Cognition (Ch. 5) - Taylor & Francis

Download Report

Transcript Cognition (Ch. 5) - Taylor & Francis

Cognition (Ch. 5)
Understanding SLA
Lourdes Ortega (2009)
www.routledge.com/cw/ortega
Published by Routledge © 2009 Mark Sawyer
Some contextualization




Positive evidence vs. negative feedback
Why not negative evidence?
A simple example: Put the adverb often into
the sentence I drink coffee
The logical problem of language acquisition:



Language seems logically unlearnable
How can we explain the fact that all children learn it?
UG, (skill learning), emergentist answers
5.1 Information processing in
psychology & SLA
Representation/access
1.


Controlled vs automatic processing
2.


3.
Knowledge/processing
Symbols/computation
Effortful vs. effortless
Serial vs parallel
Attention & memory are limited
5.2 The power of practice:
Proceduralization & automaticity


Proceduralization / Automatization:
Knowledge “that” (declarative/explicit)

Knowledge “how” (procedural/implicit)
Involves speed-up + restructuring
5.3 An exemplary study of skill
acquisition theory: DeKeyser (1997)




6 sessions (3 weeks) to learn
Autopractan grammar, vocab
15 sessions to practice
Conditions: (a) single or dual task;
(b) comprehension, production, mixed
Results: (1)reduction of practice effect
(2) mode-specific automatization
5.4 Long-term memory




Declarative/propositional (explicit)
Procedural (implicit)
Semantic: decontextualized
Episodic: based on experienced events
5.5 Long-term memory &
L2 vocabulary knowledge



Strength (procedural)
Size (declarative)
Depth (both)





multiple meanings (polysemy)
morpho-syntactic behavior
relations to other words
contextual/usage constraints
Nonselectivity: Simultaneous activation of
both L1 & L2 information in language use
5.6 Working memory




Manages 同時 storage & processing
 capacity limitations can be severe
 activation is temporary
Site for executive control
Site of consciousness
Also induction, hypothesizing,
analogizing, prioritizing, deciding
Working Memory: Components



Central executive
Phonological loop
Visuo-spatial sketchpad
5.7 Memory as storage:
Passive working memory tasks


Passive WTM = STM (depending…)
Limitations measured by span of recall of…
 Digits
 Words
 Non-words
 Sentences
STM Limitations: Explanations




Capacity (size)
Time passage
Increasing interference
LTM shortcomings
5.8 Memory as dynamic processing:
Active working memory tasks


Simultaneous processing/storage
Limitations in span measured by…


Reading span task
Listening span task

L2 WM capacity predicts L2 proficiency (

Capacity reflects control of processing
rather than size of storage (Engle, 2002)
5.9 Attention & L2 learning

Characteristics of attention




Limited
Selective
Voluntary
Controls access to consciousness
Attentional conditions used in research
1.
2.
3.

Incidental
Implicit
Explicit
How do these conditions influence
learning?
5.10
Learning without intention
= Incidental learning, IS possible
 e.g. vocabulary thru extensive reading
 But…intention seems to work better
(for specific purposes)
5.11 Learning without attention
Noticing vs. detection only, or…
Inside or outside focal/selective attention
 Focal attention entails consciousness
(subjective experience)
 Fleeting recognition of tree while doing
other things can be detection-only,
but evoked intuition, inference, feeling
entails noticing

5.12 Learning without awareness


Immediate think-aloud research has
shown strong effect of awareness on
learning (more so for understanding)
Indirect measures--uptake of recasts,
note-taking--have shown no effect.
5.13 Disentangling Attention
from Awareness?
Different Measurements
 Awareness


Self-reports (of subjective experience)
Attention


Dual task (learning from unattended task)
Memory tests


Direct: recognize item later
Indirect: show bias to prefer old items
5.14 Learning without rules
= learning rules without…



looking for them (process)
awareness of learning (product)
Such implicit learning can be shown by


Memorizing artificial language strings:
aabcd abbce abeec…
(*acbbe)
Later distinguishing “(un)grammatical”
ones (to some extent) w/o awareness
An exemplary study of symbolic vs.
associative learning: Robinson (1997)

Dative alternation rule: mono-syllabicity
Mitt gave his delegates to John
 Mitt gave John his delegates
 Mitt donated his delegates to John
 *Mitt donated John his delegates
What can be learned from brief training?
 John minided some hot coffee to Sue
 *Sandy bivarded Patrick some Swiss cake


Robinson (1997): Learning conditions
Implicit: Memorize word positions
 Incidental: Read for meaning
 Explicit 1: Look for rules (with help)
 Explicit 2: Receive rule, practice
Post-test: Speed, accuracy of judgments
Results: Explicit 2 fastest, most accurate

All groups better on old than new instances
Robinson (1997): Conclusion


Low-level implicit learning is possible,
allowing fast access of learned instances
Explicit learning leads to generalization with
awareness.
An emergentist turn in SLA?
3 important tenets of learning
associative: based on co-occurrences
probabilistic: not categorical, deterministic
rationally contingent: guesses based on…
1.
2.
3.




accumulated statistical (frequency) information
most relevant recent evidence
attention to cues
contextual clues
Other emergentist tenets

Usage-based: use & knowledge are
inseparable; no competence/performance,
representation/access distinctions


Grounded: language & thought are
structured by human experience
Dynamic system: identifiable patterns
emerge from a multiplicity of factors