External and Internal Attention: A Single Cognitive Move in Two Directions

External and internal attention: one selective move in two directions

Pavel Ishchin argues that external and internal attention are the same operation, applied to the world or to one’s thoughts. A 2011 taxonomy by Chun, Golomb, and Turk-Browne formalizes this split, defining internal attention as selection among memories, rules, and responses.

Why This Matters

Conventional models treat visual attention and mental focus as separate faculties, leading to design assumptions (e.g., separate capacity budgets for perception vs. thought) that fail under real-world dual-task loads. If both draw from one limited pool, systems built on separate budgets will underestimate interference—costing up to proportional throughput loss when a single act of attention must divide between the world and the mind.

Key Insights

• Posner’s spotlight model of attention (1980) describes attentional selection as orienting to a location, speeding processing there; Desimone and Duncan’s biased competition theory (1995) explains the neural mechanism as competition among stimuli for neural representation, biased by attention.
• Internal attention selects among internally generated content like memories and rules, not just sensory inputs; Chun, Golomb, and Turk-Browne (2011) organized the field around this external/internal split.
• The strongest test for shared attention is a single bottleneck: loading external attention should interfere with internal attention in proportion, and cross-domain switching should cost as much as within-domain switching.
• A shared frontoparietal resource pool could explain overlap between external and internal attention, but this measurement alone does not confirm identity—a single bottleneck signature is required.

Practical Applications

• Designing user interfaces: external attention selects among on-screen elements (e.g., icons, alerts) via biased competition; pitfall: cluttered interfaces cause competition among multiple features, reducing processing of critical alerts.
• Multitasking software: internal attention selects among task rules or held memories (e.g., working memory for a code snippet); pitfall: switching between a visual task (external) and a mental calculation (internal) incurs a cross-domain switch cost, reducing throughput.
• Cognitive load modeling (e.g., in developer tools): treat attention as a single resource shared across sensory input and thought; pitfall: assuming separate budgets leads to oversubscribing one domain, causing system slowdown under load.

References:

• https://dev.to/poushwell/external-and-internal-attention-one-selective-move-in-two-directions-2o9c