Didactics - 12 : Overview

We have spent the last few weeks looking at the work of Stephen Krashen, and through examining his five hypotheses, we have traced out an overview of what is known about how a Second Language is acquired. Let us go back through the five hypotheses once more.

1. The Acquisition/Learning Hypothesis

Krashen distinguishes between acquiring a language and learning about a language. The former occurs without our paying any attention to the process, whereas the latter is conscious. Furthermore, he sees acquisition as the real road to mastery of a FL. Therefore, teaching about the language is of only marginal utility - it enables learners to develop what he calls a Monitor, which can only be helpful to them in certain limited circumstances.

There are a number of criticisms that can be made of the basic distinction. Anderson, in his model, makes what appears to be a similar distinction - between Declarative and Procedural knowledge. However, according to Anderson, the differences are between two stages in the assimilation of knowledge, rather than between two totally different forms. Knowledge is first of all taught, and rehearsed, as a set of rules - this is declarative knowledge. Then, through assimilation to already existing knowledge, the new behaviours become routinized as procedural knowledge, and the rules themselves may be forgotten, and no longer available for recall. That is why I think that I simply absorbed my mother tongue - I have forgotten the rule-learning stage.

Gagné offers a rather similar model, only there is less stress on the formal presentation of rules, and rather more on active participation in rule-getting.

1. The learner is offered a model - this may be through watching an expert perform, or through reading a book or listening to a talk, etc. The learner constructs hypotheses about how the skill is accomplished.
2. The learner actively attempts to carry out the task, while aided by an expert - this is 'cued performance'. It gives the learner the opportunity to verify the hypotheses that she has constructed.

Thus we may explain to a learner how the 'have -en' form of the verb is used. She listens to our explanation, and forms her own ideas of when the form is or is not appropriate. She will then try them out to see if they work.

Neither Anderson's nor Gagné's model sees language as being different in kind from any other skill which we learn, whereas Krashen's hypothesis is based upon the double assumption that

a) language acquisition is facilitated by a specific pre-programmed LAD.
b) the LAD is in some way still available for the acquisition of a second language

Neither of these assumptions is accepted by all observers.

K's first hypothesis, then, points us to some consideration of the mechanisms by which learning is accomplished. We will have to return to these matters when we consider cognitive development.

The Natural Order Hypothesis

Linguists working on the acquisition of L1 by children have discovered that the children appear to go through a series of stages, acquiring the different grammatical morphemes in regular and predictable sequences. Krashen holds that the same is true of SLA, and in particular leans upon the work of Dulay and Burt, in their analysis of acquisition orders. This order is not displayed by the learner under all circumstances - thus, for example, the 's' of the third person singular, which is a fairly late acquisition in the natural order, may be placed correctly when the learner is carrying out a task such as a grammar test, where her attention is directed to linguistic form. But when she speaks in order to communicate, the 's' will not be present.

There are several problems with this research.

1. Dulay and Burt's samples were not longitudinal, so that we cannot say that the same learners exhibited different behaviours at different stages of their development.

2. The morphemes themselves do not form any linguistic unity, so that there can be no unitary hypothesis to explain why they should follow in sequence.
3. The research takes no account of the fact that there may be considerable variation in the behaviour of any language learner, so that a morpheme that is present today may disappear tomorrow.

4. There are studies which have not found any such order, or which have found different orders. The order may depend to some greater or lesser extent upon the situation in which the learner acquires her L2.

Perhaps more interesting is the well-established observation that particular grammatical structures are acquired in stages. The example that we looked at was the acquisition of the English negation, which appears to follow a very similar path for learners, no matter what their L1 might be, and which is very similar to the sequence observed amongst L1 learners of English.

Another interesting finding is that of implicational hierarchies. According to some linguists, when we compare different languages, we discover that we can make predictive statements such that if language A contains feature X, then it must also contain feature Y. Thus Hawkins tells us that 'if a language has a noun before a demonstrative, then it has a noun before a relative clause.' The reverse is not true.

For some features, this kind of relationship can be organised hierarchically. Take the example of the relative clause. In English, this can function in a number of ways :

Function	Example
1. subject	The man who lives next door ...
2. direct object	The man whom I saw ...
3. indirect object	The man to whom I gave a book ...
4. object of proposition	The man about whom we spoke ...
5. genitive	The man whose wife had an accident ...
6. object of comparative	The man that I am richer than ...

(from Ellis, The Study of Second Language Acquisition, OUP, 1994, p. 102)

Now, it would appear that a language that allows the construction of a relative clause of type 6, also allows the construction of all the other types. But a language that does not allow the construction of clause type 3, will also disallow types 4, 5 & 6. But if 3 is allowed, then it follows that 1 & 2 are also possible.

It has been observed that when we teach English relatives, if we present one of the rarer forms first, the learner will *fill in* the unlearned forms automatically. This implies that it is not always the case that we should go from the simple form to the more difficult one, and that in some circumstances, it is more rational to teach the more complex form, and to leave the learner to work out the simple one for himself.

Krashen's Natural Order hypothesis implies that teaching can do very little to change the order. To some extent, his contentions are backed up by the research - although it should be pointed out that the research is neither as extensive nor as robust as some applied linguists have implied. Nevertheless, it does appear to be the case that implicit instruction - even if it does not change the order - may help the learner get through the stages more quickly.

Another argument that K extends from the Natural Order Hypothesis is that ease of learning is not the same thing as apparent simplicity or complexity. Something that may appear to be difficult may in fact be absorbed without much trouble, whereas a language feature which is simple to explain may be difficult to acquire. Too many grammar programmes are based upon whether the points covered are easy or difficult for the teacher to formulate, rather than on whether they are easy or difficult for the learner to acquire.

The Natural Order hypothesis led us to a consideration of how we should characterise the language produced by FL learners. We saw that most observers today refer to Selinker's concept of interlanguage. According to Selinker, we should think of the learner not as someone who fails to speak the language he is learning, but as someone who successfully speaks the interlanguage that he has constructed. What we perceive as errors are, in fact, evidence of the fact that the learner is not using the same grammar as the native speaker. The question for the teacher then becomes one of knowing how to move the learner from one grammar to another.

The dangers that await the learner are considerable. Selinker points out that the majority of adult learners never achieve full fluency in L2. Almost all of them fossilize at some point or other. Fossilization may be induced by teaching methods - if a learner is forced to produce forms that are beyond his present IL, then these may fossilize, and prevent the learner from making any further progress. On the other hand, the learner may fossilize as the result of an economic judgement, in which she decides that the added effort needed to bring about fairly marginal increases in mastery are not worthwhile.

Selinker suggests that the learner constructs his I/L through the employment of a series of learner strategies, which include :

Transfer - the use of the L1 to construct a model of the L2. While this was once believed to be mainly negative in its consequences, there is now agreement that the adult learner may use the L1 positively, and that some comparison of the two languages is useful in language learning. It is no longer considered realistic to insist that the learner should have no recourse whatsoever to the L1.

Overgeneralization - the learner applies the rules of the L2 to situations in which they are not used by native speakers. (In fact, the strategy itself should be referred to as 'generalization', which is a fundamental way of acquiring or constructing knowledge. It is only when the process leads to erroneous conceptualizations that one can properly use the term 'overgeneralization').

Simplification - the learner often begins by utilising a simple subset of language rules. However, this is not so much a simplification of the TL - one cannot simplify what one does not know - but a retreat to a simplified grammar very similar to the basic rules of many pidgins. Some linguists believe that this is in fact a set of rules given by the Universal Grammar.

This view leads us to consider the question of what it means to make an error when speaking an L2. As we have seen, some errors may occur as a result of transfer - interference - while some errors may result from the other two processes. Generally, we can say that the making of errors appears to be a natural part of the learning process, and is not to be understood as evidence of perversity on the part of the learner.

It is likely that the errors which are the most difficult to treat are those that are over-determined - that is, those that are traceable to more than one of the possible causes, such that the different strategies reinforce each other.

One should also note the existence of 'avoidance errors', which occur when the learner avoids a particular structure or set of vocabulary items, because he or she is not sure of his capacity to use them correctly.

The Input Hypothesis

Krashen states that all that the only way for a learner to make progress in an L1 is for her to be exposed to a sufficient amount of comprehensible input. The learner needs to be provided with input that is just a little beyond her present capacity ; it is in the effort that she makes to comprehend the material that lies the source of her linguistic progress.

There are two kinds of problems that arise from this hypothesis ; the first is to define what is meant by the term comprehensible. The second is the question of whether it is true that input alone leads to progress, or whether the learner does not also need to speak or write the language.

We may provide comprehensible input by ensuring that the learner is given enough non-linguistic cues to enable her to decipher the message without understanding everything in the text. However, if we give too much help, this may lead to pseudo-comprehension.. Lydia White, for example, suggests that the input should not be comprehensible, as this makes things too easy for the learner.

Moreover, a number of researchers have argued that output is necessary. It is necessary at a fairly mundane level, because it is the only way that the teacher can judge the needs and progress of the learners. At a more complex level, output is necessary, according to Swain, to Boulouffe, and to White, because it is when the learner is forced to find adequate expression for her thoughts that she will reorganise her language, and make real progress. As Boulouffe suggests, it is only when we have to express our own inner feelings, opinions and desires, that we go beyond the surface of language, and enter into the depths of modality.

The arguments do appear to suggest that output should be meaningful, and that the learner should feel a real need to provide it. It is also necessary for the teacher to react to the learner's language in a meaningful way, and for the teacher to signal those occasions upon which a break-down has occurred in the communication.

The Monitor Hypothesis

Krashen's belief that we should distinguish between learning and acquisition, and that only the latter provides communicative language, does not mean that there is no role for learning at all. The rules of that we learn formally are embodied in the monitor, which we use to check on our output as it arrives in the mind. The monitor, however, is a tool that should be used sparingly. If overused, it can lead to the learner being unable to say anything, because she is too afraid of making a mistake, or to her speaking so slowly that no-one has the patience to listen to her. This, says K, as a pity, because it will deprive her of output. The monitor is best used in situations where we have the time to apply it - formal situations, writing and so on.

Even then, it is of limited use, for the language learner can only apply the formally learned rules. There are a number of reasons why these rules may be wrong. They may have been badly learned, badly taught, badly formulated, or not taught at all.

Krashen's ideas on the monitor have been criticised because they push the monitor to the periphery, whereas studies that have been carried out on how good learners organise their learning suggest that these people in fact employ the monitor quite extensively, and that it is not only useful for ensuring that output is correct, but that it also aids the learner in her acquisition.

The good-learner studies, which suggest that attention to form is an important aspect of language acquisition, and that good learners do monitor their output, and also ask others to correct them, are not above criticism, as we saw. They mainly concentrate on those learners who are successful in classroom terms - and often enough, classroom criteria are exactly those which using the monitor would help ; this is, however, no guarantee that what we look for in the classroom is necessarily what makes for good language acquisition, if we mean by that the mastery of the L2 as a means of communication.

The Affective Filter Hypothesis

Finally, the fifth hypothesis suggests that input alone is not sufficient ; the input must be experienced under conditions which lower the anxiety, and raise the motivation and the self-image of the learner. We saw that all three of these factors are influenced by conditions outside the classroom - some individuals appear to have high anxiety and a poor self-image as part of their basic personality, and these characteristics have been laid down by their upbringing, the ways that their families have treated them well before they arrive in our classes.

Nevertheless, we have also seen that there are a number of things that teachers can do to lower anxiety, to raise motivation and to encourage a positive self-image.

We saw that motivation could be of several kinds. In SLA, some researchers have distinguished between integrative motivation, and instrumental motivation. The former indicates that the learner wishes to get closer to the culture and people of the language to be studied, and the latter that the learner simply needs the language in order to better herself in one way or another. Gardner and his colleagues have suggested that the former kind of motivation is more efficient.

However, there are problems with Gardner's theories ; it is not as simple as it might seem to distinguish between integrative and instrumental motivation, and it is often the case that the learner in fact has a mixture of the two. Moreover, Gardner constructed his model on the basis of his observations of language learning in Canada, a bi-lingual society in which relationships between the two language communities are such that one might expect learners who were positively oriented towards the other community might be more likely to be successful in learning the other language, while those whose orientations were negative might resent the imposition.

A number of researchers looking at language acquisition in other societies, where conditions were different, have found that instrumental motivation can be just as powerful - if not more powerful - as integrative motivation. Neither should one forget the interesting finding that in certain cases, motivation of the Machiavellian kind can be quite sufficient.

It can also be said that such studies, because they view motivation as something that the learner brings with her into the classroom, leave aside what actually occurs in the learning situation itself. We have seen that the tasks that learners are asked to carry out, the way the syllabus is presented, and the way the teacher interacts with her pupils can also influence motivation, so that classes that begin with very low motivation can be persuaded to change their minds. Where learners are offered the opportunity to communicate, where learners are asked to take responsibility for their own learning, where learners are treated with respect, they will acquire higher motivation, and will make more progress.

Conclusion

Krashen's hypotheses were interesting, and provided some insights into the way we might present a foreign language. However, they suffer from under-definition and over-generalisation, and we need to recognise that they cannot represent the last word on language teaching. It is my opinion that they had the merit of directing teachers' attention to the need for varied and interesting input, to the need to take care over such questions as error-correction, meaningful communication, and the provision of a learner-centred classroom that encourages learning, rather than punishing failure.

(If you wish to comment or ask a question, please write to tmason@timothyjpmason.com)