Theories of Language Learning

The behaviourist perspective

Behaviourism sees learning in terms of habits formation, the creation of systematic association between a stimulus and a response (without reference to mental processes or thinking). Learning happens through experience with environment, in a passive way.

Classical and operant conditioning

- Classical conditioning (Pavlov): automatic response.

- Operant conditioning (Skinner): voluntary response followed by a reinforcing stimulus with positive vs negative reinforcement.

Learning is defined as a change in behaviour in the learner: positive reinforcement becomes habits, negative responses produce the abandonment of the "habit".

L1 and L2 learning

Language is a set of patterns or habits learned through experience with the environment. The child learns by imitation, and then makes inferences on simple grammar rules, testing and verifying these hypotheses in speech production (trial-and-error system).

Learning in L2 consists in replacing the responses consolidated in the L1 with those of L2.

Contrastive analyses hypothesis (CAH)

Errors were assumed to be the result of transfer from learners' first language. Mistakes made by learners are caused by the differences between L1 and L2, and by the transfer of L1 habits to L2. Languages can be analyzed in contrastive ways:

• Symmetries that cause positive transfer
• Asymmetries that are a source of negative interference

Key figures: Robert Lado, Charles Fries, Robert Stockwell, Donald Bowen.

However, many L2 errors were not the result of transfer.

Audio lingual method (ALM)

Language learning could be trained through a system of reinforcement. Based on drills, mechanical and pattern language practice that required repetition, manipulation, transformation of a particular form or structure. Classroom activities emphasized mimicry and memorization. It is the product of three historical circumstances:

• Behaviourism
• American Structuralism
• World War 2

The finding that many aspects of learners’ language could not be explained by the CAH led researchers to a different approach for analyzing errors.

Theories relying on behaviourism

1. Error Analysis (Stephen Pit Corder, 1967)

The goal is to discover what learners really know about the language. Emerges as a reaction to the CAH. When learners produce correct sentences, they may simply be repeating something they have already heard. But, when they produce sentences that differ from the target language, we may assume that these sentences reflect the learners’ current competence.

In this view, errors are a necessary feature of learner language. Not all errors could be attributed to L1 influence, and L2 learners were active creators of linguistic systems. Key figures: Martina Burt, Heidi Dulay, Elaine Tarone, Jack Richards, Jackelyn Schachter.

Like children, perhaps L2 learners came equipped with something internal (the internal syllabus), and like child language, learner language is a system in its own right. Errors are the result of a discovering and testing mechanism.

Distinction between errors and mistakes

- Errors (true errors of competence): errors of performance say nothing about the underlying speaker's competence.

- Mistakes should be excluded from analysis.

Types of errors:

• Omission (the absence of an item that must appear in a well-formed utterance)
• Addition (the presence of an item that must not appear in a well-formed utterance)
• Misinformation (the use of a wrong form of the morpheme or structure)
• Misordering (the incorrect placement of a morpheme or group of morphemes in an utterance)

Corder's types of errors

According to Corder, there are 3 types of errors:

• Presystematic errors: the learner is unaware of a particular rule in the target language and can’t explain why he has chosen a particular form.
• Systematic errors: the learner has discovered a rule, but the wrong one; he can explain the mistaken rule used.
• Postsystematic errors: the learner knows the rule but uses it inconsistently; he can explain the correct rule.

Further distinction:

• Overt errors (deviation in form)
• Covert errors (correct form, but pragmatically odd)

Procedure for error analysis

• Selection of a corpus of language
• Identification of errors in the corpus
• Classification of the errors identified
• Explanation of the psycholinguistic cause of the errors
• Evaluation (error gravity ranking) of the errors

Problems of error analysis

1. Identification: existence of a “middle ground” between completely acceptable language and completely erroneous language.

2. Classification: existence of “global errors” that seem to extend over the whole sentence.

3. Assigning psycholinguistic causes: difficulty in deciding whether an error is caused by language transfer or not.

4. Error evaluation: differences between communicative and formal criteria, native speakers, and language teachers.

5. It can only provide negative evidence.

6. Low-risk takers tend to avoid structures they are unsure of, therefore they produce fewer errors.

7. If learners do not produce a certain structure, EA cannot say whether they have mastered that structure or not.

Ultimate distinction in input vs intake

- Input: the language available from the environment.

- Intake: the language that makes its way into the learner’s developing competence.

Highly influential for EA:

• Chomsky’s view of FLA: there is an inbuilt syllabus that determines the order in which the language system is acquired.
• Weinreich’s position of a psychological or psycholinguistic explanation for language interference.

2. Interlanguage (Larry Selinker, 1972)

The learners’ developing second language knowledge. L2 learners possess an internal linguistic system worthy of study in its own right. It is called Interlanguage because the system is neither the L1 nor the L2, but something that the learner is building from environmental data.

It has characteristics influenced by:

• Previous learned languages
• The L2
• General characteristics that occur in all interlanguage systems, independently from L1 or L2.

Interlanguage is:

• Systematic (has its own rules)
• Dynamic (continually evolving)
• Variable (may show forms from an earlier stage of the IL, and even backslide into earlier stages)
• Simplified (both formally and functionally)

Five central processes underlying SLA

• Language transfer (another name for “interference”, to stress the active role of the learner)
• Transfer of training (effects of bad teaching)
• Strategies of L2 learning (attempts to simplify the system which the learner is struggling to master)
• Strategies of L2 communication
• Overgeneralisation of L2 rules

Fossilization: some features in a learner’s language seem to stop changing. Today, it refers to the end-state of SLA, specifically to an end-state that is not native-like. Some researchers prefer the term stabilization.

Questions on fossilization and motivation

What is the source of fossilization? Is motivation an important part of SLA? Motivation is one of a number of what researchers call individual differences:

• Integrative motivation: refers to the internal impetus that learners have to relate to, identify with, or integrate into another L2 culture.
• Instrumental motivation: related to purposeful use of language, such as educational or economic purposes. Language is a tool.

The innatist position

Chomsky rejects the behaviourist explanation of L1 acquisition on the basis of the poverty of the stimulus argument. Language use is assumed to be based upon an abstract linguistic system, a mental representation of grammar; much of this unconscious knowledge derives from UG.

Lydia White (1991) and others agree that acquisition of many grammatical features of the new language takes place naturally when learners are engaged in meaningful use of the language; but because the nature of UG is altered by the acquisition of the first language, second language learners may sometimes need explicit information about what is not grammatical in the second language.

Principles constrain the form of grammars. Parameters account for certain circumscribed differences across languages; input data trigger the appropriate parametric choice.

L1 acquisition = process of parameters setting.

L2 acquisition = process of parameters re-setting.

Focus of the innatist position

The focus is on the nature of the learner’s internal mental representation and what constrained it. Researchers are more interested in the language competence.

Usage and knowledge reflection

Usage does not necessarily provide an accurate reflection of knowledge or acquisition. L2 learners may appear to be highly proficient, even native-like, and yet have non-native grammar. So sentences that superficially appear to be identical to those produced by native speakers might have different underlying representations.

Grammaticality judgment task: the most common used task to determine knowledge of L2 (un)grammaticality; it taps linguistic intuition and allows the researcher to investigate whether sentences that are disallowed for native speakers because of principles of UG are also prohibited in the interlanguage grammar.

The innatist position is almost universally accepted by child language acquisitionists, but for SLA there is no agreement or little agreement in the generative perspective. Some scholars rejected this position, including Robert DeKeyser, James Lantolf, associated with cognitive and psychological or sociocultural approaches.

Major questions in SLA

• Exposure to input is necessary for SLA. Universal principles and language-specific parameter settings must be triggered by input from the language being acquired. A central question is that UG can explain why L2 learners eventually know more about the language than they could reasonably have learned.
• Learners come to know more than what they have been exposed to in the input. L2 learners come to know very subtle properties of the L2 which are undetermined by the L2 input, both naturalistic input and classroom instruction, and which cannot be explained in terms of the L1 grammar either.
• There are limits on the effects of frequency on SLA. UG allows learners to acquire properties quite unrelated to frequency; children achieve certain kinds of knowledge on the basis of little or no input.

If Interlanguage competence goes beyond the L2 input and the L1 grammar in particular respects, then UG is implicated in non-native language acquisition.

Problems with UG

• Lack of neurobiological validity (difficulty to identify an area that might constitute UG)
• The innatism issue (“poverty of imagination” fallacy, the tendency to resort to innateness whenever there is no obvious alternative explanation)
• Speed of language acquisition (L1 acquisition isn’t as fast as it is believed, but it is rather slow)
• Grammaticalization and other experience-based processes (certain grammatical rules that have traditionally been thought of as being innate can actually emerge as a result of humans stringing together language components into consolidated and often conventionalized patterns, the so-called "grammaticalization")