discusses how low levels of literacy may effect L2 language processing and describes current theories and their associated evidence, regarding literacy and adult acquisition of L2 morpho-syntax. Chapter 5 is a description of the methodology of this research study, chapter 6 the results, and finally chapter 7 is a discussion of these results and a conclusion.
2. Literacy and phonological awareness
Phonological awareness is the ‘insight into how spoken words are structured and composed of individual sounds and combinations of sounds’ (Geudens 2006), and encompasses awareness of the word, syllable, rhyme, onset and phoneme (Geudens 2006). The development of phonological awareness is known to co-occur with becoming literate (Goswami and Bryant 1990). For example, Hulme et al (2005) who showed that low levels of phonological awareness predicts low literacy levels, and (Hatcher et al. 1994) showed that extra tuition to improve phonological awareness also improves literacy level. However, there is much evidence that awareness of the word, syllable, rhyme and onset may occur prior to reading (Young-Scholten and Strom 2006), therefore phonemic awareness – awareness of the smallest unit of sound, the phoneme – is discussed. In children, the co-occurrence of learning to read and the emergence of phonological and phonemic awareness have been studied extensively (Young-Scholten and Strom 2006). For example, Lieberman et al. (1974) investigated how phonological skills, developed by becoming alphabetically literate, effect the way that children segment words. Using a tapping game, four-year-old and six year old children were asked to tap in time to spoken phonemes, demonstrated firstly by the experimenter. The four-year-old participants were unable to do this but 70% of the six-year-olds managed. Tapping in time to syllables was found to be easier for both age groups, though again the six year olds were more successful. Pre-literate children are not consciously aware of phonemes and cannot segment speech into phonetic segments (Liberman et al. 1974). The authors suggest that children become aware of speech segments in a hierarchical way, with awareness of the word emerging first (Gibson and Levin 1975), followed by the syllable, and finally the phoneme.
Classic cognitive psychology studies have tested rigorously the parallel development of phonemic awareness and literacy. Morais et al. (1979) were interested in the origins of the ‘explicit knowledge of the phonetic structure of speech’, and operationalized literacy level as a predictor of this knowledge. They studied 60 people from a poor region of Portugal; the majority of these were woman. These were separated into two equal groups, the I group were illiterate, and the R group had become literate as adults. These participants were asked to delete the first sound of an utterance provided by the researcher. The sound to be deleted was one of three phonemes. The I group achieved only 22% accuracy, whereas the R group achieved 80% accuracy. The authors concluded that the ability to manipulate phonemes does not arise spontaneously. They demonstrate that these results are not the result of misunderstanding the task or an inability to manipulate speech sounds in general by conparing the results with the result of another experiment. Carried out by Cary and Morais (1979), the other experiment showed that illiterate participants cannot reverse phonemes (eg. cha – ach) but they can reverse syllables (chave – vechá). The authors conclude that it is ability to deal specifically with the phoneme that literacy aids.
Read et al. (1986) demonstrated that phonemic awareness is not than the result of literacy or education in general, rather the result of associating a grapheme with a phoneme specifically. They compared the phonemic skills of two groups of Chinese speakers, one group were literate in only Chinese characters and the other group literate in Chinese characters and the Chinese alphabetic script, Hanyu Pinyin. Logographic or syllabic scripts do not match graphemes with phonemes, and therefore associations between these are not required. The participants were required to add or delete consonants from the start of a syllable in both words and pseudo words. As expected, the alphabetically literate group achieved on average 88% accuracy while the alphabetically illiterate group achieved only 29% accuracy. This indicates that it is as a result of alphabetic literacy that phonemes can be manipulated consciously. Lieberman believes that this occurs by being made aware through orthography that the letter <b> represents the first sound in bat and bed the person can make a direct connection between the grapheme and the phoneme. Chapter 3 now discusses how these abilities affect L1 processing
3. Literacy and language processing
3.1 Pre-literate child language processing
Pre-literate children do not appreciate that language consists of words (Olsen (2002) and assume that a written word symbolises the meaning of the word, rather than the word as a label (Berthoud-Papandropoulou 1978). When asked whether cupboard is a long word, a five year old replied ‘yes, because it has a lot of things in it’ (Berthoud-Papandropoulou 1978). A word is interpreted semantically rather than phonologically. These differences cannot be accounted for by mental maturation; literate, but not illiterate, six year olds are unable to learn ‘Pig Latin’, a ‘secret language’ that requires the speaker to consciously segment and manipulate speech sounds (Savin 1972). Literacy alone accounts for this difference in language processing in children and acts through developing phonological awareness (Olson 2002). The next section further demonstrates that these changes that occur in early childhood are not the result of mental maturation but the result of literacy acquisition by discussing illiterate adult language processing.
3.2 Illiterate adult language processing
Similarly, illiterate adults process language in terms of semantics, rather than phonologically. Reis and Castro-Caldas (1997) studied 20 illiterate and 10 literate Portuguese women matched on cultural background, intelligence and pragmatic skills. The participants were asked to repeat a list of 24 highly common words, followed by 24 pseudo words. The literate group performed equally highly on both tasks, achieving on average a near ceiling score. The illiterate group achieved a near ceiling score when repeating words but performed poorly when repeating pseudo words. Reis and Castro-Caldas (1997) conclude that through developing phonological awareness, literacy allows for assigning visuo-graphic meaning to linguistic segments smaller than the word. These can be input into working memory (WM) and manipulated at will. The illiterate participants, unable to assign visuo-graphic meaning to pseudo-words, were unable to input them into their WM. The participants also completed a semantic and phonological fluency test, involving listing as many words as possible in a set time limit which, either a) were names of animals (semantic fluency), or b) begin with phoneme /p/ (phonological fluency). Again, the literate participants performed significantly better on both tasks that the illiterate participants, and as a group they performed slightly better on the semantic fluency task than the phonological fluency task. The same pattern was observed in the illiterate group, though here semantic fluency was much stronger than phonological fluency. The authors postulate that the semantic system is the major system with which language is processed, and the written word constitutes a secondary system. Literates use both systems in parallel, which, according to Reis and Castro-Caldas (1997), accounts for their superior language processing skills.