Stephen van Vlack

Sookmyung Women`s University

Graduate School of TESOL

Second Language Learning Theories

Spring 2007

Week 4 - Answers

M&M, Chapter 3: Linguistics and language learning: the Universal Grammar approach

1. What is linguistic knowledge made up of?

Linguistic knowledge is made up of everything that someone knows about their language. In the UG model this includes the integration of UG (i-language) and the periphery (e-language). Chomsky in his pursuits of universal grammar limits his discussion to knowledge of what he believes to be universal grammatical systems or i-language, but there certainly many more things involved. One thing we need to understand right from the beginning is that i-language is limited to just UG-governed properties. Many of the form issues which concern us as foreign language teachers are, in fact, not part of Chomsky’s idea of I-language. I-language does not include any language specific or what linguists call idiosyncratic systems. Thus, things like morphology (tense, number and agreement marking or complex word formation) or even our arch enemy articles are not seen as being part of I-language because they are language specific properties of English. UG theorists have made vague and conflicting claims about such linguistic properties with most believing that they are affected by UG but are not part of the core as such.

In addition to the form concerns (both in and pout of I-language), there is also USE (e-language). All native speakers are claimed in the UG mode to have the same level of control over i-language forms (competence), but they are all different in how they are able to use those forms. This is the big question - where is the information about how the language is used housed? Language use is indeed a type of knowledge, both the knowledge that is strictly linguistic (E.G. distributional knowledge such as how certain forms can be used with and ordered with others) as well as knowledge about how to process and articulate those forms. Chomsky says that this type of knowledge has to be separate because it is not innate and while innate systems might help us to learn other non-innate systems they are still learned in a different way. For foreign langyage teachers like us who see language in a different way e-language can also be part of our linguistic knowledge but UG people choose to simply ignore it.

Another reason i-language researchers like Chomsky choose to ignore e-language research is that he feels it is not scientific. I-language and e-language research differ primarily on the level that the research is done. E-research is done by analyzing surface representations of huge amounts of language samples. I-research, on the other hand, tries to analyze language from a deeper place. These researchers try to figure out the underlying process of language by creating a system of language generation which can be studied. They do not look at the sentences people actually say as e-researchers do. They make up their own sentences and check themselves if the sentences are possible or not and test their system that way.

E-language people think of language as a social phenomenon. They are concerned with many different aspects of language. They think it is as important to see how that knowledge is actually used in the real world. I-language people, on the other hand, are only concerned with trying to describe the knowledge base that language is formed from (linguistic competence).

2. What UG made up of?

UG is a highly formalized model intent of describing how languages work by focussing on the core, or what Chomsky refers to as competence. Chomsky sees Universal Grammar (UG) as a set of conditions that all languages share. These conditions are part of the hardwiring of that brain and in that respect are most certainly innate. It is important to note that in somewhat later publications (Chomsky, 1981) Chomsky claims that the properties of UG itself make up what he referred to earlier in Chomsky (1965) as the LAD. Thus, there is no LAD separate from UG. Chomsky does not see universals as constraining possible grammars (although they certainly do that), rather Chomsky sees universals as the building blocks of language. They are what define and make up language. Universal grammar, for its part, is the pursuit of these universals. For Chomsky and his followers the search for universals means working abstractly within a single language to find the basic and highly abstract principles from which that language is built. Cross-linguistic co-occurrences of these properties are only usefully sought and examined after a certain stage has been found by looking internally at a single language.

Since the 1980s) UG is widely regarded as being composed of two parts: principles and parameters (Chomsky, 1981). Principles are general and abstract rules that apply equally and necessarily to all human languages. Once more, they begin to work from the moment the child is born. Parameters, on the other hand, apply differently in different languages. They are generally seen as being bi-polar. Different languages, thus, may have different values (+/-) for a specific parameter. Parameters are more specific than principles in that they make more direct statements (and, indirectly, limitations) on the types of structures a language will allow. It has also been claimed that parameters take time to develop completely, so children might not have full, adult -like parameter settings for several years. In more recent models, (Chomsky, 1995) UG has been simplified somewhat and is seen as being composed of a computational system (principles) and a lexicon (parameters).

3. How is SLA different from FirLa in the UG approach?

SLA and FirLA are often seen as being potentially the same linguistically, but they differ in two ways. The first way they are different is in the affective/personal side and this is obviously an e-language issue. This has been already discussed previously, but is still a big and spreading issue in SLA studies. A newer issue in the way FirLA and SLA are different is in cognition and in a related vein, language processing. The idea here is that since adults have matured cognitively they will approach the learning of a second or subsequent language in a different way cognitively than they approached the process of learning their first language. Of course, it should be realized that cognition and language processing are both affected by the (emotional) affective side as well, so it is a bit difficult to clearly define what is affective and what is cognitive. Entwined in this are concerns related to processing.

4. What are the main views on the role of UG in SLA and what are the repercussions of each?

Parameter setting in SLA is based on the idea that the learner needs to either switch or simply carry over the parameter settings from the L1 to the settings of the L2. That is basically it, but it is, of course, not quite that simple. Parameters have been claimed to be related only to functional categories (DP, TP, AGRP) and not lexical categories (NP, VP, AP, PP) (Radford et al., 1999). As we read in this chapter, there is a large amount of debate currently on how functional categories develop in children. Since second language learners have been claimed to go through the same or similar sets of stages as L1 learners then we might be able to observe some interesting behavior in relation to parameter assignment and functional categories.

Basically three different types of models for second language acquisition in relation to UG theory have been proposed (Cook, 1985). They are the direct access model, the indirect access model, and then no access model. Actually, none of these are actually hard to understand. As successful and maybe sometimes unsuccessful second language learners we can intuitively get a feel for how these models might have been proposed. This does not however mean that they are valid. Well, and the first place all the models are completely reliant on the fact that UG actually exists and that UG is a specific or separate component. Cook and Newson (1996) point out that if UG actually does exist that UG can't really be separated from language learning. UG is language learning. What also seems to affect all this is the fact that second language learning is extremely complex and variant. In fact, we can go so far as to describe second language acquisition as diverse in it's very nature. This means there are an awful lot of variables which could affect what appears to be access to UG, which we have already established is basic language learning.

A couple of the main factors which might affect this are the language learning environment and the methodology for language learning. In the term `environment` we make a very strong distinction between second language and foreign language environments. It is assumed that in a second language environment a lot of the language learning will actually occur outside the classroom in the real world and therefore in a much more `natural` or holistic way. In the foreign language environment it is assumed that the language learning out of necessity will take place in the classroom away from the world world without any actual or meaningful language use. Aside form the obvious processing differences these diverse environments will create in the foreign language environment it would seem very difficult to be able to activate UG because, as Chomsky has often claimed in order to get parameters to set or indeed reset a certain amount of meaningful input is required. It could be argued that such meaningful input simply doesn't exist in most foreign language learning environments. If we take Korea as an example, we can argue that in most school settings English is not treated as a language, English is treated as a subject. Students in schools do not learn English: they learn about English. The distinction is extremely important. It is only when the students are actually learning English that they will be exposed to type of input they need will probably activate UG, that is if UG actually exists.

So you can see from all this but we're dealing with an extremely sticky problem. It is one that is not easily solved. Nor has the issue of environment and input really been well discussed in the entire literature related to setting parameters in second language learning. Ideas about access to UG have basically been acquired by assessing the errors that learners make and even more importantly the grammaticality judgments that learners have. As far as grammaticality judgments go there is no guarantee that these grammaticality judgments are based on UG. This clouds the issue even further.

Clahsen and Muysken (1989) claim that principals (the general side of UG) are always available, at least potentially, but parameters are not available and can't actually be reset. This means that learners have to start and probably just deal with the parameter settings of their first language. Once more, if the target language has different parameter settings (and it is important to remember that all languages, no matter how similar, actually do have some different parameter settings) from the target language then the learners going to have to find mechanisms outside UG to learn how that language works in relation to those parameters. This is a very interesting idea. It is neat and simple, however, I'm not sure that there is experimental evidence to support this fully. Like almost all the theories that we discussed in this chapter, there seems to be tremendous amount of variation in variability, even in first language acquisition. This just goes to show that maybe language learning is not as innate as linguists have proposed for such a long time. Even the evidence just from our dealings in class has refuted their claim. In dealing with government, it was readily apparent that everyone in this class had some idea of government and how it works in English, particularly in relation to movement. These are things that your high school English teacher could not have taught you. It could be things that you might have picked up from the environment, or it could be part of UG. Either way you somehow came up with this knowledge.

5. How does UG see language acquisition in general and what are the effects of this view on the field of SLA?

A major part of the UG model has always been to try to account for phenomenon in FirLA and SLA studies of UG are generally newer. Once more, as we saw above, there is little consensus on how UG is said to affect the second language learning process. The major problem in trying to draw conclusions about language acquisition that relate to both L1 and L2 processes is that the two are so different, at least on the surface. On the surface level the basic description of them is different. L1 acquisition has been described as being regular and fixed while L2 acquisition has been described as being irregular and various. This difference alone would force us to posit differences in relation to UG, but the differences need to be looked at more carefully. Acquisition does not occur in a vacuum or a black box, in fact it could never occur in such conditions, both for L1 and L2. It is hard to tell what differences between L1 and L2 acquisition are the result of environmental (external) differences or differences in brain functions in relations to language or other systems and functions (internal). The two different areas affect each other strongly. Thus, it is very hard to draw on parallels between the two.

Johnson, Chapter 3: The cognitive tradition and second language acquisition

6. Chomsky makes a fundamental difference between development and acquisition. What is the nature of this difference and why is it important?

Development is dependent upon other concerns like maturation whereas acquisition occurs free of other variables. Initially UG was seen as being wholly acquisitional in nature, but there are currently many maturation-based hypotheses related to UG in FirLA. There are three different hypotheses for maturation in relation to universal grammar. They are the continuity hypothesis, the discontinuity hypothesis, and the maturational hypothesis. The continuity hypothesis, supported by Pinker, claims that people are always open to UG. They have a constant access to UG throughout the acquisition process. As a result the utterances that these people, primarily children, make must and will always conform with UG. The reason why they may not conform to UG is because of other kind of limitations which exist outside of language, such as attention span, lack of vocabulary, or cognitive development. In the discontinuity hypothesis principles are realized differently at different stages in a person`s life. The acquisition process allows for the realization of parameters in a stepwise and somewhat disjointed fashion. Following this discontinuity hypothesis view, learners move through different possible stages which all feed into each other and allow for the developments to occur. Hence, development is a serial process. In the maturational hypothesis certain principles emerge in the mind at particular times. Basically the emergence of these principles is biologically determined. It is innate.

7. What is the crux of Gregg`s argument in favor of a UG model of acquisition and what is your reaction to him?

Gregg seems to love the UG model simply because it is the most descriptive model out there and he is indeed right about this. Most models of SLA do not provide any descriptive framework that can be looked at and tested in detail. Many of them are simply too vague. The UG model`s detail and precision are two of the main properties which set UG models apart from many other SLA models. Just because the model is better developed and quite broad does not mean, however, that it is necessarily correct and certainly is not correct in all its assumptions. This is a very dangerous and faulty way to go about arguing for a specific theory. A single theory may very well be wrong.

8. Why is Long so vehemently against socially oriented theories of SLA and so strongly in favor of cognitive approaches? Do you agree?

According to Long (1997) learning is about the mental state of the learner. It is, therefore, necessary to focus on that mental state. The social setting plays only a minor role in the formation of that internal mental state. Again, one of the reasons Long sees little future in socially-based models is that they are hard to observe and even harder to prove. The general idea of language learning as being testable because it is the same for everyone goes out the window with socially-based models. For this reason Long does not see them as being amenable to scientific study.

References

Chomsky, N. (1965). Aspects of the theory of syntax. Cambridge, MA.: MIT Press.

Chomsky, N. (1981). Lectures on government and binding. Dordrecht: Foris.

Chomsky, N. (1995). The minimalist program. Cambridge, MA.: MIT Press.

Clahsen, H. and P. Muysken (1989). The UG paradox in L2 acquisition. Second Language Research 5, pp. 1-29.

Cook, V. (1985). Chomsky`s Universal Grammar and second language learning. Applied Linguistics 6, pp. 1-18.

Cook, V. and M. Newson. (1996). Chomsky’s Universal Grammar (2^nd edition). Oxford: Blackwell.

Long, M. (1997). Construct validity in SLA research: A response to Firth and Wagner. Modern Language Journal 81, pp. 318-323.

Radford A., M. Atkinson, D. Britain, H. Clahsen, and A. Spencer. (1999). Linguistics: An introduction. Cambridge: Cambridge University Press.