Working Memory and SLA

Definition:

Working memory — the cognitive system responsible for temporarily holding and mentally manipulating information during active cognitive tasks — is a central limiting factor in second language acquisition and use, determining how much linguistic information can be processed simultaneously during reading, listening, speaking, and writing in an L2. Working memory capacity predicts L2 learning rate, particularly in the early stages of vocabulary and grammar acquisition; phonological working memory (the ability to hold novel sound sequences in short-term store) is among the strongest predictors of vocabulary acquisition speed; and the resource demands of L2 processing (which is inherently more effortful than L1 processing for non-native speakers) mean that working memory load shapes which features of language learners notice, process, and acquire from input.

Baddeley’s Multi-Component Working Memory Model

The dominant working memory model (Baddeley and Hitch, 1974; expanded Baddeley 2000) has four components:

Phonological loop: Holds verbal/phonological information in a short-term phonological store; a sub-vocal rehearsal process prevents decay. Most relevant to language: this is the system that holds a new word’s sound-form while you look up its meaning, and that maintains partial sentence material while the rest of the sentence is being processed.

Visuospatial sketchpad: Holds and manipulates visual and spatial information. Less directly relevant to oral language but important for reading (word form recognition, text layout processing).

Central executive: Attentional control system; coordinates the subordinate systems, allocates attentional resources, controls inhibition of distraction, and manages task-switching. Critical for fluent L2 use: the central executive manages competing demands of meaning-construction and form-monitoring.

Episodic buffer: Temporary storage integrating information from multiple sources (phonological, visuospatial, long-term memory) into coherent episodes. Important for discourse comprehension.

Phonological Working Memory and Vocabulary Acquisition

The most consistent working memory finding in SLA: phonological short-term memory (PSTM) predicts new vocabulary acquisition rate.

Evidence:

Children with better nonword repetition (NWR) scores — a PSTM measure — learn vocabulary faster
Adults with higher PSTM scores learn new vocabulary-form associations faster, especially in paired-associate tasks
PSTM is among the components of language learning aptitude (Carroll’s MLAT includes phonemic coding ability)

The mechanism: acquiring a new word requires holding the novel sound-form in phonological store long enough for it to be encoded into long-term memory. Better PSTM = longer stable hold = better encoding.

Working Memory and Grammar Processing

Working memory is critical for sentence processing in L2:

Complex sentence structures (relative clauses, embeddings, long-distance dependencies) require holding partial sentence material in working memory while processing continues
L2 processing is more effortful than L1, placing greater working memory demands for the same sentence
When working memory is taxed (complex sentence + unknown vocabulary + novel phonology simultaneously), processing fails earlier and comprehension suffers
This is why graded input (comprehensible input) works: it reduces working memory load so acquisition processing can occur

Cognitive load theory application: instructional materials that minimize extraneous cognitive load (overly complex grammar explanations alongside new vocabulary in unfamiliar scripts) produce better learning outcomes than materials that overwhelm working memory capacity.

Individual Differences

Working memory capacity differs across individuals and is:

Substantially heritable (genetic)
Correlated with general intelligence (particularly fluid intelligence, g)
Age-sensitive: declines in processing speed and working memory from middle adulthood
Trainable to a limited extent (working memory training research shows effects but limited transfer)
Partially offset by automatization: high-proficiency speakers process L2 more automatically, reducing working memory load per unit of input

History

1974 — Baddeley and Hitch model. Working memory reconceptualized from simple short-term store to active processing system with multiple components.

1990 — Just and Carpenter. Working memory capacity linked to reading comprehension in L1; foundation for SLA applications.

1998 — Miyake and Friedman. Working memory models applied to foreign language aptitude; phonological working memory identified as key predictor.

2000s–present — SLA working memory research. Studies linking PSTM to vocabulary learning rate, working memory to grammar uptake, and cognitive load to input design; major researchers include Doughty, Williams, Juffs, Harrington.

Practical Application

Reduce cognitive load in study materials. Learning new grammar AND new vocabulary AND new script simultaneously exhausts working memory. Separate new learning streams: build vocabulary before tackling complex grammar; master script before adding grammatical complexity.

Use working memory-friendly input. Comprehensible input at i+1 level optimizes working memory loading: known vocabulary and familiar grammar structures free working memory for processing new elements. Content above i+1 overloads working memory and impairs acquisition.

Common Misconceptions

“Working memory capacity is fixed and cannot be improved.”

While the fundamental capacity of working memory is relatively stable, the efficiency with which it is used can be improved through automatization and strategy development. As L2 processing becomes more automatic, it requires less working memory, freeing capacity for higher-level processing like comprehension and planning.

“Working memory is the same as short-term memory.”

Working memory includes not just storage (short-term memory) but also processing — the ability to manipulate, organize, and integrate information in real time. Language processing engages working memory’s processing component as much as its storage component.

Criticisms

Working memory research in SLA has been critiqued for the difficulty of isolating working memory from general intelligence and language aptitude, for the inconsistent measures used across studies (digit span, reading span, operation span measure somewhat different constructs), and for the unclear direction of causation — does greater working memory capacity facilitate L2 learning, or does L2 learning develop working memory capacity? The ecological validity of laboratory working memory tasks for predicting real-world L2 outcomes has also been questioned.

Social Media Sentiment

Working memory is rarely discussed by name in mainstream language learning communities, but its effects are widely felt — learners describe “running out of mental space” when trying to process complex L2 sentences, forgetting the beginning of a sentence by the time they reach the end, and the exhaustion of extended L2 conversation. The concept resonates with comprehensible input advocates who argue that overwhelming working memory (with incomprehensible input) prevents acquisition.

Last updated: 2026-04

Related Terms

Research

1. Wen, Z.E. (2016). Working Memory and Second Language Learning: Towards an Integrated Approach. Multilingual Matters.

The comprehensive treatment of working memory in SLA — presents an integrated framework for understanding how working memory components contribute to different aspects of L2 learning and processing.

2. Juffs, A., & Harrington, M. (2011). Aspects of working memory in L2 learning. Language Teaching, 44(2), 137–166.

Critical review of working memory research in SLA — evaluates the evidence for working memory as a factor in L2 development, discusses measurement issues, and identifies key areas for future research.

Mikey Does