3.2 Early approaches to SLA
3.2.1 Contrastive analysis hypothesis (CAH)
Before the SLA field was established, scholars from the 1940s to the 1960s had conducted contrastive analyses by comparing two languages. They believed that L2 teaching would be more effective if similarities and differences between L1 and L2 could be identified. This belief was best summarized by Lado (1957):
Individuals tend to transfer the forms and meanings and the distribution of forms and meanings of their native language and culture to the foreign language and culture --- both productively when attempting to speak the language and to act in the culture and receptively when attempting to grasp and understand the language and the culture as practiced by natives.
(From Larsen-Freeman & Long, 2000:53)
It was possibly because of the occurrence of transfer that Lado further claimed that ‘those elements that are similar to his native language will be simple for him, and those elements that are different will be difficult’. The assertion that linguistic differences could be used to predict learning difficulty led to the emergence of CAH: Where two languages were similar, positive transfer would occur; where they were different, negative transfer, or interference, would result. Then followed a detailed structure-by-structure comparison of the sound system, morphological system, syntactic system, and even the cultural system of two languages, just for the purpose of discovering similarities and differences.
CAH was heavily influenced by structuralism and behaviorism, which were dominant in linguistics and psychology in USA through 1940s and 1950s. The ultimate goal is to predict areas that will be either easy or difficult for learners. Teaching of a second language under the influence of contrastive analyses at that time was based on the following assumptions:
(1) Language is a habit and language learning involves the establishment of a new set of habits.
(2) The major source of error in the production or reception of an L2 is the L1.
(3) One can account for errors by considering differences between L1 and L2.
(4) The greater the differences, the more errors that will occur.
(5) One who learns an L2 learns the differences. Similarities can be ignored.
(6) Difficulty and ease in learning is determined respectively by differences and similarities between the two languages.
There were two versions of CAH, the strong and the weak one. In the strong view, it was held that one could make predictions about learning and hence about the success of language-teaching materials based on a comparison between the two languages. The weak version gained credence largely due to the failure of predictive contrastive analysis, and then became part of error analysis.
After the appearance of the strong version, criticisms followed. Those who argued against the strong version of CAH pointed out that many areas where predictions made did not appear in actual learner production: predicted errors did not occur. Besides, errors which occurred in learner production had not been predicted by the theory. That is, the theory did not actually predict what was happening in nonnative speech. For example, one difference between French and English is that the object pronoun precedes the verb in French, while in English, the object pronoun follows the verb. However, the fact is that the French learners of English produced I see them instead of I them see, as predicted by CAH (Zobl, 1980). In other words, French learners of English never pre-pose the object pronoun. Rather, they correctly follow English word order.
Another criticism of CAH is related to the concept of difficulty. A fundamental concept of the CAH was that differences signified difficulty and similarity signified ease. Difficulty in this view was equated with error. That is, an error produced by a learner was a signal that the learner was having difficulty with a particular structure being learned. Actually, this is not always so. Let us look at an example from Kellerman (1987), a sentence written by a student,
But in that moment it was 6:00
When asked to comment on the student’s use of the preposition in, she insisted that the in was correct but questioned whether it should be it was 6:00 or it had been 6:00. Obviously, the learner had some difficulty in tense usage even though there was no error reflecting that difficulty. On the contrary, there seemed to be no doubt in her mind about the correctness of the preposition in. From this example, actually one of many, we find that difficulty cannot be simply equated with errors, which is a predicted result of linguistic differences. The difference described by a linguist or a teacher through CA is not a real measure of difficulty in reality.
Although the CAH has some problems, the above discussion does not mean that NL does not play any role in SLA. What is implied is that there are other factors affecting SL development and that the role of the L1 is far more complex than the simple 1:1 correspondence suggested by CAH.
3.2.2 Error analysis (EA)
1. The significance of errors
We have just discussed CAH and its shortcomings. Now we are going to focus on errors made by L2 learners, so as to learn more about second language development. It may appear unusual to focus on what learners get wrong rather than what they get right. However, studying errors can be beneficial to teaching and learning of an L2. First, errors are an outstanding feature of leaner language. Teachers can get valuable information from thinking about why learners make errors. Second, it is useful for teachers to know what errors learners make. Third, making errors may help learners themselves to learn an L2 better when they self-correct the errors they make.
Besides the above pedagogical significance, the study of errors can produce even greater theoretical value. Ever since the publication of S. Pit Corder’s (1967) article on “The significance of learners’ errors”, learner errors were no longer regarded as ‘bad habits’ which should be eradicated, but as sources of insight into the process of learning an L2. According to Corder, errors enable us to “see” the system of language a learner is using at any particular point in the course of L2 development and the strategies the learner is using in his “discovery of the language”. In other words, errors are windows through which we can peep into the learner’s mind. In this approach, learner language is seen as a target of analysis which is independent of L1 or L2, and the state of learner knowledge is viewed as transitional competence on the path of SLA. Further, it was claimed that making errors is significant because it is part of the learning process. As Corder said, it is “a way the learner has of testing his hypothesis about the nature of the language he is learning.” By making errors, the learner is exploring the new system of language rather than just experiencing “interference” from old habits.
2. The procedure of error analysis
（1）Collecting samples
Learner language samples can be gained from both written data and oral data from learners who are responding to the same kind of task or test. Samples can also be collected from a few subjects who are studied over a period of weeks, months, or even years in order to determine patterns of change in error occurrence with increasing L2 exposure and proficiency.
（2）Identifying errors
To identify errors is to determine the elements in the sample of learner language which are wrong in some way compared with the normal or ‘correct’ target language. For example, in the sentence
A man and a little boy was watching him. (Ellis, 2000)
We find the error ‘was’ since the correct form should be ‘were’.
（3）Describing errors
After deciding what is erroneous in the learner language, it is time to classify the errors, which enables us to discover the rule-governed system in learners’ mind. For example, the above misuse of ‘was’ belongs to subject-verb agreement error. Errors can be classified according to general linguistic category such as auxiliary system, passive sentences or negative construction. Also, they can be classified according to specific linguistic elements such as articles, prepositions and verb forms.
The difference between an error and a mistake should be made clear here. Mistakes are akin to slips of tongue. They are only one-time events. The learner who makes a mistake is able to recognize it as a mistake and correct it if necessary. But an error is systematic. It may occur repeatedly and not be recognized by the learner as an error. The learner has incorporated an erroneous form into his own linguistic system. In this way, errors are only errors from a teacher’s perspective, not from the learner’s.
In Corder’s (1974) framework for describing errors, he distinguishes three kinds of errors according to their systematicity: pre-systematic errors, systematic errors, and post-systematic errors. Pre-systematic errors are random, and occur when the learner is unaware of the existence of a particular rule in the target language, and thus cannot explain why a particular form is chosen. In contrast, systematic errors occur the learner has discovered a rule but it is a wrong one. The learner is unable to correct these errors, but he can explain the mistaken, self-discovered rule. Post-systematic errors occur when the learner knows, and can explain, the correct target language rule but uses it inconsistently.
（4）Explaining errors
Explaining why an error was made is the most important step in trying to understand the process of SLA. Two most likely causes of L2 errors are interlingual factors and intralingual factors. The interlingual errors are the result of negative transfer or interference from L1, while intralingual errors can not be traced to cross-linguistic influence. They are regarded as developmental errors and often represent incomplete learning of L2 rules or overgeneralization of them. Now let’s look at two examples from a Chinese-speaking learner of English:
Sample 1: interlingual error
Soldier should love country.
The speaker did not add determiners before the two nouns, ‘soldier’ and ‘country’. Obviously, his English is influenced by his mother tongue, in which nouns do not usually take articles or other determiners. They say ‘士兵应该热爱祖国’ instead of ‘一个士兵应该热爱他的祖国’. However in English, nouns, especially singular countable noun and plural nouns, often take determiners: A soldier should love his country. Therefore, this is an interlingual error, since the error is the result of L1 interference.
Sample 2: intralingual error
He standed beside me.
The same speaker, using ‘standed’ instead of ‘stood’, has learned a rule that the past tense verb takes an –ed suffix, and forgotten that irregular verbs have special past tense forms. He has obviously over-generalized the rule. This error, which is not influenced by the L1, shows that the learner is constructing his own rule for an L2.
（5）Evaluating errors
This step of error analysis is about what effect the error has on the listener or the reader. That is, how serious is the error? Or to what extent does it affect intelligibility or social acceptability? A sentence may have one or two grammatical errors, but it may not affect communication, such as the above two examples produced by Chinese-speaking learners of English. This is called a local error. A local error only affects a single constituent in a sentence, such as a verb, and does not create any processing problem. Hence, no misunderstanding results. However, a global error is different because it violates the overall structure of a sentence. For this reason, it may bring more problems in understanding the meaning of the sentence produced by a speaker. For example:
That computer you want to use now is impossible.
he error in this sentence is global, since there is something wrong in the structure. It is confusing in meaning and difficult to understand. It may be that The computer is impossible, which does not really make any sense; it may also be that It’s impossible for you to use the computer, but the speaker does not seem to mean that. A close study of the erroneous sentence suggests the effect of mother tongue interference: it is a sheer transfer from the L1 “那台计算机你想现在用是不可能的”.
3. Shortcomings of EA
EA is a useful procedure for the study of SLA. However, it has encountered some criticisms.
One problem of EA is that it totally focuses on errors to the exclusion of correct forms. It is argued that one shoud also consider non-errors as well as errors to get the entire picture of learners’ linguistic behaviour. A study (Schachter, 1974) of errors in English restrictive relative clauses (RC) by four groups of ESL learners (native speakers of Persian, Arabic, Chinese, and Japanese) has shown that those who make more errors (Persian and Arabic speakers) also produce far more correct RC sentences than Chinese and Japanese speakers, who produce a comparatively smaller number of RC sentences. The fact that is otherwise ignored by EA is that the percentage of correct use of RC sentences by Persian and Arabic speakers is much higher than by Chinese and Japanese speakers. The results are illustrated in Table 3.1.
Table 3.1 Relative clause production
NL speakers Correct Error Total errors%
Persian 131 43 174 25
Arabic 123 31 154 20
Chinese 67 9 76 12
Japanese 58 5 63 8
American 173 0 173 ---
(Gass & Selinker, 2008:105)
It must be pointed out that absence of errors may result from learners’ avoidance of difficult structures. This is not revealed by EA. The above examples show that Chinese and Japanese L1 speakers make fewer errors in English relative clauses simple because they avoid using them.
A second problem with EA is that it is difficult to determine the type of error. Let’s look at the following examples from Chinese learners of English.
There are so many Taiwan people live around the lake.
There were lots of events happen in my country.
There is a mountain separate two lakes.
There are so many tourist visit there.
( Gass & Selinker, 2008:106)
Superficially, these sentences look like relative clauses without relative pronouns (that, who, which). Another explanation is that they are constructions that are similar to topic-comment constructions in Chinese language. That is, these learners are following a Chinese pattern of establishing a topic ad then making a comment about it, as shown in Table 3.2.
Table 3.2 Topic and comment
Topic Comment
Taiwan people they live around the lake
lots of events they happen in my country
mountain it separates two lakes
tourist(s) they live there
Such sentences produced by Chinese learners of English are very commonplace in both speaking and writing, and is like the following construction in English You see that boy? He just broke the window but not Did you see the boy who just broke the window? Therefore, there can be a difference between what a researcher determines to be the TL structure and what the learner is attempting to produce.
Another problem of EA is that we tend to believe that a learner has formed the correct rule because he has temporarily produced correct forms. But actually, the absence of errors does not mean correct rule formation. For example, when a learner produces correct sentences like I asked him to come, and I enjoyed reading the book, we might say that the learner knows which verb takes which kind of complements. But the error in I saw him to cross the street clearly suggests that the learner has not mastered the correct rule.
3.2.3 Monitor Model
Another early approach to SLA is the Monitor Model, which was proposed by Stephen Krashen (1978). The Monitor Model adopts the notion of language acquisition device (LAD), a metaphor used by N. Chomsky for children’s innate knowledge of language. This model has five hypotheses which present claims and assumptions about how an L2 is acquired. The five hypotheses are:
1. Acquisition-learning hypothesis
According to this hypothesis, acquisition is different from learning. Acquisition is subconscious while learning is conscious, and usually takes place in classroom contexts. As for L2 acquisition, Krashen has the following definition:
Acquisition [is] a process similar, if not identical to the way children develop ability in their first language… Language acquirers are not usually aware of the fact that they are acquiring language, but are only aware of the fact that they are using the language for communication. The result of language acquisition, acquired competence, is also subconscious. We are generally not consciously aware of the rules of language we have acquired. Instead, we have a ‘feel’ for correctness. Grammatical sentences ‘sound’ right, or ‘feel’ right, and errors feel wrong, even if we do not know what rule was violated.
(Krashen,1982:10)
The second way to develop competence in an L2 is by learning, as is described below:
… refer to conscious knowledge of a second language, knowing the rules, being aware of them, and being able to talk about them. In nontechnical terms, learning is ‘knowing about’ a language, known to most people as ‘grammar,’ or ‘rules.’ Some synonyms include formal knowledge of a language or explicit learning.
(Krashen, 1982:10)
As can be seen, L2 development takes place in different ways. In Krashen’s view, learners also use the language developed through these two means for different purposes. What’s more, learned knowledge cannot be converted into acquired knowledge. This position has become known as non-interface position. These two separate systems of knowledge are like the two pockets in a person’s jacket. You can put things into one, but you can never take them out from the other. The acquired system is used to produce language. When learners generate utterances they focus on meaning, not the form. The learned system serves as an ‘inspector’ of the acquired system. It checks to make sure that the utterance is correct against the knowledge in the learned system.
2. Natural order hypothesis
This hypothesis suggests that learners acquire the rules of a language in a predictable order. The order is the same regardless of whether instruction is involved or not. This order is like the order of acquisition we discussed in 2.3.2.
3. Monitor hypothesis
According to Krashen, the learned rules are only used as a monitor, for the purpose of editing or making changes in what is produced through the acquired system. However, the monitor can not be used at all times. There are three conditions that must be met to activate the monitor. The first condition is time. That is, a learner needs time to think about and use the rules in their learned system. The second condition is focus on form. A learner must also pay attention to how we are saying something in addition to what we are saying. The third is know the rule. That is, a learner has to know a rule in order to use it. In other words, a learner must have a learned system in order to apply it. Therefore, the monitor is intended to connect the acquired system and the learned system in language use, as can be shown in Fig. 3.1.
Learned system (monitor)
Acquired Output
system
Figure 3.1 The monitor function of learned system
4. Input hypothesis
According to Krashen (1985:2), language acquisition takes place because there is comprehensible input. If input is understood, and if there is enough of it, grammar can be picked up automatically, as Krashen says, second languages are acquired “by understanding messages, or by receiving comprehensible input”.
Defined by Krashen, comprehensible input is the language that is heard or read and that is slightly ahead of a learner’s current state of grammatical knowledge. If the input contains structures a learner already knows, it serves no purpose in acquisition. Similarly, input containing structures far beyond a learner’s current knowledge is not useful either. Krashen described a learner’s current state of knowledge as i and the next stage as i+1. Thus, the input a learner receives must be at the i+1 level so that it can be acquired. “We move from i, our current level to i+1, the next level along the natural order, by understanding input containing i+1” said Krashen (1985:2). According to Krashen, a teacher’s main role in a classroom setting is to provide students with comprehensible input.
5. Affective filter hypothesis
One explanation of why some are less successful in learning an L2 is, in Krashen’s view, “inappropriate affect”. Affect include factors such as motivation, attitude, self-confidence, and anxiety. Krashen proposed the existence of an affective filter. When the filter is up, input is prevented from passing through; if input is stopped, acquisition does not occur. In contrast, if the filter is down or low, and the input is comprehensible, it will reach the acquisition device and acquisition will take place. One reason why children can successfully acquire their mother tongue may be that affective filter does not exist or is very low.
According to Krashen, acquisition can be explained by two of the above five hypotheses, the input hypothesis and the affective filter hypothesis, besides the LAD.
Although the Monitor Model was severely criticized by researchers, it has had a significant influence on language teaching in the USA in 1980s and 1990s, which was conducted by avoiding explicit teaching of grammar in the classroom settings. Since then, the pendulum began to swing back in the opposite direction: formal grammar teaching was introduced for adult L2 learners, who benefit from an explicit explanation of grammatical structure.
Above we have discussed three early linguistic approaches to the study of SLA, CAH, EA, and the Monitor Model. Although there exists a heated debate among proponents of different approaches, there was widespread agreement on some important points. First, L2 learner is acquiring a “rule-governed” language system. Second, L2 acquisition involves creative mental processes. Third, age is a primary factor in explaining why some more (or less) successful in L2 acquisition(Saville-Troike, 2008: 45-46). During 1980s, new ideas in Chomskyan theoretical linguistics had a major influence on SLA. Universal grammar became the dominant approach focusing on internal factors.
