Reading focus
Syllabus: How Do We Direct Attention and Control Thought?
Focus on what the instance account says about the source of a response and its dependence on attention and experience. Connect selection history to memory-based performance without treating the two theories as identical.
A question to consider. A response becomes faster for practiced items. What result involving new items or response-time variability would help distinguish retrieval of instances from a generally faster procedure?
Why this reading
Practice may change how a task is accomplished. Logan proposes that retrieving prior solutions can come to replace calculating them afresh, connecting a theory of representation with predictions about speed, variability, and transfer.
Before you read
An instance is a stored trace of an encounter with a task. An algorithm is a procedure for obtaining an answer. The model treats retrieval and the algorithm as competitors in a race. A power function describes diminishing gains with practice; the explanatory question is what mechanism produces that pattern. This substantial paper can take several sittings; understanding its predictions does not require deriving every equation.
Read for
- Identify the assumptions about encoding, retrieval, and separate traces. Ask what each contributes beyond the observation that repeated tasks become faster.
- Distinguish model derivations and simulations from human evidence. Follow why predicting both average response time and its variability constrains the account more than fitting average speed alone.
- Compare repeated and novel items, and consistent and changed interpretations of familiar items. Examine what the transfer comparisons imply about what has been learned.
Consider afterwards
- A person rapidly answers familiar arithmetic problems but gains little on new ones. What evidence would distinguish retrieval of learned answers from faster use of a general calculation procedure?
Editorial context
This is a foundational theoretical and experimental contribution, not an account of every skilled action. Its task-specific retrieval proposal must be distinguished from a general claim that automatic behavior is unconscious or uncontrollable. Braun and colleagues (2009) examine how experience can support new motor adaptations. The two papers address different tasks and representational questions; placing them together does not make one a direct experimental refutation of the other.