Universal Grammar (Chomsky)

Universal Grammar is the answer to the ‘logical problem of language learning’: how do young children, with their limited cognition and limited and ‘messy’ input, manage to learn their L1 so fast and so effortlessly, producing output that they cannot have heard in input?

The cognitive aspect is interesting: young children think in concrete terms and cannot deal with abstract terms, yet they can apply the seemingly abstract rules of a language.

Children also acquire their L1 regardless of their environment. They may take different lengths of time, but as long as they have no specific learning difficulty, all children will acquire their L1.

UG theory says that humans are biologically endowed with the ability to learn language. Learning a mother tongue is what infants do; they can’t help it (a bit like spiders spinning webs).

UG is the initial state of the L1 language learner. It is a genetic blueprint, made up of principles and parameters that guide and constrain language learning, thus making language learning much easier. Principles are what all languages have in common; parameters are areas where these principles vary. This means that language learning is highly constrained in advance, making the language learning task much more manageable for the child.

A good analogy of principles and parameters is driving on one side of the road. The principle is that we drive on one side of the road. The parameter is left or right.

One example of a principle in universal grammar is that language is structured into phrases, such as noun phrases, verb phrases and prepositional phrases (structural dependency), and these phrases are dependent upon each other for meaning. This principle applies to ALL languages. The learner does not need to learn this principle because it is part of the UG blueprint.

[Language is understood as phrases because phrases (not words) carry meaning, so it’s not surprising that children learning their L1 (which is initially orally) process language as phrases and not words. This also gives more credence to the Lexical Approach?]

Related to this principle is the head parameter. In a noun phrase, the noun is the head: e.g. in the NP, ‘the boy in the striped pyjamas’, ‘boy’ is the head; in a verb phrase, the verb is the head: e.g. in the VP, ‘ kissed the pretty girl’, ‘kissed’ is the head; in a prepositional phrase, the preposition is the head: e.g. in the PP, ‘on the cheek’, ‘on’ is the head. (the VP and the PP also contain NPs, known as ‘complements’).

However, the position of the head in the phrase can vary between languages. In English, the head goes at the start of the phrase, whereas in Japanese, the head goes at the end of the phrase. So we can say that some languages take a ‘head first’ parameter and some languages takes a ‘head last’ parameter. So the only thing the learner needs to work out is whether their L1 is a head first or a head last language.

So children don’t need to work out that their L1 is structured into phrases because it’s in their language blueprint. They also don’t need to work out that these phrases have a consistent word order related to the head, because that’s also in their language blueprint, but they do need to know whether the head comes first or last in a phrase. In theory, this last bit can be worked out with very little input.

Most western Romance and Germanic languages, including English, are head first. Most languages from north-east Asia (including Japanese and Korean), languages from the Indian sub-continent, Turkish and Basque are head last. However, languages tend to be on a continuum from ‘strongly head first’ to ‘strongly head last’, but the first-last dichotomy is more useful for the L1 learner.

Another example of a principle is that language consists of ‘syntactic categories’, such as determiners (the, a, some, my), genders and inflections (word endings to show meanings, e.g. -ed, -ing). Some languages have these, others do not, so the parameter is yes or no. If the parameter is ‘no’, then this syntactic category can be ignored. In English, we don’t have genders for nouns, so the L1 learner of English can drop this category.

Another principle is that phrases in all languages are made up of specifiers, heads and complements: e.g. my mother’s holiday in the Caribbean is a noun phrase made up of ‘my mother’s’ (specifier) ‘holiday’ (noun) and ‘in the Caribbean’ (complement). However, different languages put the specifiers, heads and complements in a different order: this order is the parameter, and learners need to work it out.

Another principle concerns reflexives: e.g. Mark wanted Tom to kill himself. Who does ‘himself’ relate to. In English, we know it’s Tom because in English we know that reflexive pronouns refer to the nearest preceding noun (local binding); however, in some languages such as Chinese, ‘himself’ can refer to Mark or Tom, depending on the context (long distance binding). So the parameter here is either local or long distance binding. (I guess the principle is that reflexive pronouns are bound to a preceding noun).

Another parameter is called the pro-drop parameter. In all languages, you can replace a noun with a pronoun (the principle). In some languages, you can omit the pronoun from the phrase, while others must have the pronoun. In Italian, Chinese and Arabic, you can drop the pronoun (pro-drop languages), while in English, French and German, you must have the pronoun (non-pro-drop languages). For example, in English you have to write ‘she eats the lemon’ rather than ‘eats the lemon’).

These principles and parameters constrain the task of learning the L1. All that a child needs to work out is which parameter their L1 takes, but because parameters have only a few options (e.g. head first or head last) it is fairly easy. Principles and parameters therefor make the task of learning the L1 much more manageable.

Take the sentence:

‘The Brighton train leaves London at five.’

As a form of input, it allows several parameters to be set:

  • SVO word order: it tells you that the subject goes before the verb and the verb goes before objects
  • it tells you that prepositions go before nouns (‘at five’)
  • it tells you that the language is non pro drop

The Minimalist Program

In the mid 1990s, Chomsky revised his theory of universal grammar, saying that language learning is concerned with the properties of vocabulary (i.e. parameters are linked to the lexicon, especially its functional features – e.g. determiners, connectives and morphemes). This means determining which parameter a language takes depends on knowing some of its lexicon, e.g.

[animate subject] gave [direct object] to [indirect object]

The girl gave the book to the teacher

NOT The rock gave…..

The implications of the Minimalist Program for teaching: “words should be taught…as items that play a part in a sentence by dictating the structures and words they may go with in the sentence.” In other words, vocabulary drives language, words dictate grammar and vocabulary comes in chunks, not as isolated items. This seems to reflect the Lexical Approach.

Universal Grammar and Second Language Acquisition

A key question now is do learners of a second language use Universal Grammar? In other words, can they apply the parameters to a second language, something which must involve resetting the parameters to the L2? If they don’t reset the parameters, L1 parameters may often get in the way of learning the L2. This may mean learning an L2 requires the use of cognitive skills to overcome L1 ‘interference’.

From the perspective of UG, the L2 learner is faced with one major problem: the need to reset parameters when the L2 has different parameters to their L1. Someone who has English as their L1 has their head parameter set to head first; if that person is going to learn Japanese, they need to set the parameter for the L2 to head last. Is this possible? Can UG parameters have different settings for different languages? Or does the learner have to use their cognitive skills rather than UG to learn where heads are placed in Japanese? Or are L1 parameter settings also applied to the L2, meaning the learner has to subsequently revise the parameter settings according to input. Does the L2 learner even have access to UG? Does the older L2 learner with their more developed cognitive skills, even need UG? Maybe UG is only needed by infants because they have lousy cognitive skills.

Research findings are pretty ambiguous and conflicting theories abound with some academics saying UG is only available for the first few years of a child’s life (the critical period).

Criticisms of Universal Grammar

It’s a very clinical, abstract theory, examining the mind rather than the learner, not taking into account sociolinguistic aspects of language learning, especially learner variation or motivation.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s