Stephen van Vlack

Sookmyung Women`s University

Graduate School of English

Introduction to Linguistics

Fall 2003

Answers for Cook and Newson, `Chapter 3`.

1. How does Chomsky basically see language acquisition as taking place?

1. According to Chomsky there are two main parts to the language acquisition process: They are Universal Grammar, which for the purposes of acquisition is restricted to the assigning of parameter values, Peripheral Grammar, which is the acquisition of lexical items and idiosyncratic values. As you should definitely know by now, UG is broken down into principles and parameters. Both begin to work automatically to aid in and direct the acquisition as well as constrain possible structures. Parameters, in contrast to principles, have an additional side to them. They are usually composed of binary features (that means that they either carry a plus [+] or a minus [-] values). Each language has different values for these parameters and the only way for a child acquiring a language to make the parameter settings is the be exposed to the language. Chomsky believes that they make the judgment of what the parameter settings are through positive input. Lexical acquisition is acquisition based on lexemes or words and belongs to the periphery in that it must be learned. It is not automatic like the UG features. Chomsky believes that the lexical entries for words are more then just dictionary-like list entries. The entries in the mental lexicon contain all sorts of information about the word, including syntactic as well as phonological information. Basically, all language specific data is housed in the lexicon. The lexicon is, therefore, the center of peripheral learning. The acquisition of this language specific-data is, to a certain extent, directed by the principles and parameters of UG, but most of it simply occurs as the byproduct of the structure of the brain in response to positive input.

One thing that should be mentioned here because it is purposely NOT part of Chomsky`s plan for acquisition is e-language phenomena. Chomsky has no way of accounting for how children come to know the pragmatic rules of language used for their language. He purposely stays away from this, just as he rejects acquisition along lines of cognition (Piaget). Again, he may be skirting the issue, but he says he has a reason for doing so. Since he sees the knowledge of language as being separate from everything else, he sees no reason for him to have to explain other related phenomenon. He is only interested in the development of competence. After all, he can`t do everything. This is a weakness (I think) to Chomsky`s whole model of language. It leaves a bad taste in my mouth because so much is left unexplained. I and E language must integrate somewhere.

All human children start off as a blank state. This means their brain has not housed any data yet. We don`t want to say that children`s brains are empty because we know they have something in there (UG), but they have not yet any raw data on their language. As they are exposed to real language data they then, with the help of UG, move through a series of states which take them to the final, static, state of competence that they will have most of their lives.

2. What arguments does Chomsky use to support his claims about language acquisition?

2. The main argument that Chomsky uses to support his claims about language acquisition is the Poverty of the Stimulus Argument. This argument is quite simple. What Chomsky says is that the input that children get is simply not enough to account for what comes out. If children merely repeated what they hear, then there would be no creativity in what they say. We know this is not true. Language is by its nature creative. So, let`s say that it is creativity is innate. Then we have children making creative utterances which would not necessarily have to fit into any patterns. We know that all human language is constrained. Children never make utterances in all the different possible patterns as we would expect if creativity was what was innate. OK, now if children are creative and are also somehow able to pick up basic patterns in what they hear, that will account for overgeneralizations, but it will still will not account for the fact that children do not say anything. From the very beginning there seems to be something constraining what a possible language can be. Chomsky`s idea is that since the stimulus children get is limited there must be something concrete already there. Children must be born with a specific set of knowledge that directs their acquisition. After all, parents do not teach their children language. That would be impossible because the parents (in almost all cases) have no idea how language works and would not have the slightest idea how to actually teach it. Thus there must be a set of knowledge that we are all born with that only needs a bit of positive input to get it started.

3. How do specific principles and parameters fit into the Chomskian model?

3. How principles work is really quite simple. Parameters are a bit more tricky. Most of this was answered above, but I will reiterate here. Both of them are automatic. As soon as the child is exposed to input, the knowledge they have in the form of principles and parameters starts to work. Principles are simply applied universally and unconditionally to all instances of language they hear and produce. As soon as they start to hear their language they will use principles to help interpret it. The exposure to real language is a necessary catalyst or a trigger. While parameters start to work from the beginning by constraining possible forms in language, children still need their language-specific values to be set and the only way to do this is for the child to be exposed to the language. So, exposure to the language or input is a must. Peripheral (lexical) learning accompanies linguistic exposure and drives parameter setting. The former, while only marginally a part of UG, is absolutely necessary for the process of acquisition to take place. If a child were, for some reason raised separately from humans and did not get exposure to a language then they would simply not acquire one. Likewise, two babies raised alone would not end up inventing a new language. They would be mute. Don`t believe what you see in the movies. Prolonged exposure to authentic forms the language at no matter what level

4. What are some of the other major claims about first language acquisition that are critical of the UG approach?

4. There are many other claims about first language acquisition that are critical of the UG approach: The main ones are, the cognitive approach of Piaget, the physiological approach of Lenneberg, behaviorism or more properly the neo-behaviorism of Skinner, and finally the cognitive science approach. All these approaches, including UG of course, are based on Chomsky`s innateness hypothesis. Nowadays nobody argues about whether language learning/acquisition is innate or not. Most researchers no matter what their field believe it is. What people argue about is the nature of what is innate.

Piaget, back in the 1960s, made a big splash with his theory, which simply states that language acquisition (linguistic development) is reliant of cognitive development. He showed that all children, no matter what culture they belong to, go through the same cognitive stages of development in the same way. Once more, these cognitive stages seem to correlate with stages in linguistic development. A similar theory is that of Lenneberg. He pointed out that children reach physical milestones at the same time that they reach linguistic ones. Surely it cannot be a coincidence that children learn to walk and utter their first words at the same time. The correlations, as with Piaget`s cognitive development are very strong. Actually, I don`t see these views as challenging Chomsky directly and neither does he as long as they are received as operating separately. Chomsky is only worried about I-language, so he isn`t worried about this. In the long view, however, what Piaget and Lenneberg say works to put language in a larger more comprehensive system and that is something that Chomsky does not agree with. We will get back to this in a second when we get to the Cognitive Science approach.

The claims of Neo-behaviorists directly challenge Chomsky`s approach. Neo-behaviorists like Skinner originally believed that it is creativity which is innate (remember the unconditioned responses we talked about). Today, however, in response to some of Chomsky`s arguments, they believe that children also have some innate ability to generalize (break down into parts and operations) what they hear. Above I showed how this is still not enough to account for the regularity of what children actually do with language and we will leave it at that, but with a warning that many psychologists still follow this viewpoint.

The final alternative viewpoint is the newest and most promising: it is the view of the Cognitive Scientists. They believe that learning is innate. Language is not a separate system of learning. It is part of one abstract system of learning which is supported and constrained by the actual composition and functioning of the brain. Cognitive scientists embrace the ideas of Piaget and Lenneberg as well as some of the basic ideas of Chomsky and weave them into a system of integrated development where everything is based on this innate system of learning.

5. What is the physical basis for language acquisition according to Chomsky?

5. For Chomsky language acquisition has a definite physical basis. This physical basis is the LAD and the LAD is composed of the principles and parameters of UG. For Chomsky UG is like any other organ in the body, like an arm or a lung and is only different because it is not readily observable. He believes that UG has undergone evolutionary development just like all other parts of the human body. That fact that UG is not physically observable now is not a problem because the Poverty of the Stimulus Argument shows that it must be there. Once we have gotten a better understanding of the brain, Chomsky believes, we will be able to find the LAD.

6. How are the values for specific parameters set in first language acquisition?

6. How the values for specific parameters are set in first language acquisition is one of the most important questions for Chomsky. There are two differing viewpoints. The first view is that the setting is set at neutral and with enough positive input the learner will be able to decide if their language gets the plus or minus setting. Pretty simple. The other view is that the parameter is set to the unmarked setting, which may vary from parameter to parameter. The unmarked setting for the pro-drop parameter is plus because a plus setting means that the language can or cannot have overt subjects. It is more general and, therefore, unmarked. The minus setting is marked because it is more specific and only allows overt subjects. According to this second view, all children would be born with the plus (unmarked) setting for pro-drop. Children in pro-drop language environments (Korean, Chinese, Turkish, Spanish, etc.) need to hear positive input to make sure that there language is definitely pro-drop. Once they have made sure the setting is finalized. Children in non-pro-drop language environments (English, French, German, etc.) are also born wit the plus setting and require positive input to determine that their language is minus and once they have done this the setting is changed and finalized.

Both viewpoints are very similar in that they both assume that the child goes through a period of experimentation where the parameter settings are not permanently set. In fact there is really only one reason to pick one over the other. The second viewpoint is better, I think, because it can be used to explain how parameter settings are made in second language learning. After all, when we go to learn an L2 we already have parameter settings and it has been shown that these settings carry over from the L1 to the L2.

7. How does second language acquisition work differently from first language acquisition?

7. According to Chomsky, from an i-language viewpoint second language acquisition works almost the same as first language acquisition. The LAD and UG have to work in SLA because if they didn`t then learners of second languages would be making utterances that are unconstrained. Research has shown that second language learners still do not create any utterances that violate UG principles and parameters, therefore, the LAD must still be working. Once more, as was shown above, parameter setting works the same way in SLA as in FLA. The main difference there would be the lack of positive input in the SLA process. So, there we have our first main difference students learning a second language are exposed to a very small amount of positive input compared to input learners of a first language are exposed to. The input is often also unnatural and contrived both in its nature and ordering. The other difference is transfer from the L1 to the L2. Transfer can be both positive an negative. That definitely has an effect on SLA, but much of SLA is in the realm of e-language and Chomsky doesn`t need to worry about that.

Here`s more!!

What I want to do here is give you just a brief overview of what we did on Tuesday, September 25. This will basically be a brief summary of how Chomsky sees language acquisition.

Chomsky`s main argument is the poverty of the stimulus argument. The main problem with Behaviorism, as Chomsky sees it, is that it cannot account for what comes out in the form of language. This is the essence of the poverty of the stimulus argument. Skinner`s idea of unconditioned responses does manage to weakly account for the all important creativity issue, but it cannot account for the types of structures that come out of children`s mouths. The basic idea is this. What is going into the children`s heads cannot account for what comes out, especially when we go to look at structure. Children are capable of, and usually do, create novel utterances and even incorrect ones. Adults do the same. All this, however, can be explained at least in general by Skinner, but the problem, as Chomsky sees it, is the sheer volume. Creation in language is not the exception, as Skinner would lead us to believe. It is the norm. Here is an example:

Hieronymous skinned the armadillo and made a boat out of it for his child to play with.

None of you has ever encountered this sentence before. I`m sure because I just made it for the first time. Despite this fact, I am sure you had no trouble understanding it just as I had no great difficulty creating it. Based on this, There has to be some sort of inherent system working which allows us to be creative and comprehend creativity on a grand scale. The input we get as children, according to Chomsky, just isn`t enough to do this. Language is unpredictable and creative by its very nature. The fact of the matter is almost all utterances made are new and unpredictable. Very little of language is actually formulaic. Creativity is not something that occurs only when we are feeling rowdy or rambunctious. It happens constantly and with everybody.

Taking this point further we can find an even bigger argument. Additionally, and this is the big point, if Skinner were right and humans have this innate creativity (and only this innate creativity), then languages would have to be unconstrained. Children would utter sentences like:

A) cat bit for the the dog I.

Such sentences do not exist for a reason that Skinner cannot explain, but Chomsky can. The idea is UG. Few people nowadays questions that fact that language learning is innate. The innateness hypothesis was the one idea of Chomsky`s which psychologists immediately picked up and used as their own. Everyone agrees that something in human language learning is innate, but people disagree vehemently on what that is. Is it just this inherent creativity? It is the ability to break language down into units and chunks? Well, now we are getting close to Chomsky`s idea. Chomsky asks us to compare language acquisition with a factory. Let`s look at a car factory. From the outside all we see is a bunch of different steel, plastic, and rubber parts which are designed by and for the factory. In the end, however, a new car drives out of the factory. We have to assume that something happens inside the factory which turns these raw parts, which on their own are useless, into something useful and very complex.

Language acquisition works the same way. Children are exposed to this language, but the input they get is not very useful in determining structure. Since the language that comes out follows a set of abstract rules which most speakers are not at all aware of, then this input is really quite inadequate. The child must be getting help. This help is the LAD. For Chomsky, then, the idea of innateness revolves around the LAD.

Deciding about what is in the LAD is easy for us because we have already studied this. We know that Chomsky believes that it is the principles and parameters of UG which are in the LAD. Originally, before he had a clear view of UG, Chomsky said that the LAD was a black box. It was unclear what was inside but over time, but looking at what comes out we can figure out its composition. Well, Chomsky has done just that. Principles of UG (structural dependency, etc..) start to function as soon as the child is born. Parameters are a different story. We said before that languages have different values for parameters. Obviously a child cannot be born knowing what their language is like. Language is biologically programmed in the form of UG, but individual languages are not. Any baby learns any human language in the same way and at the same basic pace. It takes a child a certain amount of time to learn the parameter settings for their particular language. interestingly, since first language acquisition is the same for all languages, exactly what values the language has does not affect acquisition at all. Next week we will go deeper into the parameter issue. For now it is enough to know that parameters are part of UG even though it takes the child a fairly long time to straighten them out. This brings us to a related issue in acquisition.

UG is simply not enough for language acquisition to take place. There has to be something more. We have often talked about the role of the lexicon in language learning. In Chomskian theory the lexicon has grown more important in each new version of the theory. Based on that, it seems impossible to talk about language acquisition without mentioning lexical learning. Lexical learning is the trigger for UG and certainly part of the LAD has to account for it. Much of what we know about our language is encoded in the lexicon. Therefore, it is impossible to try to use or speak a language without lexemes (words) and their corresponding grammatical information. Chomsky wants to separate lexical learning from UG because the former is very much language specific while the latter is universal and works the same way in all languages. If we take this further, we can say that the principles of UG are there to ensure that lexical learning proceeds well. Indeed, lexical learning would be impossible without the principle of Structural Dependency and the Projection Principles as well as all the phonological UG principles which we will not get to cover in this course..

Some people (Wexler and Manzini) have even gone so far as to claim that parameters are also part of lexical learning, but they are missing a crucial point. Yes, parameters and their settings in individual languages are reliant on lexical learning, but all languages share the same parameters. Lexical learning is all about the acquisition of language-specific data; like lexemes and grammatical information about the use of each lexeme. To claim that parameters are part of lexical learning is to say that the parameters themselves are language specific, which they are not. Based on this, we can reject at least that part of their theory. Talking about lexical learning brings us to one more point and it will be the last one here. We have talked a lot about the core and the periphery and it is probably important to know how lexical learning fits into that. This is a tricky subject basically because the term periphery is used more or less like a dumping ground for everything that is not core.

Chomsky believes that any theory of language acquisition has to account for the system of language (the core). That is his main goal. He wants to account for the fact that all languages are so similar and that children acquire all languages the same way. That is the core. There is, however, much more to it than that. In order to figure this out we have to make one more distinction and that would be between linguistic and pragmatic knowledge. Under Chomsky`s system linguistic knowledge (lexemes, grammar points, etc.) whether it is core or not is part of lexical learning while pragmatic knowledge (knowledge about how to use the language) is learned somewhere else. Pragmatic knowledge is not part of Chomsky`s language acquisition model. It has to be learned in some other component of the brain.

To sum up, language acquisition, according to Chomsky, is basically controlled/determined by the LAD. The LAD consists of the principles and parameters of UG. The principles of UG work to ensure that lexical learning takes place, which is necessary for the setting of the values of the parameters and fuels language acquisition in general. Additionally, there is the learning of pragmatic/socio-linguistic knowledge of the language. This occurs elsewhere in the brain and is not a part of language acquisition, according to Chomsky.