Cognitive and Social Models of Language Acquisition
Interactionism and Social Constructivism
From 1990 onwards, Interactionism (Social Constructivism) posits that social interaction leads to both language acquisition and knowledge development without fixed stages, as thought is mainly internalized speech. Social interaction pushes us into our Zone of Proximal Development (ZPD) as the basis for learning. Conversation serves as the scaffolding for acquiring new language and knowledge. There is a strong emphasis on the pragmatic and communicative aspects of language. Group work and interaction are considered essential for Second Language Acquisition (SLA).
Information Processing (IP)
The Information Processing (IP) model, prominent from the 1970s onwards, suggests the brain acts like a computer that represents images. It receives outside input and manipulates that information. This involves Representation (knowledge) and Access (processing).
- Representation: Stored in Long-Term Memory (LTM).
- Access: Managed by Working Memory (WM), which acts as the reception desk.
Long-Term Memory (LTM)
Long-Term Memory has unlimited capacity regarding how long you keep information; theoretically, you never forget things. It is divided into:
- Declarative / Explicit: Knowledge that we can verbalize.
- Semantic memory: Concepts, facts, and words.
- Episodic memory: Personally experienced events.
- Procedural / Implicit: Knowledge we are not “aware” of (inferable from behavior) or can execute automatically without conscious thought.
- Skills and habits: Walking, speaking, swimming, playing a guitar, or feeling sad when someone cries.
Working Memory (WM)
Working Memory has a limited capacity of approximately 2 seconds or 7 items at a time. It is truly “working” as it encodes information, controls attention, and hosts reasoning (induction, analogy), decision-making, and prioritization. It handles two complementary processes:
- Controlled processing: Conscious, voluntary mental work by the “central executive.” It takes up most cognitive resources and is serial in nature.
- Automatic processing: Unconscious and involuntary. It is fluent and parallel. This process makes a difference among students and is a factor that correlates with success.
Working Memory capacity is smaller in an L2 than in an L1, but it becomes increasingly similar as L2 proficiency develops.
Skill Acquisition Theory
According to Skill Acquisition Theory (McLaughlin, 1986), SLA is the building up of a complex cognitive skill (or knowledge) that gradually moves from controlled to automatic performance for both reception and production. Complex skills are composed of lower-order and higher-order skills. Learning progresses from easier to more complex tasks.
Perception and Controlled Processing
McLaughlin (1986) discusses the distinction between focal attention (controlled, conscious) and peripheral attention (automatic, unconscious). He criticizes this distinction, suggesting you only control knowledge and questioning if peripheral attention is necessarily unconscious.
In controlled processing, lower-order skills demand significant mental space and effort (focal attention, and sometimes peripheral in First Language Acquisition). Because human attention is limited, we can only process small amounts of information at one time.
Practice and Automatic Processing
Through practice, we achieve automatic processing, which requires little mental space (lower neural activation patterns) and relies mainly on peripheral attention. This frees up processing capacity for new information and higher-order skills, such as general ideas and pragmatics. When tasks or information are automatized, they remain out of focal attention unless controlled processing is called upon, which often disrupts performance. Automatized skills are more difficult to change.
This explains fossilization, which is viewed as premature automatization (automatizing incorrect forms fluently).
Mechanisms of Automaticity
How is the automaticity of so much data accomplished?
- Chunking: Language is not processed word-by-word but in strings of prefabricated patterns or lexical units (Sinclair, 1991).
- Transfer-appropriate processing: Information is best retrieved in situations similar to those where it was acquired.
Restructuring
Restructuring is the reorganization of mental representations to make them more coordinated, integrated, and efficient (i.e., fluent). This explains:
- Sudden steps forward when no relevant change has occurred in instruction or exposure.
- Evidence of restructuring (Stauble, 1978): Seen in Jorge’s development of English negations.
- Backsliding / U-shaped development: Where two correct utterances are of a different nature, moving from exemplar-based to rule-based representations.
Production Rules
Production rules follow an “if x, then y” logic:
- Generalization/reduction: Productions become broader or narrower in their application.
- Strengthening: Certain rules (mappings) are applied more often as shortcuts, leading to automaticity.
Declarative knowledge is generalizable and can be transferred to different tasks. In contrast, procedural knowledge is difficult to apply to different tasks (DeKeyser, 2006). When rules are proceduralized and applied automatically, they become inflexible, which can lead to fossilization.
The Information Processing Model Summary
- Learning moves from easier to more difficult, and from controlled (slow) to automatic (fluent) use.
- Learners have control over their own acquisition rate through focal attention and practice.
- Prefabricated units should be made part of declarative knowledge to ease cognitive effort and favor automaticity.
- For automaticity and restructuring, the model suggests meaningful encodings/decodings of declarative knowledge plus associations with production rules to favor generalization before automatization.
Multidimensional Model and Processability Theory
Processing Strategies (Clahsen, 1984)
- Canonical order strategy: Relies on basic word order (SVO: actor-action-acted upon), focusing on meaning without requiring grammar knowledge. (This has been criticized as it is not true for all languages).
- Initialization/Finalization strategy: The beginnings and ends of sentences are more salient, so discrepancies between input and the learner’s own rules are noticed, and initial/final elements are moved around.
- Subordinate clause strategy: Reordering in subordinate clauses is avoided until later, as it requires identifying sub-strings and holding hierarchical material in memory. Moving elements out of sub-strings into other positions is considered the most advanced processing.