Our guest lecturer will be BobHale.
—Ceci n'est pas un seing.
I'll be starting tonight with a quick piece about what we mean by "language acquisition". After that I'll post about some of the various theories so you should get to know a little about those names like Chomsky and Krashen that pop up from time to time in other threads.
I don't want to rush at this and I'm a bit pushed for time this evening.
So I'll give you a quick list of the things that I'll try to cover.
1. What do we mean by "language acquisition"?
2. What theories are there about how we acquire our first language. For example Chomsky's contention that we have a kind of built in grammar that gets activated by the language that we hear. This is based on the idea that our productive skills far outstrip, even at an early age, the total number of sentences and structures that we could possibly have heard.
3. Cognitivist theories of language acquisition. Treating the mind as a kind of computer.
4. Functional/pragmatic theories. Language acquisition is related to the need to communicate and the environment in which that need exists.
5. Input, Interaction and Output Theories. Focussing on the interactive aspects of Second Language Acquisition.
6. Sociolinguistic theories. The effect of external factors and the sociolinguistic setting on the achievement of the learner.
7.Socio-Cultural theories which see language learning as being a socially mediated problem-solving process.
That's all I have time for now.
Next time I'll do a quick definition and have a look at first language acquisition.
This looks great, Bob! I'm looking forward to it.
Right then, let's get started.
Before I begin let me list my reference sources so that I don't need to go through the business of referencing everything in the subsequent text.
The following books are the main references used.
Rod Ellis – The Study of Second Language Acquisition, OUP 1994
(This is a very thick very detailed book and probably not a good starting place for the student.)
Rod Ellis – Second Language Acquisition (Introductions to Language Study), OUP 1997
(A thin book with a readable overview but not much detail.)
Mitchell and Myles – Second Language Learning Theories, Hodder Arnold 1998
(Quite a good book for new learners.)
Lightbown and Spada – How Languages Are Learned,OUP 1999
(quite similar to Mitchell and Myles)
OK references out of the way.
What exactly do we mean by language acquisition and is it different to language learning? Ellis, after a lengthy explanation, decides to treat the two terms interchangeably but I prefer to follow Chomsky who sees the difference as being between "naturalistic" acquisition and "instructed" learning. That doesn't mean that for a typical learner –especially in ESOL – both processes aren't happening simultaneously, but it does, to my mind anyway, preserve a useful distinction. "Acquisition" means getting your language by whatever process while "learning" is receiving instruction. To give an example. If I tell my students that the present continuous is formed by the simple present of the verb "to be" plus the present participle and they write that down and remember it, that is learning. If, on the other hand, they hear people say "I am coming" or "he is talking" and copy it or adapt it (perhaps even incorrectly) that is acquisition.
Some theorists make this distinction. Others don't. You pay your money and you take your choice.
Most theorists agree though that we acquire our first language (or in the case of bilingualism, languages) in a way that is quite different to the way that we acquire subsequent ones as adults. Infants receive, at best, only minimal formal instruction – pointing to a picture and saying "doggy", for example.
One thing that is generally agreed upon though is that the sequence in which children learn appears to be well defined. To quote Ellis
"Children typically begin with one-word utterences which function as holophrases (i.e. express whole propositions). They gradually extend the length of their utterances, passing through stages when the bulk of their speech consists of first two-word, then three- and four-word utterances. At the same time, they systematically acquire the various syntactical and morphological rules of the language."
Now there are quite a few different theories as to what's going on to allow the child to do this – probably at least as many as there are theorists. Chomsky, for example, (and I hope the knowledgeable here will forgive an insane oversimplification) suggests that we have a kind of built in Universal Grammar and that learning your first language consists of flipping the switches and setting the parameters. In essence, you can't NOT learn it. Your brain was built for it.
A more common theory is that everything is learned by imitation of what is heard. The trouble with this, is that even at a relatively young age we can produce vastly more syntactically correct utterances than we ever could by imitation alone. At the very least there would need to be a cycle of imitation, inference, substitution, error correction.
To me even that seems to be inadequate.
Personally I confess that I've read the theories and am really no wiser now than I was before. The way that children acquire the full range and depth of their first language is a mystery to me. It's a black box with everything they see and hear as inputs and comprehensible language as the output. Or maybe it's magic.
Next time – Adult language acquisition: Cognitive Theories
Since I'm most familiar with Chomsky I hope you won't mind if I add a bit. I hope I get this right. Chomsky's argument is known as the poverty of the stimulus: children's input contains many ungrammatical utterances. How do children know which input to pay attention to and which to ignore? So to take an example from English, Jack, it would make sense to assume that given input like this:
She sent a letter to him.
She sent him a letter.
Children should generalize and produce utterances like this:
She explained the letter to him.
*She explained him the letter.
Children get very little if any negative evidence - evidence about what is ungrammatical.
Chomsky's answer is that we have an innate linguistic capacity, which he calls the language acquisition device. In other words, a lot of the rules are already built in. The structure we are born with is called Universal Grammar, and when we acquire a language we are comparing the input to UG and seeing what fits, and changing certain settings to produce the actual grammar of whatever language we're acquiring.This message has been edited. Last edited by: goofy,
They don't? I thought children made mistakes like this all the time.
OK, the observation that children and adults acquire language differently is pretty trivial: as you noted, I don't spend my day the way an infant spends his day. But if I could spend my day the way an infant does, with no responsibilities and an attentive Mommy talking to me all day, would I learn language like an infant does?
Studies (I'm at work so don't have the specific data to hand) suggest not. It seems that initial language acquisition is a one shot deal. Once the parameters for your first language are set they are set forever. I'll try to look out the specifics tonight.
Good point. They may make mistakes like this. But it does not end up being part of their grammar. How do they know?
Children overgeneralize plurals (mouse - mouses) and past tense (break - breaked) and then later correct when they learn the standard forms. But in a case like this, they will not receive any evidence that their mistake is ungrammatical. So the theory goes.
Universal Grammar confines the hypothesis space - it limits the choices a child has to make in constructing their grammar.This message has been edited. Last edited by: goofy,
I haven't found a citation lately but some years ago I read that during Charles Berlitz's childhood each of his various adult relatives spoke to him in a different language. He allegedly said, "I thought everyone in the world had his own language and I kept wondering how and when I would discover my own."
My kids are older now, but in remembering their acquisition of language, they rarely made grammatical mistakes. They had trouble with pronunciations, and at the beginning with putting together words (between 5 and 18 months), but later on (2 years) they put together grammatical sentences. I always thought the reason was that they were modeling what they heard. That, I have always thought, was similar to the ways adults learn. That's why most adults I know who have done well in learning a language have done so when put into the environment. Shu, for example, spent 6 months in Israel when he was in college. He was forced to speak Hebrew and French and to this day has great command of both those languages. When we went to Paris a year and a half ago, having not spoken French for years, Shu still was able to comprehend French and to speak it quite fluently. I was amazed at his command of the language.
I remember my 8th grade daughter going to France to visit a family after a year of middle-school French. By the time she returned home, having not spoken an English word for a month, she was dreaming in French. I realize these are isolated examples and therefore don't support my theory, but I've been skeptical of Chomsky's view. Is it commonly held within the linguistic community? If so, does this community see a difference between babies learning a language, versus young children, older children, or teenagers? Does Chomsky acknowledge that there may be some variance with his theory?This message has been edited. Last edited by: Kalleh,
It is a widely held view, but there are many linguists who disagree. Personally, I have yet to encounter another theory that explains how children acquire grammar without negative evidence.
Chomsky's theory has nothing to say about how adults learn language. There are differences between children's and adult's language acquisition/learning, I'm sure Bob will talk about them.
It is a widely held view, but there are many linguists who disagree. Personally, I have yet to encounter another theory that explains how children acquire grammar without negative evidence.
I studied linguistics at Cal, which at the time was a hot-bed of anti-Chomsyan (or more correctly anti-generative grammar) sentiment. So I was reared academically to have a knee-jerk negative reaction to Noam's theory. There was and still is much disagreement with Chomsky's theories (which some parts of have changed over the 50 years or so that they have been under development). My intro syntax and semantics class was taught by George Lakoff. There were about 35 people in attendance, not all of whom were undergrad linguistics majors such as myself. We had at least 7 post-docs / professors who were auditing the class. Then there was a sizable contingent of computer science students for whom the class was a requirement. (The former asked Lakoff questions which were far over our heads, and the latter were a sullen and angry bunch who couldn't quite fathom why their department forced them to take this social science class.) The summer after I graduated I attend the LSA summer school in Salzburg, Austria, and wouldn't you know it my roommate in the dorms was a Japanese grad student at MIT, and Chomsky was his advisor. I got a crash course in pro-Chomsky thinking, and survived the experience. (I also took a class in diachronic syntax from David Lightfoot who was (and still is) about as fervent a generativist of the Chomsky camp as one can get.) Years later, a dear friend of my wife's who was born and grew up in Philadelphia, informed me that her family used to attend the same temple as the Chomsky family. The thing that caught my ear was that she (and they) pronounced the name /'xamskiː/ (see the IPA thread) rather than /'tʃampskiː/.
I have always felt, without much evidence, that much of the mechanism for language acquisition was based on the more generalized pattern matching skills of humans.
—Ceci n'est pas un seing.
I'll be coming to adult acquisition in my next lesson ( ) but let me make a couple of further comments/clarifications about infant acquisition.
As others have commented there are pro- and anti-Chomsky schools of thought. I'm cautiously pro because I can't think of another mechanism that would allow the vast array of "wrong" sentences to be identified. A pattern recognition mechanism would allow you to identify the correct sentences but you are never exposed as a child to the considerably larger number of wrong ones.
Nobody comes along and says "I is go to bed now" accompanied by a slap to show that it's wrong. And yet we can identify wrong structures as unerringly as we identify right ones. Something is going on and until someone comes up with a better idea about what it is then "Universal Grammar" is as good a model as any. Or "magic" as I like to think of it.
My second comment will be dealt with fully later but essentially it's a trivial observation that adults can't possibly learn their second languages in the same circumstances or by the same method as their first because they already know their first language. Anything that they learn about L2 is through the filter of L1. They have something to compare it to. To give a trivial example when a child first learns that the brown hairy thing in the corner that says "Woof" all the time is called "dog" she is associating the thing with the token. When, later she visits France and learns that it's now called "chien" she is associating a new token with the old token. She doesn't need the actual animal to be present. Same goes for grammar. Learning a second grammar has the advantage (or perhaps disadvantage) that you have an old grammar to map it to
To give another analogy. Imagine that you have led a sheltered life and never learned to cook. If I put you in a kitchen with all the ingredients and equipment required to make a shepherd's pie and let you play eventually you will probably come up with something that's quite similar to the ones you have previously bought in the supermarket. If I now take away those ingredients and give you the ones needed for a rhubarb crumble it's unlikely to take anywhere near as long as you have already discovered some of the principles - you have your L1 - "shepherd's pie" - and can compare your L2 - "rhubarb crumble" - to it rather than discovering everything again from scratch.
A faulty analogy for all sorts of reasons but I like it.
More later...This message has been edited. Last edited by: BobHale,
So then , to adult learning theories.
The first group is called "Cognitivist Theories" and can be broken down into Processing Theories and Connectionist Theories. There is a pretty good treatment of all of this in Mitchell and Myles.
Processing Theories liken the brain to a computer and the acquisition of new information, including languages, to writing a program.
McClaughlin proposes an information processing model which sees the mind as a computer and complex behaviour and knowledge as being composed of simpler processes (like subroutines) and views the whole shebang as modular. Each module is independent and as they can be studied independently and as they take finite time to process estimates can be made about the overall capacity to acquire new knowledge.
We learn words and rules and apply the rules to the words to create meaningful utterances. Repeated use of these utterances turn them into well-defined formulae that are placed in long term memory.
One thing that this model explains well is a phenomenon known to every second language teacher in the world – fossilisation. That is to say, the situation where – despite frequent repetition and correction – students are unable to get rid of certain productive errors. To give an example, I have been marking papers and several students are still – after three years of contrary instruction – producing the form – subject + present to be + infinitive without to as a past tense form: "Last week, I am go to London." McClauglin explains this as a rule that has prematurely been placed in long term memory, perhaps by inappropriate analogy with the students first language, and which cannot now be easily dislodged.
Personally I'm not over-enamoured of this theory as it seems to me the reduce language to a sequence of set formulae and my own observation is that although some chunking goes on and there is certainly fossilization, the range of student speech production is far too complex for such a simplistic model.
Anderson has a different model – ACT:Active Control of Thought. This proposes three stages of acquisition.
Cognitive: we learn facts about language – declarative (or descriptive) knowledge. For example we learn the rule that present perfect is formed from present of have + past participle and is used for actions starting in the past and continuing now or having consequences now.
This is followed by Associative: where we seek to discover a technique that will allow us to reliably apply the declarative knowledge and then by Autonomous where we produce the structure with no deliberate effort.
Note that even though this model doesn't explicitly mention computers , it does implicitly correspond to a programming, testing, running sequence.
It does well at explaining errors such as failure to apply a third person –s to verbs. The rule has been learned as declarative knowledge but has failed to become procedural and needs to be consciously activated each time.
Another theory is Pienemann's "Processability Theory". I don't really know very much about this one and although I've tried to read up on it I'm struggling. Roughly, as far as I can tell, it separates a Chomskian style "Lexical Functional Grammar" from associated processing rules. Language is seen as a set of built in rules and the processing as learning how to exchange lexical items for the ones in the rules to produce valid sequences. As I say, I found the descriptions in various books to be rather confusing. Perhaps someone else can help out on this one.
Piennemann also proposes a "Teachability Hypothesis" which I understand rather better. Rather like the first language acquisition research it suggest that language acquisition generally takes place in a quite rigid sequence and cannot easily take place unless that sequence is maintained. You cannot learn to produce the later structures until you have fully integrated the earlier ones into your Lexical Functional Grammar.
There's lots about it in Mitchell and Myles. Good luck with that.
One last bit that ought to be mentioned is connectionism. This views everything in terms of neural networks. The mind is a series of nodes and connections between those nodes like a vast net. Information is represented as nodes and processes as connections. The more you encounter a particular structure (process) the more reinforced it becomes and the stronger the connection. All the connections are initially there but failure to encounter forms weakens the bad connections. This is a kind of pattern recognition. The good patterns are encountered, the bad ones aren't and each time a connection is strengthened. The stronger the connection gets the more automatic the production of similar patterns becomes.
This explains well "over-generalisation" errors taked instead of took, for example.
I need to do more reading on this as the books I've consulted give it only a very superficial treatment.
Coming Next Time: Pragmatic and Functional Theories
Interesting. From your explanation, to me the most plausible theory seems to be the ACT model.
Thanks Bob. This is really interesting.
Sorry for the delay. And sorry for the fairly scrappy note form of this entry. Time has been too tight to organise it better.
So, on to Functional and Pragmatic theories.
At it's simplest level the clue is in the name. Very broadly speaking pragmatic theories of language acquisition say that students learn easily things that are useful to them and less easily things that are less useful to them. The order of acquired elements of language is determined by the need to convey a meaning in a real life situation.
This of course has an impact not only on theories of how language is acquired but on how language is analysed. To give an example, "The fat lady sings" would conventionally be analysed as a subject, in the form of a noun phrase, and a verb. Pragmatically speaking it would be analysed as a description, and agent and an action.
Before I go on I need to define one particular term that is important in language acquisition – interlanguage.
Interlanguage is the characteristic development of a second language influenced both by the learned grammar of the new language and the previously known grammar of the first language. It is dependent on the languages in question and the stage of development the learner has reached but also has some characteristics that are quite general.
An example of how interlanguage develops and changes can be seen in the typical patterns of negative formation usage among learners. English is quite useful to look at in this respect because our way of making sentences negative is quite unusual. Other languages commonly use a negative marker of some sort. Je mange, je ne mange pas. Ich esse, ich esse nicht.
English is unusual in that we negate all verbs except the verb to be with the use of a negative auxiliary. I eat. I don't eat. This isn't simply a negative marker as there are two different positive forms, I eat, I do eat. The latter is used only to answer a negative question or to add a specific emphasis.
Very interesting. But what does this have to do with language acquisition?
It explains the use, by lower level learners, of phrases like "I not work" or "I work not" using not as a negative marker, often in imitation of the way their first languages mark negatives. Some of these interlanguage structures are notoriously hard to shift. One such, from my experience, is, among Indian students, the use of the present continuous with a past time marker – "Yesterday I am going to the market." Despite repeated correction this often persists throughout even quite high levels. Why should it persist? Because, on a day to day level, it may mark them out as non-native speakers but it will be understood. Once that incorrect structure has been incorporated as part of the student's interlanguage there is now social or functional need to replace it with the correct "Yesterday I went to the market."
Givón suggests that there is a distinction to be made between pragmatic and syntactic modes of expression and that there is a continuum between them. Learner speech has a heavy reliance on shared knowledge and context whereas more formal modes of expression rely on syntactic structures to carry meaning more explicitly with little if any input from context being required. Language acquisition, he argues, is a process of moving from one end of this continuum to the other, replacing the "defective" pragmatic structures with the "accurate" syntactic ones.
Another typical pragmatic structure is the topic/comment sequence in discourse which leads to another commonly seen error in presentation – double subjects – which is also very difficult to shift.
Sentences like "My daughter, she is six." and "The college, it is very good." litter the work of most ESOL students, right the way up to Level 2 (the level required to take the IELTS University entrance test.)
This topic comment sequence is extremely common among second language learners and appears in most students interlanguage at some stage in their development.
Most of the other discussion that I have found in the books for this particular model discuss specific, and often very limited, case studies. Once again Mitchell and Myles has a pretty good section.
Language acquisition is context driven.
Development of second language skills goes from the purely functional (often topic/comment) forms towards more structured syntactic forms.This message has been edited. Last edited by: BobHale,