Is adaptive control in language production mediated by learning?

Area of Science:

• Cognitive Psychology
• Psycholinguistics
• Neuroscience

Background:

• The Congruency Sequence Effect (CSE) in language production, similar to non-verbal tasks, indicates adaptive control but its underlying mechanism is debated.
• CSE is defined as a reduced congruency effect following incongruent trials, suggesting adjustments in cognitive control processes.

Purpose of the Study:

• To investigate the mechanism of adaptive control in language production by testing predictions of a learning model against activation models.
• To evaluate within-task CSE and cross-task CSE within a task-switching paradigm using the Picture Word Interference (PWI) task.

Main Methods:

• Two experiments employed a task-switching paradigm, assessing CSE on PWI trials based on preceding PWI trials (within-task) and trials from a different task (cross-task).
• Experiment 1 used a visuospatial task, while Experiment 2 used a self-paced reading task with varied inter-trial gap durations.
• Predictions from a learning model (robust within-task CSE, absent/reversed cross-task CSE) were contrasted with activation models.

Main Results:

• A robust within-task CSE was consistently observed in PWI tasks across both experiments, unaffected by intervening trials or gap duration.
• Within-task CSE was also evident in the visuospatial and sentence reading tasks.
• Cross-task CSE, examined between PWI and the other tasks, was either absent or reversed.

Conclusions:

• The findings strongly support a learning model of adaptive control in language production.
• The results suggest that adaptive control mechanisms, as indexed by CSE, operate differently across tasks and are influenced by learning processes.