Stephen van Vlack
Sookmyung Women`s University
Graduate School of TESOL
Developing Bilingualism
Fall 2007
Week 8 - Answers
Hamers and Blanc, Chapter 7: Information processing in the bilingual
1. What is the basis of the much discussed coordinate/compound distinction in bilinguals? What evidence exists for or against this? (pp. 163-167)
When considering information processing in bilinguals the main question that needs to be answered revolves around the relative amount of integration or separation of the systems. What this really means is that by looking at this we need to try to determine whether the two languages of a bilingual are separate or integrated and, if integrated, where, how, and to what degree they do manage to integrate. This might seem like one of the those, what`s the point questions, but the relative integration and/or separation of a bilingual`s two languages is of great importance to us as teachers of English in Korea because it relates to cognitive processes and learning. Ultimately, it tells us about whether our current models of language are correct or incorrect and it reminds us of the important notion that what we do in the classroom changes the brains of our students in a fundamental way. How we teach will have a direct effect on how information is stored and made available for use in our students.
Based on the confusing array of data presented in this chapter we would have to lean toward the conclusion that the current models (inspired by the generative school of MONOlingual linguistics) are generally unacceptable. In this way we are following the claims of Herdina and Jessner (2002) and Cook (1993) that linguistic theories have suffered because they are based on monolingual native speaker models, which are particularly useless because such a model does not exist in reality. Following their suggestions we study such bi/multilingual phenomenon because they bring us closer to how real language users all over the world actually integrate their languages. Thus, integration, rather than separation is the norm and it is the thing that we as language teachers should aim in fostering in our own students.
Since the coordinate/compound distinction, at the surface at least, is about concepts and lexical representation, then we need to look towards lexical items (words) to try to determine what it really is. The type of word, whether concrete or abstract, seems to be very important to the compound/coordinate distinction. In bilinguals, generally, there will be more compound-type organization for concrete words and more coordinate-type organization for abstract words. This result is not hard to predict when we think of the nature of abstract and concrete words. Most concrete words have a direct physical representation in nature, which, for the most part, does not vary too much from place to place. For this reason it is relatively easy to set up direct translation equivalents for concrete words. There is very little cultural interference there at the conceptual level. In abstract words, on the other hand, culture is at the core. They represent concepts that do not exist clearly in the world. Culture has either created them entirely (as in the case of such words as jealousy or hate) or interpreted them from some unseen source (as in the case of pre/postpositions). These are concepts that do not transfer neatly from language to language because they are so tied in with cultural concepts and do have a simple physical form.
Again, it would be nice if things were this simple, but unfortunately they are not. Concrete words do not only acquire meaning based on imagens. They also have other types of associations, some of which may be abstract, which give them their meaning. Likewise, abstract words sometimes have very concrete associations and even imagens associated with them. In fact there is a whole field of study called semiotics, which is devoted to exactly this abstract concept-image-meaning connection.
We can think about the compound/coordinate distinction in a more general way by trying to determine if it is a psychologically real phenomenon. Psychological reality, here, is meant whether there exists a real difference (a physical one, as determined by observation of particular behaviors in experimental situations) in the minds of bilinguals. The best way to determine this would be to run some tests. Tests were done (Lakobovits and Lambert 1967, Lambert and Rawlings 1969) based on word associations. This is a good type of test because it can determine how closely associated words are in the lexicon and what type of connections link them. The results of many studies show that there is indeed some degree of psychological reality to the compound/coordinate distinction. There are, however, a number of other studies which show just the opposite; that no real distinction exists. Part of the reason for these results is that the representational system of a bilingual is not stable over time, but changes according to environment of learning/exposure and level of competence (Opoku 1983). A bilingual will draw on differences cognitive resources related to the same concept based on small situational differences which are virtually impossible to replicate in the laboratory. Based on this it is very hard to say that there is real psychological reality to the compound/coordinate distinction because we can`t actually test it very well. Any reported differences may be attributable to experiential differences in acquisition (language acquisition history) and the amount and type of experience with the language but a large amount of variation persists as that is the very nature of language. It is also important to realize that the distinction is by no means absolute. There are no language users who are entirely compound or entirely coordinate. In reality we are all a mix and the mix changes constantly.
What it all comes down to really is not only the type of connections, but the amount of connections between entries in the mental lexicon of a speaker. Important, or essential really, are not just the connections between the two linguistic systems, but the connections within each system. Thus an L2 lexicon that has little or few links among its own members or entries will have a greater percentage of all its connections running to the L1 lexicon. Without going too far into this, it should be clear that the compound/coordinate distinction is a tremendous oversimplification which doesn`t really help us too much in figuring out the reality of bilingual; processing/production.
Research in the field, as mentioned above, has shown correlations between language acquisition history and the compound/controlled distinction in bilinguals. By language acquisition history we mean how, when, and where the person acquired both their languages. We have already talked a great length about balanced and dominant bilinguals. Gekoski has argued that only true balanced bilinguals show compound organization. This is because it necessary for the learning environment of the two languages to be the same, for compound bilingualism to occur. To take this one step further it seems that not only must the environment be the same but the timing must be the same as well. It is only under such circumstances that the words can be associated together. If the learning of the L2 occurs in a different place or at a different time than the L1, then the words will be associated in a coordinate way. At the same time, however, it is important to realize that an individual`s linguistic skill is constantly changing. As people use a language more and better they are better able to interact with the world defined by the speakers of that language. Based on this observation, we would expect the types of connections in the mental lexicon to change even as the conceptual base undergoes a neutralization process. Simply more, more and more complex connections would be added with more exposure, thus altering the shape and working of the mental lexicon, but such changes would occur in a diffusional way. That is, not immediately, but from one entry at a time.
2. What are the two main models proposed to account for memory storage in bilinguals and how is each supposed to work? (pp. 167-173)
The two main models that have been used to account for memory store in bilinguals are single store and the separate store model. I know this sounds boring and like another who gives a damn question, but think about it. Are the memories you have capable of switching codes (languages)? For example, if you grew up in Korea and spoke and only Korean, but move to New Zealand and assimilate into that society, it might be possible that the language of your memories might change. You will remember your family members in English and not Korean. You may even remember them as speaking in English and not Korean. Do you think it is possible. That`s what we are trying to find out. Additionally, it is not just memories that we are concerned with, but with conceptual/lexical storage as well.
What we basically come up with at the end of all this is that there seems to be both separate and single store mechanisms in the brain. This means that at some level the two languages interface, they are connected, but at another level they are separate. This seems to make sense due not only to the experimental evidence (which will be swiftly reviewed below) but also because the model is so common. From a cognitive view of the brain, we posit that the brain functions based on a limited set of highly abstract principles. It is, therefore, not only acceptable, but also desirable that we find the same models used over and over again to describe different systems.
The separate store model is supported by evidence that intralingual and interlingual behavior are different. That is there is a limited, slow, and difficult transfer from one language to another. The tests designed to examine this are often based on priming effects. Now, let`s take a step back and figure out what priming effects are.
We all know that the brain works on electricity. When you think of a certain word a certain amount of electricity is circulated in certain parts of the brain in a specific pattern to activate a word. That patterned flow of electricity triggers the word to pop into our conscious thoughts. It does this by activating specific neuronal synapses along a long series of connected pathways. When a synapse if activated so that electricity can pass over it this is called firing. It is also true that some of the electricity stays behind and keeps the synapses active (or primed = ready to fire again easily and smoothly and with an overall much smaller amount of electricity), so to speak. This is priming. The bit of electricity stays behind to get the synapses that call up a specific thought (a word, for example) ready to be used again in that it makes the word much easier to trigger again. This is a very efficient part of how the brain works and means the more we use a specific pattern or neuronal firing used to activate a word or phrase (whatever), the easier it is to use. Now, we also know that words are not stored in the lexicon like entries in a dictionary. Words are stored according to a number of different principles simultaneously. Also, words in the lexicon do not have dictionary-like definitions which make their meanings clear. Words are defined in a large part, by their relationships with other words. So, for example hot and cold are closely linked in the mental lexicon due to their antonymic relation. Not only words with semantic relationships but also words with situational or collocational associations are linked in our lexicon. OK, now when electricity is sent to synapses to trigger a word it also primes other links to other related words which are linked closely to the triggered word. It is like dropping a pebble in a pond. The pebble hits what you want, but the ripple effects from the initial pebble drop disperse and effect many other things in the lake are affected. So it is with priming.
Now, according to the standard argument (and I fear it really is not this simple), if the languages are stored together (common store) then we should be able to see examples of cross-linguistic priming effects, but if they are stored separately then priming effects should be limited to one language only.
Much of the proof against the single store model comes from very simple arguments, such as there are different associations for translation equivalents. This is hard to swallow in that it would create a very messy system in which a lot of the priming would go haywire. In addition when addressed in a certain of their languages and then later asked what the language was the bilingual is always able to remember correctly which language which the input was given in. If there were a single store we would expect some mistakes and confusion about this.
The common store model is supported by evidence that intralingual and interlingual behavior are the same. That is there is a neat, quick and easy transfer from one language to another. Evidence for a common store comes from Young and Navar (1968). They report that a Spanish-English bilingual had different associations for a single set of translation equivalents. The word in question is cat - gato. In the subject`s Spanish, gato (En. cat) is associated with comer (En. eat). In their English, cat is associated with house (Sp. casa). When asked to give the translation equivalent of gato the subject came up with casa. It is clear that there were cross-linguistic priming effects. This could only happen if there were some connection between lexicons.
3. Why were `switches` proposed as a means of controlling both input and output and do you think they are real? (pp. 174-180)
The idea of a magic switch somewhere in a bilingual`s brain which would allow them to switch back and forth from one language to the other without influence from either side is a very nice idea. It represents a simple, neat idea to a complicated problem and we always want to find the simplest (most elegant) description we can. We have read that in balanced bilinguals there does not seem to be any problem with language interference form either language. Children do go through a short stage early in bilingual development where they confuse the two language, but it is a short period that never returns. Based on this type of observation a switch was proposed as the simplest solution to this phenomenon.
A switch seems like a great idea, so simple and effective, but on a closer look it does not make sense. Well, first of all there is all the evidence for the single store model. Also we have the major problem of bilingual-specific behaviors (translation, code-switching, code-mixing). All of these behaviors show a large amount of integration between the two language systems. If there were a switch it would seriously impede the bilingual`s ability to engage in these activities.
The evidence form syntax supports neither the single, nor the common store exclusively, but argues that there must be both. This seems to be same regardless of the model of syntax (generative -vs- cognitive) we are following but for very different reasons. For generative syntax the distinction is in the different components of UG. We mentioned briefly earlier in this class about the two main components of UG (universal grammar): principles and parameters. We know that principles are general concepts that work the same way in all languages. Following this a single store would make it easier to transfer universal syntactic principles from one language to another. Since students of a second language do not create utterances that violate UG, even to a very early stage of development, then there must be some integration. Parameters are concepts that have language-specific settings. When a person is first exposed to an L2 they carry over the parameter settings from their L1, which, of course, results in severe negative transfer (Braidi 1999). As they are exposed to the L2 more they change the parameter settings. Based on parameters the languages have to be stored separately so that the different parameter settings of the two languages do not interfere with each other. The nett result is that we need a system that is integrated at some point and separate at another. For cognitive linguistics which sees syntax as being generated from an underlying cognitive/conceptual representation of language, as with the common -vs- separate store hypothesis, some of these concepts will be the same or similar across languages and some will be very much language specific. So, here too, we seem to find support that we do in fact need both.
4. How does the dual-coding model differ from the hierarchical models? (pp. 184-189)
The dual-coding model is based on the belief that words are stored in two different ways. This model draws on the belief that there are imagens, which are non-verbal and represent physical objects or concrete images and logogens, which represent words (concepts) (Paivio and Begg 1981). Bilinguals have two verbal representations, one for each language and an additional representation for imagens (Paivio and Desrochers (1980). The two types of representation correspond to different organization in bilinguals according to this model. Logogens are housed separately, but imagens are integrated.
Hierarchical models are different in that they give a more theoretical description of the same phenomena and as such can describe things in a broader, yet more powerful framework. Instead of just saying that there are imagens and logogens, hierarchical models call for two different levels of representation. They claim that the words themselves are stored at a lexical level. The semantic features of the words, however, are stored at a conceptual level. This is quite different than the dual-coding model in that the concepts are stored singly and the lexemes themselves (even translation equivalents) are stored separately. The hierarchical models seem to account better for data which seems to indicate that translation equivalents often do not have a direct association. The association is interpreted through the common semantic features of the words.
5. What is the strength of the asymmetrical storage models? (pp. 189-193)
The primary strength of the asymmetrical models in that they are better able to account for the differences between different levels of bilinguality (whether balanced or dominant and the relative degree of proficiency of the L2 in the dominant category) all within the same system, which is also quite simple and easy to understand. In reality the asymmetrical model is quite similar to the hierarchical model but with a little quirk. There are two levels (lexical and conceptual), as in the hierarchical models, but there is an additional difference. Within the conceptual level a difference is placed between shared and language-specific concepts. As the amount of bilinguality goes up, so too does the amount of shared concepts (common store). This occurs as a direct result of the type and amount of form-to-function mapping the speaker has. Lower level speakers, which have little to no form-to-function mappings in the L2 alone, will have to rely on translation as a way of trying to communicate in the L2. They have English forms (for example) mapped to Korean forms which, in turn, are mapped to Korean functions. Thus in the initial stages of acquisition all English forms are mapped only indirectly, through Korean forms, onto any kind of functions att all. Later, with more and more varied patterns of use, these English forms might be able to form their own more direct connections to Korean functions. In the next stage of development English functions might come into play and/or a process or neutralization will ensure which prunes out some language specific elements of the functions themselves so that a single function can be used for both or all languages involves regardless of differences in surface level forms. Following this model, balanced bilinguals will be able to communicate easily and effectively in both languages because they can communicate directly from the conceptual level. This will also allow them to code-switch and code-mix without any great hesitation or delay.
6. Do bilinguals really have two different personalities? (pp. 195-196)
Just as we know that a bilingual is more that the simple sum of two native speakers in the realm of pure language use, so too do we know that a bilingual/bicultural person is not the same as two different native speakers. Yes, bilinguals adapt to follow the cultural rules of their two languages, but each side still retains the core of their personality. In this way, it is unfair to say that bilinguals are like dual personalities, split down the middle, or that they are strange or that bilinguality might result in a kind of mental disorder. This is simply not true. At a young age the balanced bilingual comes to terms with the cultural and personal side of both languages and finds a way of dealing with each effectively. If this did not happen, then they would not be able to develop into balanced bilinguals. It is not that different from the acculturation process that consecutive bilinguals have to go through. We know that in the middle of this process anomie starts to emerge as a force which can seriously hinder the learning and cause the learner to give up on learning the target language entirely. All successful language learners, whether balanced or dominant bilinguals have found a way to deal with the issue of personality. As with the other, more formal elements, we have discussed above it seems that bilinguals are able to draw from an integrated store (the nature of which depends on the type of bilingual) of thoughts and behaviors, but there are also special thoughts and behaviors reserved for each cultural situation.
Kecskes and Papp (2000) Chapter 7
1. What is pragmatic transfer and how can it be stimulated?
In order to understand pragmatic transfer we first need to understand what pragmatics is. A person`s pragmatic skill relates to how they actually use the language they speak. Pragmatics relates to how a person uses language to achieve certain purposes, whether these be communicative or cognitive. Now, it should also be clear, and we have stated it earlier in this course, that language is deeply connected to culture. How, why, when, where, as well as the specific forms used by language are all determined by the culture in which the language is used. Use determines usage. Pragmatics is the skill that allows us to produce comprehensible language as well as understand all language use. In a very real sense we use our language to make sense not only of language but of the world itself. Our L1, then, determines our world view, the way we understand the world to work, the way we interpret things in the world.
Now, some of the ways in which language is used (pragmatic skills) will be universal. Some pragmatic skills apply to all languages because all human cultures, and languages as well, are much more similar than they are different. Pragmatic skills, therefore, are prone to transfer. What generally happens is that the pragmatic skills of the L1 are transferred onto the L2 usage and with little difficulty. In fact, it has been argued in Kecskes and Papp (2000) as well as Kasper and Rose (2002) that pragmatic skills are the easiest to transfer among all the linguistic skills because of their real world nature. This does not mean that there are not many language specific elements of pragmatics, but it is more universal, than lexis for example.
The more universal nature of pragmatic skills means that they should be easy to activate for transfer. But activation requires language use, not even authentic language use initially, but really any kind of language use, even reading can be used to develop pragmatic skills. We saw in this class how reading skills can transfer. As we look below at how pragmatics relates to actual writing then this will become clearer.
2. What is pragmatic competence and how does it relate to norms of conversation?
As mentioned above, pragmatic competence is the ability to use language effectively in the real world, both for production and reception. Now, in order to do this there are universal ideas such as the idea that all instances of language use (language use events) have a purpose and a meaning. So, for example, if I walk into the classroom and scream something like, `that`s it` all of you are going to start looking around desperately to find out what I might be talking about. It is your universal pragmatic competence telling you that I wouldn`t say something unless there was a good reason, therefore you must find the reason.
Now, there are also language or culture specific norms which should be followed for effective communication and/or interpretation. For example, how one apologizes in Korean and English varies a bit. In Korean silence is an accepted norm for apology, while it is not in English. Norms determine how people within a specific cultural group approach certain situations and how they can deal with them effectively through the use of not only certain forms but of strategies and these differ quite a bit from language to language/culture to culture. Once we take all these different norms of language use together we can begin to understand how it is through pragmatics that we not only understand how language could also come to understand the world. This should be obvious because language is always used in context. We do things with language. Language is a way that human beings, and only human beings, deal with life on a very general basis. Linguistic behavior and social behavior are really different sides of the same coin and it is in this way that our knowledge of linguistic pragmatics really determines how we view what is normal in the world.
3. How is written discourse related to thought and why is this important for language development?
As mentioned above, language becomes not only the way one uses language itself, but the way one perceives and interprets the world. Language is the way we think. Our logical structure is reflected in and affected by the language we use. The easiest way to deal with this is by simply looking at how different speakers of different languages go about writing. Studies have shown that different languages, based not only on their structure but also on their cultural norms, go about organizing their writing in different ways. Here of course we`re talking about more formal types of writing in longer texts.
These systematic differences were shown in studies of how students of diverse language backgrounds structure their written discourse. Think about your own experience in learning how to write formally in English. I am sure you have had and maybe even continue to have difficulties in writing well in English and not just because of grammar and vocabulary. Based on my own experience as a teacher of English writing/composition in Korea I can tell you that the students really do come with a different logic. They bring with them the norms of Korean discourse to English and even more. Because writing, on an even grader and much more controlled scale than speaking, is about problem solving how a person approaches an empty piece of paper is determined by much more than just the linguistic forms they are going to use. There are strategies for how they suppose to fill that paper up. This is how pragmatics works in writing and I am sure all of you have felt it..
Being able to deal with and use the deep seeded effects of the L1 is not an easy matter. As reported in Purves (1986) students need to reach a threshold level in their proficiency before they can begin to merge systems and use a CUCB for regular production in both languages and this applies to writing as well as to speaking. Interestingly, exposure to the L2 in the L2 environment alone was not a determining factor in the development of the CUCB, but rather the amount of formal exposure to the L2 prior to exposure to the L2 environment. This is where we get the optimal transfer, as was discussed in previous chapters. We need skills to flow downhill form strong areas to weak areas. Once this CUCB is established through an intensive exposure to the L2, leading to a process of concept neutralization, then the learner will be able to make all skills available in all the different languages they have and that is exactly what we want.
Herdina and Jessner, Chapter 1: Introductory remarks
1. Explain the DMM model of multilingualism. How is it different from traditional approaches to multilingualism?
Since the majority of the world`s population is multilingual, we have to build our concept that the speaker`s language system should be flexible to adapt the order of more than one language. Dynamic Model of Multilingualism is a psycholinguistic model that sees language change on an individual level as a function of time, focusing on the variability and dynamics of the individual speaker system. According to the idea of multilingualism presented it here must be accepted that language is closely related to the concepts of personal identity, ethnicity, and multiculturalism, not only realized as the linguistic norm, and DMM includes a psycholinguistic focus on the systems-theoretic approach, which is based on the behavior research of living systems and a dynamic interpretation of the systems model.
DMM by Herdina and Jessner sets several purposes as follows. It provides a new set of concepts for the interpretation of psycholinguistic phenomena observed in speakers of more than one language, and also provides a useful model that enable us to think about a multitude of contradictory related to multilingualism in a more coherent way. It determines the relations among various factors within the language systems and makes predictions concerning the typical development of these variables. DMM provides an innovative theoretical framework as well in which it is possible to ask meaningful questions concerning multilingual development and obtain more satisfactory answers to the questions surrounding multilingualism as a psycholinguistic phenomenon with sociolinguistic consequences.
Compared to traditional approaches to multilingualism, a systems-theoretic interpretation of multilingualism suggested by DMM allows for a more realistic view of the phenomenon of multilingualism. In accordance with this systems-theoretic principle, DMM views a multilingual speaker as a complex psycholinguistic system comprising individual language systems. It also tries to create an explicit model of multilingualism specifying variables (dependent and independent) and making predictions about the development of multilingual systems.
Research into multilingualism is expected to exert considerable influence on linguistic theory in the future because it certainly has not do so to date. It is probably true to say that recent research into multilingualism has raised rather more questions concerning the acquisition of languages and the nature of human language ability than it has provided answers. Traditional conceptions and explanations of language learning and resulting cross-linguistic effects have at the very least been called into question. Based on this our interpretations of language learning, what it is to know a language, etc., urgently require revision. The authors believe that a large number of theories currently discussed in research on language acquisition and multilingualism lack an overall theoretical foundation.
The authors assume that the phenomenon of bilingualism on which most of the research available has been conducted is essentially a variant of multilingualism. One of the great advantages of DMM is a holistic conception and understanding of multilingualism.
Additional Notes
Multilingual acquisition can be defined as the process of acquiring more than two languages. It thus comprises the consecutive and simultaneous acquisition of three or more languages. Multilingual acquisition is linked to multilingualism because to acquire a third or additional language is to acquire some type of multilingual competence and therefore some type of multilingualism. Multilingual acquisition presents more diversity than second language acquisition and its study presents greater complexity. This diversity is increased if we consider that the process of acquiring an additional language can interrupt the acquisition process and then be restarted again (L1 LL2 LL3 LL2). The processes used in third language acquisition may be very similar to those used by L2 learners but the additional language complicates the operations of the processes. The diversity and complexity of multilingual acquisition give rise to situations which are unique in language acquisition and which justify the need to conduct research in order to identify the characteristics of multilingual acquisition and the specific operations that affect this process.
The main research areas in multilingual acquisition are relating to answering the following question:
1) How do children acquire more than two languages simultaneously?
2) Does bilingualism influence third language acquisition?
3) How do individual factors affect multilingual development?
4) How does proficiency in L1 and L2 affect the acquisition of a third language?
5) How is cross-linguistic influence reflected in third language acquisition?
To sum up with research findings to some of the questions above;
The influence of bilingualism on multilingual acquisition
Studies on the effect of bilingualism on third language acquisition conducted in different contexts tend to associate bilingualism with advantages in third language acquisition. The findings of these studies on third language acquisition share some characteristics with the results of other studies on the effects of bilingualism, because those studies in which bilinguals present no advantages involve subtractive contexts (Lambert, 1974). However, the comparability of the results of these studies is severely limited by the diversity of the specific areas of language proficiency tested and by their differing research methodologies. Furthermore, given the diversity and complexity of third language acquisition these findings need to be confirmed by more studies involving a wider variety of acquisition contexts.
Bilingual Competence and Multilingual Acquisition
Even though research studies seem to confirm interdependence among the different languages involved in multilingual acquisition, more of them are need in this area if we are be able to specify the nature of the relationship between the languages and the way this might affect multilingual acquisition at different stages. More studies on multidirectional interdependence among the languages involved in multilingual acquisition and on specific areas of language proficiency (phonetics, lexis, grammar, pragmatics, discourse) will provide insights into the characteristics of multilingual acquisition.
Individual and Contextual Factors in Multilingual Acquisition
Even though the number of studies on the role of individual and contextual factors in multilingual acquisition are still very limited, results indicate that multilingual learners may benefit from being more experienced language learners.
Cross-linguistic Influence in Multilingual Acquisition
Transfer is more likely from the first language than from those learned later on, such effects seem to be less potent than typological similarity between the languages. Research on cross-linguistic influence in third language acquisition indicates that linguistic typology is an important factor, one that determines the choice of a specific language as the source language of influence. It also indicates that bi/multilingual learners could use one of the languages they know as the base language when acquiring an additional language. Research on multilingual acquisition presents interesting findings regarding the relationship between bilingualism and third language acquisition, linguistic interdependence and cross-linguistic influence. Research into second language acquisition and bilingualism can certainly befit multilingual acquisition, which shares some characteristics with these phenomena. On the other hand, a thorough understanding of the diversity and complexity of multilingual acquisition itself is certain to provide insights that are relevant not just to the study of second language acquisition and of bilingualism, but to all researchers interested in language processing.
Some Features of Multilingual Development
One of the common denominators in current theories of language learning seems to be the idea of a gradual sequence of language improvement leading to an acceptable degree of mastery of a language system. Language acquisition has thus traditionally been considered a linear process. However, according to biological principles language development is seen as a dynamic process with phases of accelerated growth and retardations. The development is dependent on environmental factors and is indeterminate. The reason why the growth curve slows down is because in real-world terms the learner is confronted with limited resources. Within a psycholinguistic context, these limited resources are expressed in the amount of time and energy learners are able to spend on the acquisition and the maintenance of a language. If learners do not continue to refresh their knowledge of a particular L2 or L3, a gradual process of language attrition will set in.
Learners achieve significantly differing levels of competence due to individual learner differences. At least two types of factor determining the rate of language growth or language attrition exist: The psychological factors in sociological factors-aptitude, perceived language competence, anxiety, motivation, esteem-. The complex relationship between the factors determines the rate and direction of the development of the new language system. These factors not only relate to language growth, but also relate to each other in more than merely unidirectional relationships. As soon as the system contains feedback loops, it obviously has the ability to determine its own conditions of growth. The input provided by the individual factors will determine the direction of growth and the size of the input will determine the rate of growth. This behavior can be termed `auto-dynamic`.
Maintenance, reversibility and stability
Acquired language systems do not exist side by side in `mutual harmony` but start to interfere with each other. A few of the influencing factors are: the limitations of available resources, the pressure created by competing language systems, the duration of language maintenance, and the age of language acquisition. Stability will depend on the degree of flexibility present in the maintenance effort that the learner of a specific language has to put in. Gradual language attrition is an inversion of language growth. Lack of maintenance of a language system results in an adaptive process by which language competence is adjusted to meet the perceived communicative needs of the individual speaker.
Interdependence
The development of each individual language within one multilingual speaker largely depends on the behavior of previous and subsequent systems. For example, two other bi-directional relationships can take place in third language acquisition: the L3 can influence the L1 and be influenced by the L1 (L1-L3) and cross-linguistic influence can also take place between the L2 and the L3 (L2-L3).
Change of quality
Multilingualism that is not automatically interpreted as bilingualism, results in a qualitative change in the speaker`s language system. It leads to an enrichment of the individual language system but, as the whole system adapts to meet new environmental and psychological requirements, also changes its nature. The acquisition of a further language leads to the development of new skills which begin to form part of the multilingual repertoire. These fall into three main categories: language learning skills, language management skills, and language maintenance skills. All the skills could be seen as contributing to language awareness or metalinguistic awareness which has been identified as one of the cognitive advantages bilingual develop due to contact with different language cultures. Applying Cummuns` threshold hypothesis (1976) to multilingual acquisition, the following might be expected: attaining an upper threshold of bilingual proficiency leads to cognitive advantages which in turn benefit third language acquisition hypothesis according to which the cognitive outcomes of bilingualism mediate the relationship between bilingualism and multilingualism:
additive bilingualism benefits in
or à cognitive advantages à multilingual
upper threshold acquisition
A dynamic Model of Multilingualism
Investigating the complexity of language systems
A multilingual, the system not only exhibits its own kind of behavior but, due to certain parameter changes, is also subject to change over time. In order to describe the nature of multilingualism in its various patterns and to be able to explain the changing nature of the linguistic phenomena linked with multilingualism, Herdina and Jessner have developed a dynamic model of multilingualism (DMM). The process involved in the learning of a third language is characterized by complexity, which is a characteristic property of an auto-dynamic system. This complexity is the result of interaction between various factors, linguistic, social and individual. The complexity of a system should be taken to imply that due to the recursive relations between the factors contained in it-the system develops its own dynamism or in other words becomes auto-dynamic.
Theoretically the understanding of multilingualism has to precede the understanding of monolingualism, is probably most aptly identified with the philosophical tradition of holism. Psycholinguistic holism must assume that a multilingual system is not to be considered as no more than an accumulation of the effects of individual language systems and their acquisition, and that it cannot be adequately understood by adopting this approach.
The need to develop a new understanding of multilingualism that breaks with the traditions of interpreting language acquisition as a linear process, and to formulate new applications of multilingual theories such as the investigation and encouragement of trilingualism as a linguistic and sociocultural phenomenon, present new challenges to theoretical and applied linguistics. Cross-linguistic influence in multilingual language acquisition will be very useful for language policy makers, teacher trainer, teachers and students interested in language acquisition and bilingualism because it provides information that can be useful for curriculum development and offers insights into the processes of acquiring and using several languages.
References
Cenoz, J. (2000). Research on multilingual acquisition. In J. Cenoz and U. Jessner (eds) English in Europe: The acquisition of a third language. Clevedon: Multilingual Matters.
Cenoz, J. and Genessee, F. (eds) (1998). Beyond bilingualism: Multilingualism and multilingual education. Clevedon: Multilingual Matters.
Cenoz, J., Hufeisen, B., and Jessner, U. (eds) (2001). Cross-linguistic influence in third language acquisition: Psycholinguistic perspectives. Clevedon: Multilingual Matters.
Herdina, P. and Jessner, U. (2000). The dynamics of third language acquisition. In J. Cenoz and U. Jessner (eds) English in Europe: The acquisition of a third language. Clevedon: Multilingual Matters.
Hoffmann, C. and Ytsma, J. (eds) (2004). Trilingualism in family, school and community. Clevedon: Multilingual Matters.