Attention

Origin domain

Psychology

Also from

Economics & Finance, Computer Science & Software Engineering, Organizational & Management Science, Neuroscience

Aliases

Attention Control, Selective Attention

Core Idea

Attention is the selective allocation of a limited cognitive, organizational, or computational resource to a subset of available information, options, or tasks — what James (1890) classically described as "the taking possession by the mind, in clear and vivid form, of one out of what seem several simultaneously possible objects or trains of thought." ^[1] It is the gating mechanism that permits selected items to be processed deeply, while unselected items are filtered, delayed, or discarded, as Broadbent (1958) formalized in his filter theory of selective attention. ^[2] The scarcity is absolute: no agent (human, organization, algorithm) can process everything simultaneously; attention is how scarcity surfaces upstream of decision-making.

How would you explain it like I'm…

The Mind's Flashlight

Imagine a flashlight in a dark room full of toys. The flashlight can only light up one or two toys at a time — you can't see them all at once. Attention is like that flashlight in your head: it shines on one thing, and the other things stay dim.

Picking What to Think About

Attention is choosing what to think about. Your brain, your computer, even a busy classroom — none of them can process every single thing happening at once. So they pick a few things to handle deeply and let the rest fade into the background. It's like a gate at the front of a busy hallway: only some signals get to come in and be worked on, the rest get held up or dropped. Attention is how a limited thinker copes with an unlimited world.

Selective Focus

Attention is the selective allocation of a limited cognitive, organizational, or computational resource to a subset of available information, options, or tasks. William James called it 'the taking possession by the mind, in clear and vivid form, of one out of what seem several simultaneously possible objects.' Because no agent — human, organization, or algorithm — can process everything at once, attention is the gating mechanism that decides which items get processed deeply and which are filtered, delayed, or dropped. Broadbent (1958) modeled it as a filter applied to incoming signals. Attention is where the bottleneck of bounded processing becomes visible upstream of any decision.

 

Attention is the selective allocation of a limited cognitive, organizational, or computational resource to a subset of the inputs, options, or tasks available at a moment. William James (1890) described it canonically as 'the taking possession by the mind, in clear and vivid form, of one out of what seem several simultaneously possible objects or trains of thought.' Broadbent (1958) gave the first influential mechanistic model: a filter applied to incoming signals, admitting some for deep processing and screening others out. The underlying scarcity is absolute — no agent can process everything in parallel — so attention is the gating layer through which finite-capacity systems cope with overwhelming input. The same structural role appears across substrates: foveal vision, organizational triage, transformer attention layers, OS interrupt handling. Attention is therefore the mechanism by which bounded capacity surfaces upstream of decision-making.

Structural Signature

Attention exhibits consistent structural properties across domains:

Resource bottleneck. Processing capacity is finite and fixed per unit time, a constraint Kahneman (1973) developed into a unified capacity model of attentional effort. ^[3]

Allocation mechanism. Inputs compete for selection via salience, goals, emotion, learned filters, or interrupt protocols, a duality Corbetta and Shulman (2002) mapped onto distinct goal-directed (top-down) and stimulus-driven (bottom-up) brain networks. ^[4]

Consequence asymmetry. Selected items receive deep processing and influence decisions; unselected items have zero causal effect, regardless of their objective value — a consequence Mack and Rock (1998) demonstrated empirically in their studies of inattentional blindness. ^[5]

Cost of filter failure. Misallocation (attending to noise, ignoring signal) degrades outcomes; opportunity cost of wrongly allocated attention compounds, as Wickens (2008) shows in his multiple resource theory of mental workload and dual-task interference. ^[6]

What It Is Not

Attention is distinct from but often conflated with:

— Focus: a narrower outcome referring to concentration quality and depth on a single object. Attention gates what can be focused on; focus measures the quality of that gate's output.

— Prioritization: an ordering or ranking of options by value. Prioritization determines which items should receive attention; attention determines which items actually do. An excellent priority list with poor attention allocation yields poor results.

— Vigilance: sustained readiness to detect a rare signal. Vigilance is a mode of attention deployment; attention is the broader resource-allocation phenomenon.

— Consciousness: the subjective experience of processing. Attention is a functional mechanism; consciousness may or may not accompany it.

Broad Use

^[7] Cognitive psychology (Broadbent 1958, Treisman, Posner & Petersen 1990): selective attention in perception, attention spans, dichotic listening, attentional bottlenecks, top-down (goal-driven) vs bottom-up (stimulus-driven) attention networks, attentional control disorders.

^[8] Economics & finance (Simon 1971, Davenport & Beck 2001): attention economy, "wealth of information creates poverty of attention," bounded rationality, market anomalies driven by retail investor attention, algorithmic attention to market microstructure.

^[9] Machine learning & AI (Vaswani et al. 2017, transformer architectures): attention mechanisms in sequence-to-sequence models, self-attention, multi-head attention, cross-attention, scaled dot-product attention as differentiable allocation.

^[10] Computer science & software engineering (Tanenbaum & Bos 2014): task scheduling, interrupt handling, event-driven systems, cache coherence (memory attention), resource allocation in operating systems.

^[11] Organizational management (Ocasio 1997 attention-based view of the firm): executive attention as the scarcest organizational resource, how leadership focus shapes strategic decisions, the role of information channels and issue interpretation in directing attention.

^[12] Neuroscience (Desimone & Duncan 1995): neural correlates of selective attention (parietal and frontal networks), attention as gain modulation in sensory cortex, cholinergic attention system, salience networks, attentional disorders (ADHD, neglect syndrome).

^[13] Advertising & marketing (Davenport & Beck 2001): eyeballs as a commodity, attention capture as core business function, ad placement optimization, algorithmic feeds designed to maximize engagement (i.e., attention extraction).

Clarity

Attention is the gating mechanism—the structural process that selects which inputs are processed. Distinguishing it from related concepts clarifies where scarcity operates and what drives outcomes. A decision-maker with excellent prioritization but poor attention allocation will implement the wrong strategy. An organization with clear objectives but no attention management will scatter effort across low-value signals. Naming this layer makes the cost of inattention visible and measurable.

Manages Complexity

Attention transforms an overwhelming problem into a tractable framework. Instead of "I can't process everything" (paralyzing), the framework asks: What is the resource limit? What triggers or guides allocation—salience, emotion, learned heuristics, organizational norms, interrupt protocols? What is the filter rule, and what does it exclude? What is the opportunity cost of misallocation? This applies whether managing human workload, network bandwidth, GPU memory, or organizational focus.

The framework also clarifies where intervention can occur: alter the input stream, change salience cues, modify the allocation rule, design better filters, or measure and reward correct attention patterns.

Abstract Reasoning

Attention invites thinking in terms of:

— Signal-to-noise ratios and how salience distorts them.

— Opportunity cost: the value of what is not attended to.

— Attentional capture: involuntary shifts (e.g., sudden loud noise, threat stimulus) that override top-down goals.

— Attention as a competitive landscape: information, stimuli, and claims fight for processing capacity.

— Attentional debt: the accumulation of unprocessed important signals until the system fails (e.g., diagnostic miss, strategic surprise).

— Domain transfer: a mechanism effective in one domain (e.g., interrupt prioritization in operating systems) may transfer to another (e.g., meeting protocols in organizations).

Knowledge Transfer

The structural insight recurs across personal productivity, clinical diagnostics, military command, financial markets, neural networks, and organizational hierarchy. A radiologist allocates attention to anatomical regions; a CEO allocates attention to board reports and market signals; a neural network allocates attention weights to input features; a manufacturing plant allocates supervisor attention to bottleneck lines. The mechanism is identical: bounded capacity, selective allocation, feedback loops, consequences of misallocation.

Mechanisms from one domain transfer across others: salience weighting (how a clinical flag redirects a radiologist's attention), interrupt protocols (how urgent signals preempt routine processing), curated feeds (how information architecture shapes allocation), and measurement of attention metrics (e.g., dwell time, engagement, processing latency).

Examples

Formal/abstract

In formal terms, attention is a weight vector w ∈ [0, 1]^n where Σw_i = 1, applied to an input set x to produce y = w ⊙ x. The weight distribution is determined by a query q and the input features, via a function f: w = softmax(f(q, x)), where f can be a learned neural function or a heuristic rule. The allocation rule f encodes the salience function, goal alignment, and learned filters. In humans, this is approximated by the Posner attention networks; in transformers, by scaled dot-product attention; in organizations, by meeting agendas and reporting hierarchies.

Applied/industry

Clinical radiology: A radiologist reads 40 chest X-rays per hour. Each image contains ~50 anatomically distinct regions and millions of pixels. Attention capacity is the bottleneck. A small nodule might be present but escape notice amid vascular shadows unless a clinical flag (patient smoking history, prior imaging, AI detector highlighting the region) redirects attention. The radiologist's diagnostic ability is sound; what changed was the allocation of attention. False negatives are attention misallocations; the signal was visible but not selected for processing.

Market trading: A hedge fund manager cannot monitor all 5,000 stocks simultaneously. Attention gets allocated to a curated watchlist, sector rotation signals, and macroeconomic indicators. When a major news event (earnings surprise, regulatory change) breaks, attention shifts via interrupt protocol. Profitable trading depends not on perfect knowledge but on attention being allocated to the right signals at the right time. Opportunity cost of attention is quantifiable: every minute spent analyzing Stock A is not spent on Stock B.

Executive strategy: A CEO receives 200+ emails per day and has 8 hours of meeting capacity. Attention allocation is filtered by executive assistants, meeting agendas, and board priorities. A disruptive competitive threat might be missed entirely if no information channel surfaces it (lack of salience) or if attention is directed elsewhere (competing priorities). The CEO's strategic decisions are constrained not by analytical ability but by which signals received attention.

Mapped back: In all three cases, the outcome (diagnosis, trade return, strategic decision) depends on what was selected for processing, not on the average quality of all available information. Attention acts as a gate. Improving outcomes requires either improving the filter rule (what to attend to) or expanding the bottleneck (increasing processing capacity, e.g., via assistants, AI tools, delegation). Ignoring attention as a structural constraint leads to strategies that look rational on paper but fail in practice because key information was never processed.

Structural Tensions

T1: Attention speed vs. attention accuracy. Fast allocation (snap judgments, interrupt-driven) risks attending to noise. Slow, deliberative allocation is more accurate but cannot keep pace with input flow. The tension is resolved by multi-level architectures: rapid filtering of obviously low-value inputs, careful processing of borderline cases. In organizations: quick triage by junior staff, careful review by experts.

T2: Top-down goal alignment vs. bottom-up salience capture. Goal-driven attention selects items that support current objectives; salience-driven attention (loud noise, threat stimulus, emotional content) captures processing regardless of goals. Neither is optimal alone. Evolutionary and organizational design resolves this via weighted combination: goals set baseline allocation, but sufficiently salient inputs override (e.g., a threat stimulus interrupts productive work; new market data interrupts existing strategy). The tension is healthy; resolution requires transparency about the weighting.

T3: Selective attention depth vs. breadth coverage. Deep focus on one item (high processing depth, high output quality) necessarily reduces coverage of other items. Distributed attention across many items (broad monitoring) necessarily reduces depth on each. This is not a trade-off to be "solved" but a fundamental constraint. Different roles require different positions on this spectrum: a surgeon needs extreme focus; a director of strategy needs broad monitoring. Team composition and task design reflect this.

T4: Attention scarcity vs. information abundance. Information production has accelerated (digital systems, social media, real-time data feeds) while human and organizational attention has not. The mismatch is structural. Naive responses (trying to attend to all information) fail; effective responses accept scarcity and design better filters and curation mechanisms. This creates a market for attention direction: news curation, analyst reports, algorithmic recommendations, information design. The tension is unlikely to resolve; managing it is a core organizational competency.

T5: Measurable attention vs. implicit attention. Attention that is conscious and explicit (task focus, listed priorities) can be managed and communicated; implicit attention (heuristic salience filters, learned associations, emotional reactions) is harder to measure and justify but often drives allocation. Organizations often manage only explicit attention while implicit attention dominates outcomes. Resolution requires making implicit allocation transparent: auditing which signals actually received processing, tracing how allocation decisions were made, surfacing hidden salience functions.

T6: Individual attention vs. collective attention. In groups, attention is fractured across members; collective attention is the intersection. A team can attend to more signals in parallel but loses depth and coherence. A single individual can achieve deep focus but has narrow coverage. Organizations oscillate between decentralized attention (individual teams focus deeply) and centralized attention (aligned priorities, loss of specialist focus). The tension is structural to hierarchy and scale.

Structural–Framed Character

Attention is a hybrid on the structural–framed spectrum, and it leans structural with only a light frame. Part of it is a bare pattern that holds in any system: a limited resource selectively allocated to some inputs while others are filtered out. Part of it is a vocabulary inherited from psychology, where it was first studied as a property of the mind.

The structural core — a fixed processing bottleneck, gating, and selective allocation under a capacity limit — applies unchanged to cognition, to organizations triaging tasks, and to computational systems weighting inputs, and stating it requires no appeal to human norms. That pattern is largely descriptive and value-neutral. The residual frame comes from the prime's psychological home: the language of the mind "taking possession" of objects, of vivid awareness and trains of thought, carries assumptions about an experiencing subject that a bare resource-allocation pattern does not need. Because the resource-bottleneck pattern carries most of the weight while the mentalistic vocabulary adds only a thin layer, it sits just on the structural side of the middle.

Substrate Independence

Attention is about as substrate-independent as a prime can be — composite 5 / 5 on the substrate-independence scale. Its signature is fully substrate-agnostic — a limited resource selectively allocated to a subset of available information, with a bottleneck, an allocation mechanism, and an asymmetric consequence. It spans psychology, economics, computer science, organizational management, and neuroscience, and the examples pair clinical radiology with the formal weight vector of an ML attention mechanism. That breadth backed by strong, concrete evidence marks it as a canonical cross-substrate prime.

• Composite substrate independence — 5 / 5
• Domain breadth — 5 / 5
• Structural abstraction — 5 / 5
• Transfer evidence — 5 / 5

Relationships to Other Abstractions

Current abstraction Attention Prime

Foundational — no parent edges in the catalog.

Children (24) — more specific cases that build on this

• Distraction Domain-specific is a kind of Attention

  Distraction is selective Attention specialized to involuntary secondary- stimulus capture against the agent's active task or goal.

• Curiosity Prime is a kind of Attention

  Curiosity is a specific kind of attention, biasing selective allocation toward information that closes a salient knowledge gap.

• Goal Shielding Prime is a kind of, typical Attention

  'attention is the broad allocation of a scarce processing resource; goal shielding is its specifically goal-protective application — the active denial of that resource to competing goals while one is held.' attention is the genus.

▸ Show 21 more

Neighborhood in Abstraction Space

Attention sits among the more crowded primes in the catalog (14^th percentile for distinctiveness): several abstractions describe nearly the same structure, so a description that fits it will tend to fit its neighbors too — transporting it usually means disambiguating within this family rather than landing on it exactly.

Family — Capacity Limits & Attention (19 primes)

Nearest neighbors

• Attentional Capacity — 0.82
• Cognitive Resource Depletion — 0.74
• Alertness — 0.73
• Processing Fluency — 0.73
• Task Interdependence — 0.73

Computed from structural-signature embeddings · 2026-07-26

Not to Be Confused With

Attention must be distinguished from Emphasis (similarity 0.708), its nearest neighbor. Emphasis is the rhetorical, perceptual, or design technique of making a selected item more salient — through formatting (bold text, bright colors, large size), prominence (top placement, repeated mention), or emotional framing (urgency, fear appeal). Emphasis is about how to make something stand out. Attention is about what gets processed in the first place. The difference is crucial: emphasis works on items already in the attended field — it amplifies or highlights them within the processing that has already been allocated. Attention is the gate that determines which items enter the attended field. You can emphasize something that no one attends to (highlighting a faint color among brighter ones, when that region is unattended); the emphasis has no effect because attention was never allocated to that region. Conversely, an item can receive deep attention without any special emphasis (a stock ticker in the corner of a trader's screen receives attention but no visual emphasis). Emphasis is a tool for increasing the salience of already-attended items; attention is the selection mechanism that determines what gets attended. In practice, emphasis tries to capture attention (by making something more salient, one hopes to trigger attention), but the relationship is asymmetrical: emphasis presumes some attention, while attention determines what emphasis can affect. A well-designed information interface combines both: it allocates attention through filtering and curation (attention architecture) and then emphasizes the most critical items within the attended field (emphasis design).

Attention is not Cognitive Load, though the two interact. Cognitive load is the total processing demand placed on a finite working-memory budget — the sum of mental effort required by all items currently being processed. Attention is the allocation mechanism that decides which items from the total available set receive processing capacity. Cognitive load measures the pressure on the system; attention determines how that capacity is distributed. A task can have low cognitive load (simple items, easy processing) but poor attention allocation (focus on irrelevant items, missing important signals). A task can have high cognitive load (complex items, many competing demands) but good attention allocation (focus on critical items despite the pressure). The relationship is sequential: attention determines which items get capacity; cognitive load measures the resulting burden. Too-high cognitive load indicates that either (1) too many items were allocated attention, or (2) the items allocated attention are inherently demanding. In the first case, attention allocation is the problem; in the second, it may be unavoidable (a surgeon in an emergency must attend to many life-critical items simultaneously, creating high cognitive load). The error is treating cognitive load as if it determines attention — in fact, attention choices create the cognitive load. An individual with good attention discipline can manage high-load tasks better than someone with poor attention who attends to low-priority but cognitively-demanding tasks. Thus, attention discipline reduces experienced cognitive load by filtering what receives capacity in the first place.

Attention is distinct from Prioritization, which is the ranking of items by value, urgency, or importance — determining which items should receive resources. Prioritization is a ranking, a list, a plan. Attention is the actual moment-to-moment allocation of processing to items right now. A perfectly-prioritized list (items ranked by importance, constraints, deadlines, strategic value) with poor attention allocation (the person is focused on low-priority items, distracted by interruptions, or attending to items in the wrong order) yields poor outcomes. The priority list is aspirational and static; attention is the dynamic, real-time mechanism. A classic failure pattern is the well-prioritized todo-list holder whose actual attention is driven by email urgency, meeting schedules, or emotional salience rather than the stated priorities. The list says "Item A is highest priority," but attention goes to "Item B is demanding immediate response." Prioritization is should; attention is is. Improving outcomes requires aligning attention with priorities — actually processing the high-priority items, not just ranking them. This distinction clarifies why many productivity systems fail: they produce good prioritization without managing the attention allocation that makes prioritization actionable. The solution is not a better priority list but better attention architecture — eliminating distractions, curating inputs, designing notification systems to surface priorities rather than letting urgency dominate. Thus, prioritization and attention are sequential: one prioritizes items, then must manage attention to ensure the high-priority items actually get processed.

Solution Archetypes

Solution archetypes in the catalog that build on this prime — directly (this prime is a source ingredient) or as a related prime.

Built directly on this prime (16)

• Activation Decay Measurement: Treat priming as a fading state: measure its useful lifetime, set an action or refresh window, and stop relying on it after it expires.
  ▸ Mechanisms (7)
• Active Goal Shielding: Protect the current goal by reducing access to competing goals, preserving only explicit exceptions, and releasing suppression once the goal window ends.
  ▸ Mechanisms (12)
• Alertness-Capacity Maintenance: Maintain the standing ability to notice important change without forcing continuous attention, alarm overload, or permanent hypervigilance.
  ▸ Mechanisms (11)
• Cascaded Hierarchical Recognition: Recognize complex cases by moving attention through a hierarchy of coarse filters and fine discriminators instead of trying to inspect every possible feature at once.
  ▸ Mechanisms (10)
• Decisive-Point Concentration: Create local superiority at the decisive point by massing finite effort there and deliberately accepting bounded weakness elsewhere.
  ▸ Mechanisms (12)
• Figure-Ground Structuring: Structure a field so the intended figure stands forward against a supportive ground that supplies context without competing for the same attention.
  ▸ Mechanisms (6)
• Kairotic Window Alignment: Match an action to the receiving system’s brief receptive state so the same action lands when it can actually take hold.
  ▸ Mechanisms (10)
• Lead-Support Channel Orchestration: Make one channel carry the foreground task while companion channels deliberately support it through calibrated salience, timing, register, redundancy, and interruption rules.
  ▸ Mechanisms (10)
• Negative Priming Avoidance: Do not let the warning, prohibition, or correction make the unwanted idea the easiest thing to think about.
  ▸ Mechanisms (8)
• Novelty-Driven Attention Capture: Make a target noticeable by breaking a stable pattern just enough to trigger attention, then connect that break to the intended meaning or action before the novelty becomes noise.
  ▸ Mechanisms (7)
• Other-Agent State Model Calibration: Model another agent as having its own partial knowledge, goals, attention, constraints, and interpretations, then update that model from evidence before routing action through it.
  ▸ Mechanisms (11)
• Perception-Comprehension-Projection Loop Design: Keep action aligned with a moving situation by continuously refreshing what is seen, what it means, what is likely next, and what decision it now supports.
  ▸ Mechanisms (10)
• Predictive-Cue Wayfinding Design: Make local cues honestly predict what lies down each path so agents can choose, continue, or recover without needing a complete map.
  ▸ Mechanisms (9)
• Salience-Significance Decoupling: Separate what got attention from what deserves weight.
  ▸ Mechanisms (12)
• Shared Attention Anchoring Design: Make the focal target mutually visible, referable, and known-to-be-shared before people coordinate meaning or action around it.
  ▸ Mechanisms (10)
• Supernormal Cue Guardrail Design: Prevent engineered cues from exceeding the range where a responder can regulate proportionate response.
  ▸ Mechanisms (10)

Also a related prime in 45 archetypes

• Accountable Gatekeeping Design: Design choke-point selection so passage decisions use explicit criteria, bounded discretion, traceable reasons, review paths, and distribution audits rather than opaque gatekeeper preference.
• Affective Contagion Modulation: Modulate emotional contagion by making affective spread visible and then adding grounding, buffering, channeling, or cooling structures before group feeling becomes runaway pressure.
• Affordance Shaping: Arrange the fit between an agent and its environment so the right actions are available, noticeable, and easier at the moment they matter.
• Ambiguity-Exploitation in Visual Metaphor: Design a visual metaphor so several plausible readings are invited, constrained, and meaningfully productive rather than accidental or confusing.
• Attenuated Threat Inoculation: Prepare a receiver for a future attack by giving it a safe weak dose of the attack, showing why that dose fails, and rehearsing how to recognize and resist stronger variants later.
• Audience-Boundary Signal Spillover Governance: Before sending a bounded signal, map who else will see it, how they will interpret it, and what response load or legitimacy spillover they may create.
• Backfire-Aware Suppression Design: Handle harmful or unwanted information without making the act of suppression more newsworthy than the information itself.
• Cue-Triggered Intention Execution: Bind an intended future action to a cue so it can sleep in the background and reappear exactly when action becomes possible.
• Encoding–Retrieval Context Alignment: Design encoding, practice, cues, and fallback so the features available at use can recover what was learned.
• Epistemic Boundary Permeability Design: Keep a belief community from mistaking a filtered environment for reality by making the filter visible and routing credible corrective signals through trusted, sustainable cross-boundary channels.

▸ Show 35 more

Notes

Attention is often treated as an individual cognitive trait ("she has poor attention") rather than a structural constraint shared across all agents. This obscures its power as an explanatory lens and mislocates the problem. Poor organizational decisions are often not due to poor intelligence or analysis but to poor attention allocation; the signal was available but not processed.

The rise of digital information abundance has made attention management increasingly central to strategy. Organizations that excel at filtering, curation, and directing attention to high-value signals outcompete those that try to process everything. Similarly, individuals who master attention management (via systems, delegation, prioritization) are more effective than those with higher raw cognitive ability but poor attention discipline.

Attention is also a site of power and asymmetry. Those who control what information is salient (media, algorithms, organizational gatekeepers) effectively control what receives attention and thus what is decided. Attention engineering can be benign (good design) or manipulative (dark patterns, manufactured consent). Transparency about what shapes salience—what algorithms are optimizing for, what narratives are being pushed, what voices are muted—is a prerequisite for defensible attention allocation.

The structural asymmetry between attended and unattended items — that what is not selected has effectively zero causal influence on downstream decisions — is the empirical core of Simons and Chabris's (1999) "gorilla in our midst" demonstrations of sustained inattentional blindness. ^[14]

The transfer mechanism — that salience weighting, interrupt protocols, curated feeds, and attention metrics are mathematically equivalent across cognitive, organizational, and computational substrates — was made explicit in Itti and Koch's (2001) computational saliency-map framework, which mechanizes the bottom-up control of attentional deployment in a form portable across substrates. ^[15]

References

[1] James, W. (1890). The Principles of Psychology (Vol. 1, Ch. 11: Attention). Henry Holt and Company. Verbatim source of the quoted definition of attention as the mind's "taking possession ... of one out of what seem several simultaneously possible objects or trains of thought"; supports FACT-031 exactly. ↩

[2] Broadbent, D. E. (1958). Perception and Communication. Pergamon Press. Foundational filter model of selective attention: a limited-capacity channel forces early selection among competing stimuli; directly supports FACT-032's claim that Broadbent formalized filter theory of selective attention. ↩

[3] Kahneman, D. (1973). Attention and Effort. Prentice-Hall. Canonical limited-capacity (effort) model of attention as a graded mental resource flexibly allocated across tasks; supports FACT-033's unified capacity-model claim. ↩

[4] Corbetta, M., & Shulman, G. L. (2002). "Control of goal-directed and stimulus-driven attention in the brain". Nature Reviews Neuroscience, 3(3), 201-215. Maps dorsal (goal-directed/top-down) and ventral (stimulus-driven/bottom-up) frontoparietal networks; supports FACT-034 exactly. ↩

[5] Mack, A., & Rock, I. (1998). Inattentional Blindness. MIT Press. Empirical demonstration that observers fail to perceive fully-visible unexpected stimuli when attention is engaged elsewhere; supports FACT-035's consequence-asymmetry claim. ↩

[6] Wickens, C. D. (2008). "Multiple resources and mental workload". Human Factors, 50(3), 449-455. Multiple-resource theory of dual-task interference by shared modality/code/stage; supports FACT-036's mental-workload / filter-failure claim. ↩

[7] Posner, M. I., & Petersen, S. E. (1990). "The attention system of the human brain". Annual Review of Neuroscience, 13(1), 25-42. Foundational decomposition of attention into posterior (orienting), anterior (executive), and vigilance/alerting systems; supports FACT-037's cognitive-psychology/neuroscience listing. ↩

[8] Simon, H. A. (1971). "Designing organizations for an information-rich world". In M. Greenberger (Ed.), Computers, Communications, and the Public Interest (pp. 37-72). Johns Hopkins University Press. Coins the attention-economy idea ("a wealth of information creates a poverty of attention"); supports FACT-038 exactly. ↩

[9] Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, Ł., & Polosukhin, I. (2017). "Attention is all you need". In Advances in Neural Information Processing Systems 30 (NeurIPS 2017) (pp. 5998-6008). Introduces the Transformer with multi-head attention as the sole sequence-mixing mechanism; supports FACT-039 exactly. ↩

[10] Tanenbaum, A. S., & Bos, H. (2014). Modern Operating Systems (4^th ed.). Pearson. Standard OS text developing process scheduling, interrupt handling, and event-driven I/O as the OS-level analogue of attentional gating; supports FACT-040. ↩

[11] Ocasio, W. (1997). "Towards an attention-based view of the firm". Strategic Management Journal, 18(S1), 187-206. Treats firm behavior as the outcome of how an organization channels the bounded attention of decision-makers; supports FACT-041 exactly. ↩

[12] Desimone, R., & Duncan, J. (1995). "Neural mechanisms of selective visual attention". Annual Review of Neuroscience, 18(1), 193-222. Biased-competition account in which parietal/frontal networks bias sensory-cortex competition; supports FACT-042's neuroscience/gain-modulation claim. ↩

[13] Davenport, T. H., & Beck, J. C. (2001). The Attention Economy: Understanding the New Currency of Business. Harvard Business School Press. Business-strategy treatment of attention as scarce commodity (capture, markets, engagement); supports FACT-043. ↩

[14] Simons, D. J., & Chabris, C. F. (1999). "Gorillas in our midst: Sustained inattentional blindness for dynamic events". Perception, 28(9), 1059-1074. Classic demonstration that ~half of observers tracking a passing task miss a gorilla-suited person; supports the consequence-asymmetry claim at FACT-044 exactly. ↩

[15] Itti, L., & Koch, C. (2001). "Computational modelling of visual attention". Nature Reviews Neuroscience, 2(3), 194-203. Canonical saliency-map framework formalizing bottom-up attentional deployment as a substrate-portable computation; supports FACT-045 exactly. ↩