Attention
in a Hurry

The Whole Thing in One Page

Attention is usually pictured as a beam you point, a span you possess or a tank the phone drains. None is quite right. Attention is a selection system. At every moment, more sensations, memories, impulses and possible actions are available than the mind can process with equal depth. Something has to receive priority. Focus is the maintained relationship between that priority and an intended task: selecting it, keeping its state available, noticing when it has been displaced and restoring it.

There is therefore no context-free number called the human attention span. A person may miss a flash during one task, wait for a faint signal across minutes in another and follow a film for hours. The famous eight-second comparison with a goldfish presents unsupported numbers as one biological limit.

Selection has consequences. Ask people to count basketball passes and some fail to see a person in a gorilla suit cross the scene. Their eyes receive the event. Their task makes it lose. The same protection that lets you read in a café can hide what is outside the chosen channel. Attention is powerful because it excludes.

Three sources of priority keep interacting. Your current goal favours what is relevant. Physical salience favours what is sudden, bright, loud or moving. Selection history favours what has predicted reward, relief, danger or social information before. The last explains why a small icon can outrank a worthy intention. It has training behind it.

The process runs on a narrow working surface. Working memory keeps the rule, sentence, number or next step needed now. A demanding task uses much of that surface. An alert can briefly disrupt processing without causing a switch. A switch can replace the task state. Returning can then require its reconstruction. These are different costs, and none supports a universal claim that every interruption consumes twenty-three minutes.

Sustained attention is not a heroic stare held at one strength. Performance fluctuates with task structure, arousal, sleep, effort, reward, uncertainty and mind wandering. Departures are normal. The practical questions are whether they are detected, whether they serve any purpose and whether the route back remains visible.

A smartphone is unusually competitive because it compresses many possible actions into one reachable object: messages, work, maps, news, games, photographs, social approval, unfinished conversations and an endless supply of something easier than the present task. Yet screen time is not one exposure, and research does not show that a phone emits a universal field of cognitive damage. Notifications can disrupt. Switching can cost. Mere-presence findings are mixed. In one 2025 randomised trial, blocking mobile internet for two weeks while preserving calls and texts produced a modest improvement in sustained attention among a motivated sample, but adherence was difficult.

The answer is neither denial nor exile. Give the period one job. Decide in advance what may interrupt it. Reduce easy routes into unrelated activity. Make the intended next action concrete. Leave a cue that stores where to return. Use the phone as a tool by refusing to let all its tools remain active at once.

You do not need perfect attention. You need better selection and a reliable way back.

That is the book.

Why You Should Care

You can look straight at something and fail to see it.

In the best-known demonstration, viewers watch two groups pass basketballs and count the passes made by one team. During the action, a person in a gorilla suit walks through the scene, stops, faces the camera, beats their chest and leaves. Many viewers miss the gorilla. Detection varies with the version and the instructions, which is why a single headline percentage is misleading. The durable finding is stranger than the number: obvious visibility does not guarantee awareness. A sufficiently strong task can make an event vanish in plain sight.

That is attention doing its job. A system that admitted everything equally would not be open-minded. It would be unusable. You could not follow a sentence while registering every colour at the edge of vision, every pressure from your clothes, every remembered obligation, every nearby voice and every possible action. Selection gives thought a shape. It also decides what never enters the shape at all.

The failures are often invisible from inside. You experience the item that won, not the alternatives that were filtered out. Confidence therefore tells you less than it seems. A driver can believe they noticed the road, a student can believe they followed the lecture and a reader can reach the bottom of a page before discovering that the words left no trace.

This matters because almost every important activity depends on choosing well. Reading requires the current sentence to outrank the urge to scan ahead. Driving requires a pedestrian to defeat the conversation in your head. Listening requires another person’s words to survive your preparation of a reply. Learning requires the relevant pattern to remain active long enough to connect with what you know. Good judgement requires an inconvenient fact to receive processing rather than being screened out by a comfortable story.

The phone enters this system at an awkward angle. It is useful, personal, socially loaded, cheap to check and close enough to reach before a conscious plan has finished forming. It carries real obligations beside engineered amusements. A message from your partner, a banking alert, a work request and an algorithmic recommendation arrive through the same rectangle. Advice that says “remove the distraction” therefore mistakes the problem. You may need the map, the authenticator, the ticket, the camera and the call. Quitting the device can be an experiment. It is not a complete theory of focus.

Nor is moral language much help. Calling yourself lazy explains nothing. Calling the phone addictive may describe severe impairment for some people, but it turns every ordinary lapse into a diagnosis and hides the mechanics you can change. The useful questions are smaller and less flattering. Which cue started the switch? What did the new action promise? What information was displaced? Why was the original task hard to resume? Which interruption was necessary, and which one was permitted by default?

The answers give you leverage. They show why silence can fail when the phone remains visible, why a vague task invites escape, why an unfinished message follows you into the next job, why a break can restore performance in one setting and become a forty-minute feed in another, and why the same person can read a novel for two hours yet feel unable to face a spreadsheet for ten minutes. There is no contradiction. Attention is sensitive to content, skill, state, value and structure.

By the end, you should be able to distinguish capture from choice, concentration from working memory, mind wandering from interruption, habit from addiction and useful connectivity from permanent availability. Focus also has four jobs: select, maintain, notice departure and return. More importantly, you should know how to change the conditions before asking character to do all the work.

The aim is not a life with nothing competing for you. It is a life in which you decide more of what deserves to win.

The Core Ideas

Attention Is Selection, Not a Substance

You are reading these words while your eyes receive far more than them. The page has an edge. Light falls differently across it. Your hands touch something. Sounds enter. Thoughts wait. Most of this is available to the nervous system and absent from the experience you would report. Attention is the set of processes that makes some information more effective than the rest.

That definition is less tidy than a spotlight, but it explains more. Selecting one voice from a room is selective attention. Staying ready for its next words calls on sustained attention. Alerting prepares you to respond; orienting directs processing towards a place or event. Executive control helps keep the governing rule in charge, especially when a familiar response would be wrong. In the basketball task, count the white shirts must survive the temptation to follow everything moving. These operations work together. Their overlapping systems are better understood as ways of controlling information than as separate supplies of mental fuel.

The need for selection comes from a mismatch. The world and the mind offer more candidates than can guide deliberate thought and action at once. A bottleneck is often treated as a defect, yet without one there would be no coherent action. To answer a question, you must privilege the question. To cross a road, you must let movement near the kerb defeat much else. To read, you must keep earlier words available while refusing most other meanings and sensations entry.

Selection also changes action before it changes conscious report. A name at a party can reorient you while the surrounding speech remains a blur. A driver can brake in response to motion before identifying its source. Experimental tasks capture fragments of these operations through reaction time, accuracy and eye movements. None is an attention thermometer. Each asks what received enough priority to alter a response under specified conditions.

The cost is blindness to the rejected material. In the invisible-gorilla experiment, the gorilla is not too faint, too quick or hidden behind scenery. It is missed because the viewer has selected a demanding stream and organised perception around it. Change the task and the same event becomes obvious. Attention does not copy the world into awareness and then choose what to care about. Choice alters what becomes reportable in the first place.

This is why “attention span” is a poor master concept. Performance depends on what must be selected, what competes with it, how often a response is required, what the person knows and why the outcome matters. Waiting for a rare target, following a conversation, proofreading numbers and watching a film are not four readings from one mental stopwatch. A person who cannot sustain one may remain absorbed in another for hours without possessing two different brains.

Nor is attention one liquid that every task drains at the same rate. Effort and fatigue are real, and limited capacity is real, but the tank metaphor hides changes in motivation, arousal, strategy and task structure. It also encourages the wrong question: how much attention do I have left? The better question is what is being selected now, what is competing with it and what would make the intended target easier to keep active.

Focus begins with selection, but it does not end there. The chosen task must remain available, departure must be detected, and the task state must be restored after interruption or drift. A person who returns well can focus effectively without remaining mentally motionless.

Control Combines Goals, Salience and Selection History

Imagine looking for a friend in a station. A red coat helps because your goal gives red priority. A dropped suitcase makes you turn because sudden movement and sound are salient. A notification badge attracts you because that symbol has predicted messages, news, approval or relief many times before. Current purpose, physical events and experience can each shape what receives attention.

Older accounts often divided control into top-down selection and bottom-up capture. The first meant guidance by goals; the second meant capture by physical salience. The distinction remains useful, but it omits selection history. What has been rewarded, feared or repeatedly chosen can acquire priority even when it is neither the current goal nor the loudest object.

Laboratory studies make this visible. Participants first learn that finding a target of a certain colour brings a larger reward. Later, in a different task where that colour is irrelevant, an item carrying it can still slow search by attracting attention. The reward is gone. Its learned priority remains. This is value-driven attentional capture: previous consequence has altered present selection.

A phone is dense with learned significance. The sound may be quiet and the icon small, but neither arrives neutrally. One cue has meant your partner. Another has meant work trouble. Another has sometimes delivered approval and sometimes nothing, keeping uncertainty alive. Opening the device has often removed boredom or the discomfort of not knowing. Repetition makes the action easier to start and the cue harder to dismiss.

The learned value can be pleasant or aversive. A cue associated with criticism, loss or an unfinished duty may attract attention without offering enjoyment. Many checks are attempts to settle uncertainty, prevent trouble or remove a nagging possibility. Reward language becomes misleading when it is heard only as pleasure. Information can gain motivational value because it reduces not knowing.

Choice has not disappeared. Deliberate goals operate alongside priorities trained by experience. Practice can strengthen useful selection too. Expertise teaches a reader, musician or clinician which pattern matters and how several elements belong together. Preparation can keep a task goal active before an attractive alternative appears. Learning supplies distractions with force, but it also helps intended targets survive them.

Physical salience deserves the same fairness. Orienting towards a bang, a face or sudden motion is often intelligent. A system that waited for conscious permission before inspecting abrupt change would be unsafe. Trouble begins when environments can manufacture urgency more cheaply than events justify it. A coloured badge borrows the visual grammar of warning. A vibration gives a routine update the bodily force of something happening nearby.

Control therefore does not require abolishing involuntary attention. It requires arranging priorities so useful capture remains possible while trivial capture becomes rarer. Emergency contacts may break through. Marketing alerts need not. A calendar warning may be time-critical. A count on an entertainment icon can wait until you choose to look.

The same principle applies beyond phones. Put the relevant document in view, state the question before opening sources, remove objects that cue another activity and rehearse the route back to the task. Attention follows intention, physical change and learned history. Focus improves when the environment stops repeatedly favouring the wrong one.

Working Memory Is the Narrow Desk

A difficult task has a hidden surface. On it sit the items needed for the next mental move: the sentence you are shaping, the number you carried, the rule you must apply, the question the meeting is meant to answer. Psychologists call the system that maintains and manipulates such immediately useful information working memory.

It is tempting to picture a tiny box holding a fixed number of objects. Research does not support one universal capacity that ignores task, representation and expertise. Familiar material can be grouped into meaningful chunks. Visual and verbal demands can interfere differently. Strategy matters. People vary. Yet the practical fact survives every theoretical dispute: only a limited amount can remain available for controlled use, and interference can push the wrong material onto the surface.

Attention and working memory are therefore close but not identical. Attention selects what receives priority. Working memory helps preserve selected information when it is no longer in front of you or when several steps must be coordinated. You can attend to a road sign without storing it for long. You can hold an address in mind while looking elsewhere. In demanding work, the two systems support each other: the goal guides selection, while the maintained goal tells selection what counts as relevant.

Expertise changes the load without abolishing it. A chess master sees familiar configurations where a novice sees separate pieces. A fluent reader groups letters into words without deliberate assembly. Knowledge compresses many elements into structures that can be handled together, leaving more control for the unusual part. This is one reason practice can make sustained performance look effortless while a beginner tires quickly on the same visible task.

This explains why vague work is fragile. “Sort out the proposal” does not specify the next operation. The mind must keep searching for what the task is while also doing it. “Rewrite the pricing paragraph so the three options can be compared” creates a smaller task set. It identifies the object, the action and the criterion. Less of the desk is spent deciding what belongs on the desk.

An interruption competes for the same space. A message can introduce a name, request, emotion and pending reply. Even if the exchange lasts seconds, those contents may replace the place you had reached in the original task. Returning then requires more than looking back at the screen. You must retrieve the goal, reconstruct the intermediate state and decide the next move. If the work had no visible structure, there may be little to cue that reconstruction.

Heavy working-memory load can also make irrelevant material harder to resist. When much of working memory is occupied, there is less control available to maintain priorities and resolve conflict. This is one reason a notification that seems harmless during routine filing can be ruinous during algebra, coding or a difficult paragraph. The alert is the same. The desk is not.

The phone is not always the enemy here. It can carry the calendar, checklist, map and external memory that remove demands from the head. Writing down an intention can reduce the need to keep rehearsing it; a precise reminder or photograph can preserve what the next step will need. The distinction is functional. A tool supports the active goal. A competing feed replaces it.

Good task design protects the surface. Put needed information where it can be seen. Close unrelated windows. Write the next step before a break. Keep a scratch line for intrusive thoughts so they need not be rehearsed. Reduce the number of unresolved decisions inside the work. These moves do not enlarge working memory. They stop wasting it on reconstruction and competition.

Sustained Attention Depends on Time, Reward and State

Imagine a job that asks you to watch a gauge and respond only when the reading changes. For long stretches, doing the job correctly looks like doing nothing. Now compare watching a film: the image keeps changing, events raise questions and the next scene helps answer them. Both can occupy an hour, but they offer attention different reasons to stay. Duration alone tells you little about the demands of sustained attention.

Performance often worsens during long, repetitive vigilance tests. Responses slow, misses rise or variability increases. One explanation is that maintaining attention uses limited resources. Another concerns control over those resources: as watching the gauge becomes familiar, the rule may lose priority and an unfinished conversation may take its place. Changes in arousal and the value of continuing also matter. These accounts are not mutually exclusive, and research has not reduced vigilance to one cause. The distinction has a practical consequence. A lapse might call for rest, a clearer reminder of the goal or a change in the task, rather than the same instruction to try harder.

Attention also fluctuates before a clear failure appears. Periods of stable, accurate performance alternate with noisier periods. This matters because the subjective feeling of concentration is an unreliable meter. You can feel busy while making variable responses, or feel effort without losing accuracy. Fatigue, boredom and effort are signals about the present relationship between person and task, not laboratory readings of one emptying tank.

Mind wandering belongs inside this model. Thoughts decouple from the current environment and move towards memories, plans, concerns or imagined scenes. During a reading test, that often predicts poorer comprehension because the internal stream competes with the text. During a walk or routine activity, the same capacity may support planning or incubation. The question is not whether mind wandering is good. It is whether it fits the task and whether you notice the departure in time.

Meta-awareness is the hinge. A lapse detected after one sentence is cheaper than a lapse detected after five pages. Practices that improve monitoring can therefore help without eliminating spontaneous thought. So can a task with visible feedback, a clear endpoint or enough challenge to keep the goal active. The aim is not to hold consciousness motionless. It is to shorten unchosen absences and place freer thought where it can do useful work.

State changes selection. Sleep loss impairs vigilance and control. Pain, stress, low mood, anxiety, medication, illness and hunger can alter arousal, working memory or the value of escape. Some people face persistent attention difficulties linked to ADHD or other clinical conditions. A productivity trick cannot diagnose or treat those causes, and repeated failure despite sensible structure may justify professional assessment rather than harsher self-judgement.

Breaks can help, but there is no sacred interval. In one experiment, brief and rare disengagement from a repetitive goal prevented the usual decline. That does not prove that every mind should work for twenty-five, fifty-two or ninety minutes. A break may restore movement and perspective, or it may introduce a stronger competitor and destroy the return path. Its value depends on when it occurs, what replaces the task and how easily work can resume.

Sustained attention is therefore managed, not commanded. Match the length and difficulty of the session to the task and current state. Add feedback where the work is flat. Move before alertness collapses. Schedule open thought rather than treating every drift as misconduct. Above all, judge focus by useful contact with the task, not by how long you remained seated beside it.

Distraction Has More Than One Cost

Two activities can coexist when their demands do not collide. Consider walking a familiar route while talking to a friend. The steps need little deliberate supervision, leaving room for the conversation. At a confusing junction, you may stop talking to work out the turn. The same pair of activities has become harder to combine because one now needs more control. Saying that humans “cannot multitask” misses this dependence on demand.

Trouble appears when tasks require the same limited operation at the same time, or when control must alternate between different rules. Reading a message and composing a sentence both demand language. Choosing a response in traffic and making a complicated decision on a call both demand response selection. Holding one set of figures while interpreting another burdens working memory. What feels like parallel work often becomes rapid alternation.

Task switching has measurable costs. The mind must activate the new task set, reduce interference from the old one and map incoming information to different responses. Preparation and practice can reduce the delay, but demanding switches often remain slower or more error-prone than repetitions. One change may look trivial. Repeated across a day, especially where complex tasks keep losing their place, switching can alter what kind of work remains feasible.

An alert, a switch and a return are not the same event. A notification may cause a short processing disturbance even when the person never touches the phone. A self-initiated check may create a switch without any alert. Resumption begins only after the original task has been suspended and must be restarted. A claim about one cannot be transferred automatically to the others, and none by itself measures long-term attention.

Resumption has its own mechanism. Before interruption, the task had a state: what had been completed, what remained uncertain and what should happen next. During the new demand, that state loses activation while other material occupies attention and working memory. On return, the old goal must be retrieved and its intermediate state reconstructed. Studies show that a brief chance to encode a return intention can help, which is why “next: compare column C with supplier invoice” may be worth more than an elaborate time-management system.

Incomplete work can also follow you forward. Experiments on attention residue found that people switching away from unfinished tasks sometimes performed worse on a later task while part of the first remained active. The result is setting-specific and does not establish fixed residue after every switch. It does capture a plausible route: unresolved goals keep prompting thought and retrieval.

Phones magnify several routes at once. External interruptions announce themselves through sound, vibration, banners and badges. Self-interruptions begin without an alert: a thought about the weather, an awkward pause, the discomfort of a hard step. Because access is immediate, the thought can become action before it fades. One check can then create fresh unfinished business, such as a message seen but not answered, which travels back into the original work.

Visible duration is therefore a poor measure of cost. A ten-second glance might be cheap at a natural boundary and expensive in the middle of a reasoning chain. There is no universal twenty-three-minute penalty. The relevant losses may be acute slowing, switch delay, error, forgotten state, lingering concern or a shallower strategy adopted because another interruption is expected.

Some interruptions deserve those costs. A medical alert, a child who needs help, a colleague in trouble or a changing market can outrank the present task. Focus is not a moral claim that uninterrupted work must always win. The avoidable loss comes from allowing every channel to claim emergency status by default.

Good control places gates around switching. Decide which people and events may interrupt immediately. Check lower-priority channels at boundaries. Before a necessary switch, leave a resumption cue. Finish a meaningful unit where possible, because a closed loop leaves less to recover. The aim is not zero switching. It is to know which cost is being paid and whether the new task is worth it.

The Phone Is an Environment, Not a Single Cause

The phrase “phone use” compresses incompatible activities. Following a map, filming a child, reading a novel, checking a glucose monitor, arguing in a group chat and watching short videos all illuminate the same screen. Their cognitive demands, social meanings and consequences differ. Counting the minutes and calling them one exposure is like measuring diet by the time spent chewing.

What makes the smartphone distinctive is the arrangement. It is carried close, wakes quickly, remembers identity, knows location and combines private tools with public channels. Useful actions and tempting actions share the same entry point. The device also makes checking cheap. A faint question that once died before you reached a computer can now be answered, replaced by another question and forgotten.

One way to catch the disturbance is to make people do something deceptively awkward. In a Stroop task, the word GREEN might appear in red: the job is to respond to the ink colour, not read the word. A 2026 experiment asked 180 university students to perform this task while smartphone-style social-media notifications appeared. Responses slowed for about seven seconds after an alert. Stronger disruption was linked to inferred relevance and more frequent habitual interaction. The finding is an acute laboratory effect, not a seven-second recovery rule for every message. It shows that an alert can interfere even without a full switch into another task.

Mere presence is less certain than popular accounts suggest. Early studies reported that a nearby personal phone reduced working-memory or reasoning performance. A 2024 meta-analysis covering fifty-six study samples and 7,093 participants found a small negative pooled effect for working-memory capacity, but no reliable effects for the other tested functions. Another 2024 synthesis, combining results across cognitive functions, found no significant overall effect. The distinction matters: a small result on one kind of test is not evidence that every mental operation suffers. Methods varied and many studies were underpowered. Moving the phone away remains a sensible low-cost experiment for demanding work. It is not proof that a switched-off rectangle drains every nearby brain.

Longer-term claims face a harder measurement problem. People estimate their own media use poorly. A meta-analysis comparing self-reports with device logs found moderate correlations and poor absolute accuracy. Logs are better at duration but still do not reveal whether the person was navigating, learning, working or escaping. Correlations between heavier use and poorer attention may reflect effects in either direction, shared causes or differences in what people do on the device.

A more direct test changes what the phone can do. In a 2025 randomised trial, 467 participants agreed to block mobile internet on their smartphones for two weeks while retaining calls, texts and internet on other devices. Only 266 installed the blocking application and 119 met the adherence threshold. Objective sustained attention improved modestly in the available follow-up data, including in an analysis that did not require adherence. Most participants were already motivated to reduce use, and the comparison group waited rather than receiving an alternative treatment. The intervention offers causal evidence, but also a practical warning: an arrangement can help in a trial and still be difficult to live with.

The clean conclusion is functional. Phones can fragment attention through alerts, easy switching, learned cues and displacement of sleep or sustained activity. They can support attention through reminders, navigation, external memory, accessibility tools and contact with people who matter. The device is a bundle of capabilities governed by settings, apps, expectations and habits.

That makes configuration part of psychology. The phone you carry after disabling commercial alerts, separating work accounts, removing feed shortcuts and defining urgent contacts is not the same attentional environment as the phone delivered by default. Hardware may match. The available actions and interruption rules do not.

Focus Is Designed Before It Is Defended

The common plan is to wait until distraction appears and then resist it. This places control at the most expensive point. The cue is present, the alternative is available and its learned significance is active. Success now requires a deliberate refusal, followed by another refusal when the same condition returns.

A stronger plan makes several decisions earlier. Research on self-control distinguishes strategies that change situations from those used after temptation has entered attention. Both can work. Situation design has one advantage: it reduces the number of conflicts that need a late response. Turning off a badge does not prove virtue. It removes a prompt that carried no right to interrupt.

Begin with the target. “Work on accounts” gives attention no clear object. Define the next visible action, the relevant materials and a finish condition. Place them in view. Remove decisions that can be made beforehand. An implementation intention can connect a cue to action: when I sit at the desk after lunch, I open the reconciliation sheet and process the first unmatched payment. Its value lies in specificity, not ritual wording.

Then write the interruption rule. Identify the people or events for which delay until the next boundary would cause material harm. Give those a trusted route. Let routine messages, news, shopping, entertainment and group chatter wait. Batching can reduce repeated capture for some people, but emergency work, caregiving and customer service require different rules. The design must count the cost of missing information as well as the cost of receiving it.

Make the wrong first action slightly harder. Disable alerts that carry no time-sensitive value. Keep feeds off the first screen. Use distance when a task state is fragile. Add a login, blocker or scheduled window where one touch has become too cheap. Friction should be selective. The map, camera or authenticator may need instant access; an endless feed does not.

Protect return as deliberately as starting. Before a planned break, leave a marker at the exact next step. Keep a capture line for unrelated thoughts so they can be stored without being obeyed. After an interruption, reconstruct the task goal before opening another channel. Departure is normal. Losing the route back is optional more often than it appears.

State comes before moral effort. Sleep, pain, anxiety, illness and overload can weaken vigilance, working memory or tolerance for difficult work. Movement, food, treatment, a lighter task or a proper stop may improve control more than another restriction. This does not explain every avoidance through state. It recognises that attention belongs to a body and a situation.

The phone can implement much of the arrangement. Current iPhone and Android guidance describes scheduled interruption modes with exceptions for selected contacts or applications. Some Android phones also offer a Focus mode that pauses chosen apps, while names and behaviour vary by device and software version. Treat the settings as tools rather than guarantees. Test whether urgent contacts can reach you, whether routine alerts stay out and whether a necessary function has been blocked.

Measure the design by behaviour and output, not purity. Did the chosen work receive enough coherent time to advance? Did necessary messages arrive in time? Were fewer switches automatic? Was the task easier to recover? A light arrangement that works is better than a severe system abandoned after three days.

The world contains more candidates than deliberate thought can serve at once. Focus does not remove that abundance. It gives one target a clearer claim, protects the state that lets work continue, notices when another action has taken over and leaves a route back. Design changes which action is easiest before the difficult moment arrives.

How It Actually Works

A difficult paragraph is open. The first sentence exists; the second does not. Several facts must be held together, and none has yet arranged itself into the right order. This is an illustrative sequence assembled from common mechanisms, not a report of one person’s day. Then the phone lights.

No hand has moved. The light is already competing with the sentence.

The target before the cue

Before the light, attention was already active in several ways. The eyes selected the text. Working memory held the intended claim and pieces of evidence. A task set specified what counted as progress: produce a sentence that connects them. Competing thoughts were present but weak enough to lose.

This state can feel like staring at a document, which makes it easy to underestimate. Much controlled work happens before an observable response. Alternatives are generated and rejected. A word is retrieved, tested against the sentence and withheld. The task goal gives these internal events priority.

The state is also fragile because the next move has not been externalised. A finished sentence can be found again. A half-formed relation among ideas exists mainly in active organisation. When people say they were “in the middle of a thought”, the middle is a working state, not a location. It can decay, be overwritten or lose priority.

Task structure changes that fragility. If the paragraph has a heading, a visible outline and a note saying what the sentence must do, the active state has supports outside the head. If the instruction is “write something good”, more of the work depends on maintaining an uncertain goal. The phone has not changed, yet the same cue now meets two different task states.

The task has an attentional shape

The task is not a passive victim. It produces cues and consequences of its own. A clear question, visible progress or approaching finish can pull attention forward. Confusion, delayed reward and fear of error can make the next step feel costly. Attention then begins to sample alternatives before any device intervenes.

This is why distraction often clusters at transitions. Starting asks the mind to replace a broad field of possibilities with one rule. A difficult middle asks it to tolerate uncertainty without immediate proof of progress. Near completion, a task may hold attention because the endpoint is visible, or lose it because the interesting problem has been solved and only checking remains. The pattern depends on what the work is demanding at that moment.

Emotion enters through the same route. An ambiguous email can occupy attention because it may signal criticism. A financial figure can invite checking because a mistake would matter. A creative task can expose the person to judgement before it supplies any reward. Escaping to the phone then removes an unpleasant state for a moment. Repetition teaches more than the value of the feed; it teaches that difficulty predicts an exit.

Good preparation changes the pull of the task. A question supplies direction. A small finish condition makes reward nearer. Immediate feedback helps monitoring. Breaking a task into meaningful units can reduce uncertainty without reducing the work to trivial fragments. The aim is not to make every task entertaining. It is to make the next operation legible enough that attention has something concrete to select.

The cue changes priority

The screen’s light is physically salient. It changes suddenly in a stable visual field. A vibration or tone adds an auditory or tactile onset, the sort of event orienting systems are built to inspect. The first turn can occur before a considered judgement about whether the alert matters.

The cue is personal as well as physical. The phone is not a random lamp. It has delivered messages, requests, reassurance, bad news, payment, amusement and relief from awkward pauses. Those outcomes are uneven. Most checks are ordinary; a few matter. Uncertainty gives the cue a question: what is it this time?

At this stage, capture need not mean full conscious engagement. The light can slow the current response, increase monitoring or place the possibility of checking into working memory. Laboratory alerts can produce a brief processing disturbance even when the participant does not pick up the device. That acute effect is different from task switching, which begins when another set of rules takes over, and from resumption, which begins when the suspended work must be reconstructed. Ignoring can still consume control when the alternative remains active.

The cue can also be anticipated. If an important reply is expected, attention may periodically sample the phone without any signal. This is prospective memory at work: an intention to notice something later keeps reactivating. Silence does not remove the open question. A visible device can become a cue for the intention, though evidence for a broad mere-presence penalty is mixed.

The hand moves

The check often begins as a reasonable promise. It will take a second. It may be important. Knowing will remove the distraction. Each claim can be true. None guarantees that checking is the best move now.

The hand movement crosses an important boundary. Before it, the phone was a candidate action. After it, a new task has begun. The eyes leave the paragraph. The motor sequence is highly practised. Face recognition or a passcode removes the final gate. The device presents several possible destinations before the original purpose has been completed.

This is where low access cost matters. Many impulses disappear when action requires standing, walking to another room and opening a machine. A phone can turn them into immediate behaviour. The thought “has the parcel arrived?” can become a tracking page before the thought is judged. The answer may then display an email count, a news banner and an icon with a number. One small question has opened a market.

The first switch requires control. The new task set may be “identify the alert and decide whether to respond”. If the message is simple, that state is easy to build. If it carries conflict, ambiguity or social consequence, it brings more material onto the narrow desk. The original paragraph does not vanish, but its activation falls as the new problem gains priority.

The check becomes a chain

Suppose the alert is unimportant. The rational action is to close the phone. Yet the device is now open, held and populated with learned cues. The cost of checking one more thing is lower than the cost that brought the phone into the hand. Behaviour follows paths as well as intentions.

A badge suggests unfinished business. A feed offers no natural stopping point because each item reveals the next. A message creates a reply that creates a search that creates another message. These designs belong to the wider history and economics of social media, but their attentional function is plain: they keep supplying new candidates before the current one becomes stale.

Uncertain outcomes can help sustain checking without requiring every outcome to be pleasurable. One refresh brings nothing. Another brings a useful answer. Another brings praise, conflict or a story that changes the day. Dopamine contributes to learning about prediction and motivational significance, but saying that each check produces a “dopamine hit” explains too much with too little. People also check from anxiety, duty, boredom, habit and the wish to close an uncertainty.

Time estimation becomes difficult inside a chain of short units. No single item feels like a decision to spend fifteen minutes. Each is one more response. The original task competes poorly because returning requires reconstruction, while continuing requires a thumb movement. This asymmetry can determine the next minute.

Not every chain is harmful. A work message may lead to a necessary call. A map check may prevent a missed train. The mechanism does not decide the priority. It shows how a justified first switch can become several unjustified later ones unless the path has a stopping rule.

Availability becomes an agreement

A phone also connects one person’s attention to other people’s expectations. Once rapid contact is possible, delayed response can acquire meaning. Silence may look careless, hostile or unsafe. In some workplaces, a green status light becomes evidence of diligence. Families create informal rules about who must answer. The interrupting cue therefore carries a social contract, even when nobody has stated it.

This changes what counts as voluntary. A person may choose to check because the cost of missing a request is real. Advice to disable everything can transfer that cost to colleagues, customers or dependants. The useful unit is not the solitary brain but the communication system around it. Who sends what, through which channel, with what promised response time?

Many systems never answer those questions. Every app may notify immediately, while senders use urgency, convenience and habit interchangeably. The recipient then performs triage at the moment each alert arrives. That is an expensive place to make policy. A better arrangement decides beforehand that calls from named people can break through, operational incidents use one channel, ordinary messages wait and promotional messages never interrupt.

The same design protects senders. An emergency route is more trustworthy when routine traffic is kept out of it. A scheduled reply window can improve communication if others know when it occurs. Focus and responsiveness are not strict opposites. Both improve when the system distinguishes a signal from the general fact that somebody has sent something.

The return is a retrieval problem

The phone closes. The paragraph is still on screen. The sentence fragment looks familiar, but familiarity is not the lost state. What relation was being formed? Which fact was meant to qualify which? Why had one word been rejected?

Return begins by retrieving the task goal. The visible document supplies cues, but the most useful intermediate state may never have been written. The mind samples the page, reconstructs context and tests possible next moves. This produces the experience of needing to “get back into it”. No single duration applies. A routine task with a checklist may resume at once. A complex proof, design or conversation may require substantial rebuilding.

Interruption research shows that preparation changes this process. When people receive a brief interval before being interrupted, they can encode where they are and what they intend to do next. A return cue then helps reactivate the suspended goal. In ordinary work, this can be a few words typed before switching: “next sentence contrasts speed with accuracy”. The note stores the bridge that the finished page cannot show.

The new task may remain active too. A message left unanswered, a disagreement or a decision postponed can generate further thought after the screen closes. This is the plausible core of attention residue. It is not a substance coating the mind. It is a goal that has not been closed and continues to compete for retrieval.

The writer may now choose an easier action, perhaps formatting the heading or checking a source already known. This can look like work while avoiding the expensive reconstruction. The distraction has taken time and may also have made the next attempt shallower.

Fragmentation changes the work

The cost of repeated switches is not confined to minutes lost. It can change the strategy used on the task. When another interruption feels likely, people may favour actions that can be completed quickly, avoid opening a difficult chain of reasoning or keep several channels visible so nothing is missed. The work becomes shallower before the next alert arrives.

This effect is difficult to measure outside controlled settings because jobs differ and people choose their own interruptions. It should not be turned into a universal claim that every fragmented day destroys thought. Some roles depend on monitoring and rapid handover. A nurse, dispatcher or market trader may be skilled at short cycles supported by external systems. The relevant question is whether the task requires a state that the interruption pattern repeatedly erases.

Learning exposes the problem clearly. Understanding a hard idea requires selecting it, connecting it with prior knowledge and testing the connection. Frequent unrelated switches can interrupt those operations even if total study time remains unchanged. By contrast, a deliberate pause after a coherent unit can support retrieval or reveal an error. “Continuous” and “broken” are not enough. The boundary and the content of the break matter.

Long work also benefits from shifts of scale. A designer may leave a problem and return with a better relation visible. That is not the same as letting every moment of uncertainty trigger a feed. A planned pause can leave room for incubation while keeping the question available. An impulsive check may replace it with competing unfinished material. The practical distinction is whether the original question remains available for return.

Repetition trains the next interruption

One episode changes little. A repeated sequence changes the field.

A hard or ambiguous moment occurs. The phone is checked. The hard moment ends, at least briefly. The action may also deliver novelty or social information. Over time, difficulty itself can become a cue for checking. No notification is needed. The blank space after a sentence, the pause before a decision or the first sign of boredom can start the motor sequence.

This helps explain self-interruption. Workplace switches can be initiated by the person rather than imposed by an alert. The environment still matters because it determines which escape is available, visible and cheap. A phone within reach can turn an internal fluctuation into an external task. Put it elsewhere and the impulse may pass, or attention may wander in another direction.

The loop can become context-specific. Someone may check constantly at one desk and read for an hour on a train. One app may pull while another does not. That contrast is hard to explain through one fixed span alone; learned relations among cue, setting and response offer another explanation. It also creates room for change. Habits are stubborn partly because familiar contexts keep presenting the first step.

Changing the loop does not require hating the reward. It requires finding the entry. If the sequence begins with visible badges, remove them. If it begins with uncertainty about urgent work, define an emergency channel. If it begins when the task becomes unclear, write the next action before that point. If it begins with fatigue, a blocker may delay the check while leaving the cause untouched.

A redesigned hour

Run the same episode under different conditions. The paragraph has one clear job. The needed notes are open. The phone is face down across the room, with calls from two contacts allowed through and routine messages held until a scheduled window. A capture line waits beside the draft.

The task still becomes difficult. A thought appears: check the delivery. It is written on the capture line rather than enacted. This does two things. The person can see that the intention has been postponed rather than forgotten, without having to rely entirely on rehearsal. The thought may return; the note gives a reason not to act on it yet.

After thirty-five minutes, concentration becomes noisy. There is no need to defend a preselected interval, so the person finishes the sentence and stops at a boundary. Before standing, they type: “next: test whether the second study used logged data”. The break involves water and movement, not the strongest competing environment in the room. It is short because it has a defined action, not because a timer possesses scientific authority.

At the message window, communication receives full attention. One issue deserves immediate work, so the switch is chosen. The paragraph’s return marker remains. When the issue is closed, the note reconstructs the next step without requiring the earlier thought to be rediscovered.

This hour contains mind wandering, an urge to check, a break and an interruption. It is still focused. The measure is not purity. The intended work retained enough priority to move, departures were detected, necessary communication was admitted and return was supported.

The design can be lighter. Some people need only silent notifications and a clear task. Others need a separate device, timed lock or mobile-internet block for a period. The right restriction is the least severe one that changes the behaviour without damaging the functions that matter. A system that depends on constant resentment towards the phone is unstable. A system that decides what the phone is for during this hour has a chance.

How we know

Attention is inferred from performance, not read from a single meter. Researchers use reaction time, errors, eye movements, brain activity, thought probes and tasks designed to stress particular operations. Each measure is partial. A Stroop task, visual search or continuous-performance test can reveal conflict or vigilance under controlled conditions without measuring a person’s complete capacity to study, drive or work.

Smartphone evidence adds further problems. Laboratory alerts have precise timing but thin meaning. Field studies contain real stakes but less control. Self-reported screen time matches device logs poorly, while logs record duration without purpose or content. “Phone use” and “attention” are both families of measures, so studies that share the labels may test different propositions.

The best current causal evidence comes from interventions, but those remain bounded. Notification experiments can isolate an acute response without measuring a working day. Mobile-internet blocks alter several activities at once, attract people willing to participate and can be difficult to maintain. Results should be carried at their measured size, task and population. The evidence supports mechanisms of capture, switching and resumption, plus practical experiments in cues, access and return. It does not establish one falling human attention span or one prescription for every phone owner.

What People Get Wrong

“Your attention span is shorter than a goldfish’s”

The claim arrives with numbers: humans once managed twelve seconds, now manage eight, while a goldfish manages nine. It sounds measured, modern and humiliating. It is also empty. Its precision helped it travel through presentations, articles and school materials, each repetition making the comparison look independently confirmed.

The best-known source is a graphic in a 2015 Microsoft Canada report. The graphic attributes the figures to a website called Statistic Brain, not to the report’s survey or neurological research. It defines neither the human task nor the fish task, gives no compatible population or method and compares unlike organisms as though attention were one timed reflex. The report’s own discussion is more careful, distinguishing sustained, selective and alternating attention.

There is no context-free span. Detection in a rapid visual task, vigilance across minutes, reading across an hour and absorption in a film measure different relations among person, target and competition. Digital environments may change habits and expectations. That question deserves evidence. The goldfish line supplies none.

The correction matters because a broken biological stopwatch invites despair. Task-specific attention can be redesigned, trained and supported. A species-wide eight-second limit cannot.

“Multitasking makes you more efficient”

The myth survives because switching feels busy. Several windows move, messages clear and no single task has had time to become uncomfortable. Visible activity is mistaken for combined progress. People also notice the message they answered and rarely observe the slower reasoning or extra error that followed, so the benefit is vivid and the cost hidden.

Some activities can overlap. Walking and talking often coexist. Practised routines can run with little deliberate control. Trouble appears when tasks compete for language, working memory, response selection or executive control. The mind then alternates rules, and experiments show slower or less accurate performance around switches even when people know the change is coming.

This does not mean every switch is foolish. Monitoring roles depend on it, urgent events deserve it and a deliberate change can help when work has reached a good boundary. The mistake is treating rapid alternation as free parallel processing.

Efficiency must be judged by output, errors and later reconstruction, not by the number of active channels. Two tasks completed in sequence can feel slower while finishing sooner. Multitasking often sells the sensation of momentum by hiding the bill.

“Putting the phone on silent solves the problem”

Silence removes one class of cue, which can help. It does not remove light, vibration, badges, visual presence, expectation or the habit of checking when work becomes difficult. It may even leave a person sampling the screen more often when they fear missing something, trading external alerts for self-initiated monitoring.

The stronger version of the claim, that a nearby phone always drains cognition even when switched off, also outruns the evidence. A 2024 meta-analysis found a small pooled disadvantage on tests of working-memory capacity, while estimates for several other functions did not differ reliably from zero and methods varied substantially. Some people and tasks may benefit from distance; the phone does not carry a measured cognitive radiation field.

The correction is to identify the active route. Sound calls for silence. Visual cues call for hiding. Anticipated urgent contact calls for a trusted exception channel. Automatic checking may require distance or friction. An unclear task needs clearer work, not another setting.

“Silent” is one tool. Treating it as the mechanism leaves the rest of the contest untouched.

“Dopamine is the pleasure chemical that makes the phone addictive”

This explanation is attractive because it turns a messy behaviour into one molecule and gives every notification a villainous “hit”. Neuroscience is less cinematic. No neurotransmitter maps cleanly onto one everyday feeling, and a brain image cannot tell whether a person’s checking is habit, anxiety, work or clinical compulsion.

Dopamine systems contribute to learning about reward prediction, motivation and incentive salience. Responses can reflect outcomes that are better or worse than expected and cues that have acquired motivational pull. Dopamine is not a direct pleasure meter. Research on addiction also distinguishes “wanting” from “liking”: a cue can drive pursuit even when enjoyment has faded.

That does not mean dopamine is irrelevant to learned checking. It means the word cannot do the explanatory work alone. People check for social information, relief from uncertainty, avoidance of effort, habit, duty, novelty and fear. Different routes require different interventions.

Frequent use is not by itself a clinical diagnosis. Serious impaired control and functional harm can occur around gaming or other digital behaviour, but ordinary distraction should not be medicalised to make it sound important. The useful question is what the cue predicts and what the check changes. “Dopamine” often ends the inquiry exactly where it should begin.

“Mind wandering means you have failed to focus”

Mind wandering often harms the task in front of you. Readers retain less when thought has left the page, and vigilance suffers when attention decouples from the signal. Calling every departure failure, however, confuses an outcome with a mental capacity. Spontaneous thought is common even in careful people; treating its appearance as misconduct adds frustration without improving detection.

Internally directed thought can rehearse a conversation, plan a future action, combine distant ideas or process a concern. Its value depends on timing. A thought about dinner may be useful during a walk and dangerous while driving. A creative association may help after sustained work and obstruct the step that makes the work real.

The key distinction is chosen placement and meta-awareness. Notice the drift, decide whether it deserves a later home and return before the task state is lost. Create periods where thought can roam without competing with safety or comprehension.

Good focus is not a blank mind. It is a mind whose departures are noticed and governed. The correction shifts effort from preventing thought to improving awareness, timing and return under real conditions.

“Willpower is the answer”

Willpower matters. People can inhibit a check, tolerate boredom and return to a task. Denying agency would be as crude as blaming every lapse on character.

The error is making late resistance carry the whole system. Once the cue is visible, access is immediate and the alternative has learned value, control is fighting on expensive ground. Repeated victories can also leave the cue central, because each refusal requires monitoring the thing supposedly being ignored. Research on self-control shows that people also succeed by selecting and changing situations before conflict peaks: removing cues, creating commitments, planning responses and avoiding repeated decisions.

This does not make environmental design a trick for weak people. Pilots use checklists, laboratories control contamination and businesses separate approval rights because reliable systems should not depend on perfect moment-to-moment judgement. Attention deserves the same engineering.

Use deliberate control where values conflict and circumstances cannot be arranged. Do not spend it proving you can resist an alert that never needed permission to appear. A good environment preserves choice by reducing pointless choices.

“You need to quit your phone”

Abstinence is clear, dramatic and occasionally useful. It can reveal what the device was displacing and break a learned sequence. It can also remove navigation, accessibility, work, banking, tickets, family contact and emergency communication. Moving the same feeds to a laptop may change little if the underlying cue and avoidance loop remain.

Evidence supports reduction experiments, not one universal exile. A 2025 randomised waitlist experiment removed mobile internet from participants’ phones for a fortnight but left ordinary calls and messages available. Objective sustained attention improved modestly. Adherence was difficult, and participants were unusually motivated to cut use. Other field work suggests that batching or reducing notifications can help in some settings.

The better rule is functional separation. Keep the capability whose benefit exceeds its attentional cost. Change the cue, access or timing of the rest. One person may need feeds removed from the phone. Another may need an hour across the room. A third may need stronger clinical or social support because the behaviour is causing serious impairment.

Quitting is an intervention, not a theory. The harder task is to make the device serve chosen purposes without allowing every purpose to compete at once.

Use It

The useful unit is not a heroic day. It is one period of work with a defined target, an interruption rule and a route back. Build those parts together. A blocker without a clear task protects an empty desk. A clear task with every channel open remains easy to displace. A quiet phone without a return cue still leaves you rebuilding after the interruption that mattered.

Diagnose the first switch

Do not begin with a screen-time total. Choose one episode in which the intended task lost priority and replay the sequence. What was the target? Which step had become difficult? Was there an external alert, an internal thought or a pause with no clear next action? What did the phone promise: information, relief, contact, novelty or escape? What made the original task hard to recover?

This turns “I was distracted” into a mechanism. Repeated alerts call for cue control. An unclear task calls for a visible next move. Worry about urgent contact calls for a trusted exception route. A feed opened at the first sign of difficulty may require friction at that transition. Fatigue may require recovery rather than a stricter lock.

Repeat the reconstruction across several episodes before declaring a universal problem. One switch may be accidental; a pattern identifies an entry point. Device logs can show which apps were opened, how often and for how long, but they do not reveal purpose. Pair the log with the task, cue and consequence. Ten deliberate map checks may matter less than one escape that dissolves a difficult hour.

Give the period one job

“Study”, “do admin” and “work on the presentation” are categories, not usable task sets. Name the object, operation and finish condition. Read pages 40 to 55 and write the three claims you would challenge. Match today’s payments to invoices until every unmatched line has a note. Rewrite the opening so the conflict is clear within two paragraphs.

Put the needed materials in the same field. Open the source, place the figures beside the draft and close unrelated tabs. Write down decisions already made so working memory does not keep reopening them. The next move should be visible enough that starting does not require a new strategy meeting.

This does not make difficult work pleasant. It makes the source of difficulty legible. You can then distinguish effort from fear, missing information from vagueness and genuine disagreement from avoidance. A phone competes most cheaply when the task offers no clear action at all.

Write the interruption rule

Most phones inherit a false constitution: any app granted notification access may request attention immediately, and the owner judges each claim after it arrives. Reverse the order. Decide which people or events become materially worse if they wait until the next boundary. Give those a route that can break through. Let routine work, news, shopping, entertainment and group chatter wait for a chosen window.

Current major mobile systems can schedule interruption modes and allow selected people or applications. Some also pause chosen apps, and names and menus vary. Use the least complicated arrangement that expresses the rule. Then test it. Confirm that an urgent contact can reach you, an ordinary notification cannot and a necessary tool still works. A setting that has never been tested is a hope, not a communication system.

Make the agreement social where needed. Tell colleagues when messages are checked and which channel means urgent. Set expectations with family rather than disappearing without explanation. In customer-facing or operational work, shared cover and escalation may protect concentration better than personal silence.

Add friction to the first wrong action. Remove feed shortcuts from the opening screen. Log out of a service opened without purpose. Put the phone beyond reach when proximity cues checking. Use a timed block where one touch has become too cheap. Preserve the camera, map, ticket, authenticator and emergency route if those functions matter. The aim is correspondence between purpose and access, not deprivation for display.

Protect the return path

Before a planned switch, write where the mind should land when it comes back. One line is enough if it preserves the missing relation: “next, test the French figure against the OECD definition” or “ask whether the customer needs speed or certainty”. A to-do list names the project. A return cue stores its active edge.

Use a capture line for unrelated thoughts. When “book train” or “reply to Sam” appears, record it and continue. There is less need to keep rehearsing the intention, and writing it down does not grant it immediate priority. After an unplanned interruption, read the return cue before opening another channel. Where none exists, reconstruct the present goal from the material on screen and write it before continuing.

Design breaks around the same principle. Stop at a meaningful boundary where possible, leave the cue, then choose an activity with a clean ending. Movement, water, food, daylight or a quiet change of view often ends more naturally than a feed. Do not worship a fixed interval. Dense reading, repetitive monitoring, creative work and routine processing fail on different schedules. Use evidence from the task: repeated rereading, rising error, restless switching, loss of the goal or a body that needs movement.

Review the system, not your purity

Judge the arrangement by behaviour and output. Did the important task receive enough coherent contact? Did necessary messages arrive in time? Which switches were chosen, and which began automatically? How much reconstruction followed the interruptions that remained? Did the system survive an ordinary week?

Change one or two conditions at a time. Keep the least restrictive design that works, because severe systems often collide with legitimate functions and are then abandoned. A complete mobile-internet block can be a revealing bounded experiment. Low adherence in the strongest recent trial is also a warning that an intervention can show an effect while being difficult to live with.

A lapse is data. Find the first point at which the sequence became cheap, unclear or socially compulsory, then alter that point. The purpose is not to prove that you can obey a rule. It is to make intended work easier to continue and easier to recover.

The limits

These lenses cannot remove the social and economic systems competing for attention. Some jobs punish delayed response. Some households offer no protected space. Caregiving, poverty, unsafe work and unpredictable demands can make clean boundaries unavailable. Advice designed for a solitary professional with control over time should not be presented as a universal psychology.

Nor can task design treat a clinical condition. Persistent attention difficulties may accompany ADHD, depression, anxiety, trauma, sleep disorders, pain, medication effects, substance use or other health problems. Pattern, duration, impairment and wider symptoms matter. Qualified assessment is appropriate when difficulties are broad, longstanding or damaging despite sensible support. A productivity system is not a diagnostic test.

Restrictions carry costs too. Missing a message can harm trust. Blocking mobile internet may remove an accessibility or safety function. Moving the phone away can increase anxiety for someone awaiting urgent news. Every intervention should be judged against the risk and burden it creates, not against an imagined ideal of disconnection.

Attention is also not the same as goodness or importance. People can concentrate intensely on fraud, resentment or trivia. A perfectly defended hour can serve the wrong goal. Psychology can improve how selection is organised. It cannot decide what deserves a life’s time.

The one thing to keep

Give every important period three things: one job, an interruption rule and a return address.

The rule works because departure is inevitable. A sound will matter. A thought will intrude. Someone will need you. Energy will change. Focus is not the preservation of one object in consciousness without movement. It is the ability to maintain useful contact with a chosen task across those changes and to reconstruct the task when contact breaks.

Your phone need not disappear. It does need a job description. During one period it may be a communicator for named people, during another a map, camera or music player, and later an open door to everything. Trouble begins when all those roles remain active at once and every possible future use is allowed to claim the present.

Consider the paragraph abandoned halfway through. You do not need to rebuild your personality to finish it. You need to recover the question it was answering, put the relevant material back in view and decide what can wait. Those are small operations, but they change who gets to allocate the next few minutes.

The phone may still be on the desk. A call may still matter more than the paragraph. The difference is that neither a flashing badge nor a moment of difficulty gets to make that decision by default.

Terms

Selective attention. Giving some input greater influence than competing input, such as following one voice in a crowded room. Selection enables coherent action but can make unattended events disappear from awareness.

Sustained attention. Maintaining useful sensitivity to a task or signal across time. It includes continuous comprehension and rare-target vigilance, which place different demands on arousal, monitoring and response.

Divided attention. Allocating performance across more than one task or source. Success depends on whether demands overlap; practised or distinct processes can coexist more easily than two controlled language tasks.

Executive attention. Control used to maintain goals, resolve conflict and inhibit a dominant response. It is tested through tasks that make relevant and irrelevant response tendencies compete.

Alerting. Achieving and maintaining readiness for incoming information. A warning cue can improve response speed, while sleep loss or low arousal can weaken readiness across a watch.

Orienting. Shifting priority towards a location, object, sound or internal representation. Orienting may follow a goal, a sudden event or a cue whose significance has been learned.

Salience. Here, physical prominence created by contrast, intensity, novelty or movement. A flashing light can attract selection without being important. Learned significance is a separate source of priority, though the two often combine.

Top-down control. Selection guided by current goals, expectations or rules. The term remains useful, but goals operate beside physical salience and the priorities built through previous experience.

Stimulus-driven capture. An involuntary shift caused by properties such as sudden onset, movement or loudness. It protects against change but lets manufactured signals borrow the force of natural warnings.

Selection history. The effect of previous choices, rewards and contexts on current priority. It explains why familiar cues can attract attention even when they are neither salient nor relevant now.

Value-driven attentional capture. Priority acquired when a feature has predicted reward or consequence. Laboratory targets can retain pull after reward ends; phone cues may gain similar learned significance through repetition.

Working memory. The limited system that maintains and manipulates information needed for present action. It keeps a rule, number or next step available while attention moves through the task.

Cognitive load. The demand a task places on limited working and control processes. Heavy demand on working memory can weaken resistance to interference, though load depends on expertise, representation and strategy rather than item count alone.

Task set. The currently active rules connecting stimuli, goals and responses. Switching work often requires replacing one task set with another, which helps explain delays and errors at transitions.

Task switching. Alternating between activities that require different rules or priorities. It can be necessary, but demanding switches often involve reconfiguration, interference and loss of the previous task state.

Switch cost. The slower or less accurate performance commonly observed after changing tasks compared with repeating one. Its size varies with preparation, practice, similarity, complexity and the measure used.

Interruption. An event or self-initiated action that suspends ongoing work before its natural boundary. Its cost includes the new demand and the later work of retrieving the suspended goal.

Resumption lag. The interval and processing required to restart an interrupted task. A visible note about the next step can shorten reconstruction by cueing the goal and intermediate state.

Attention residue. Continued competition from a previous, often unfinished task after switching. It is a useful description of lingering goal activation, not a fixed substance or universal recovery period.

Mind wandering. Thought becoming relatively independent of the current external task. It can impair reading or vigilance, yet may support planning and incubation when its timing fits the activity.

Meta-awareness. Awareness of the current contents and direction of thought. Detecting that the mind has left the task turns an unnoticed lapse into a choice about continuing or returning.

Vigilance decrement. Declining performance across time in repetitive monitoring tasks, often seen as slower responses, more misses or greater variability. Competing explanations include arousal, motivation, goal loss and mind wandering.

Attentional blink. Reduced detection of a second target when it appears shortly after a first within a rapid stream. It reveals time limits in consolidating selected information for report.

Stroop effect. Slower or less accurate colour naming when a word spells a conflicting colour. The task exposes competition between an instructed response and a highly practised reading response.

Focus. A sustained useful relation between a chosen target and action. It depends on selection, preservation of the task state, detection of departure and return; it does not require unbroken awareness or zero switching.

Signal detection theory. A framework separating sensitivity to a signal from the criterion used to report it. It clarifies why misses and false alarms depend on both ability and decision policy.

Habit. A response made more automatic by repetition in a stable context. Habits reduce decision effort but can keep running after the original reward or purpose has weakened.

Cue. An event, place, feeling or time that predicts an action or outcome. Phone light, boredom and a desk transition can each become cues for checking through repeated pairing.

Implementation intention. An if-then plan linking a specified situation to a response, such as opening the ledger after lunch. Precision helps the intended action start without a fresh decision.

Precommitment. A present choice that changes the options or costs facing a future self. Blocking access, scheduling communication or giving another person control can protect later priorities from immediate temptation.

Go Deeper

The accessible bridge: Adam Gazzaley and Larry D. Rosen, The Distracted Mind: Ancient Brains in a High-Tech World (MIT Press, 2016). This is the best next step for a reader who wants cognitive control, interference and technology brought into one readable account. Gazzaley is a neuroscientist and Rosen a psychologist, and their division of labour keeps the book moving between mechanisms and daily behaviour. Read it for the distinction between goal setting and the limits of carrying goals through a noisy environment. The technology examples are now dated in detail, and some claims predate later mixed findings on phone presence, but the operating model remains useful.

The original voice: William James, The Principles of Psychology, Volume I (Henry Holt, 1890), especially Chapter XI, “Attention”. James combined close observation with the reaction-time research of his day, long before brain imaging or smartphones. He captured the felt structure of attention with unusual accuracy: selection, effort, wandering and the difficulty of voluntarily holding an uninteresting object. The chapter is a primary source for psychology becoming a discipline while still sounding like philosophy and close observation. The terminology and physiology are nineteenth-century, so read it for questions and descriptions rather than current mechanism. It is freely available in reliable public-domain editions.

The technical map: Harold E. Pashler, The Psychology of Attention (MIT Press, 1998). This is the demanding option and the strongest antidote to slogans. Pashler surveys selective attention, dual-task interference, automaticity, perception, memory and central bottlenecks, showing how much depends on the exact task and measure. The book predates modern smartphones and much current neuroscience, which makes it useful for understanding the behavioural foundations without a technology morality play. Expect dense experimental detail. Read the opening chapters and the sections on dual-task performance rather than forcing a cover-to-cover march.

The political question: James Williams, Stand Out of Our Light: Freedom and Resistance in the Attention Economy (Cambridge University Press, 2018). Williams asks what follows when institutions can optimise for capturing attention at scale. His subject is not laboratory capacity but agency: whether repeated competition for selection changes what people are able to want, pursue and govern together. The argument is sharper and more normative than the evidence-led account in this book, so treat it as a major interpretation rather than a neutral synthesis. It is brief, provocative and useful for seeing why personal settings cannot settle every attentional problem. Pair it with the experimental books above so institutional criticism does not become a substitute for cognitive explanation.

Notes and Sources

Scope and definitions. Attention is not one agreed faculty with one measure. The working distinctions among alerting, orienting and executive control draw on Michael Posner and Steven Petersen’s network account and its twenty-year reassessment, plus Jin Fan and colleagues’ Attention Network Test. The book uses these as organising distinctions, not as claims that each task or neural network is independent. Harold Pashler, Klaus Oberauer and Michael Kane with Randall Engle support the treatment of selective control, capacity limits and working memory as related but contested constructs.

Selection and inattentional blindness. The basketball and gorilla demonstration is Daniel Simons and Christopher Chabris’s 1999 experiment, developed within the broader study of inattentional blindness associated with Arien Mack and Irvin Rock. Detection varied with the attended team, style of counting and visibility of the unexpected event. For that reason the body avoids the familiar claim that exactly half of people miss the gorilla. The supported conclusion is conditional: demanding selection can prevent a conspicuous, fully visible event from reaching reportable awareness.

No single attention span. Experimental psychology measures different operations through different tasks and timescales. Vigilance, rapid target detection, reading comprehension and voluntary engagement cannot be collapsed into one duration without defining a common event and criterion. The treatment here follows the field’s task-specific approach rather than accepting a species-wide stopwatch.

Goals, salience and learning. Edward Awh, Artem Belopolsky and Jan Theeuwes argued that the old top-down versus bottom-up split omits selection history. Brian Anderson, Patryk Laurent and Steven Yantis showed value-driven capture by first associating a colour with reward, then finding that it distracted attention after becoming irrelevant. The phone application is an inference from this learning mechanism and habit research, not a claim that every notification reproduces the same visual-search effect.

Expertise and compression. The chess example comes from William Chase and Herbert Simon’s studies of expert perception and meaningful chunks. Expertise does not remove capacity limits. It changes the representation of familiar material, letting organised structures be handled with less deliberate assembly than disconnected elements.

Working memory. Klaus Oberauer reviews the conceptual relations between working memory and attention and cautions against equating them. Kane and Engle support the role of executive attention in maintaining goal-relevant information under interference. Nilli Lavie and colleagues distinguish perceptual load from load on working memory and cognitive control; the body uses only the latter claim when explaining increased distractor interference. The narrow-desk description is a teaching model. The book deliberately avoids a fixed item count because estimates depend on task, grouping, knowledge and what counts as an item.

Sustained attention. Michael Esterman and David Rothlein review competing models involving arousal, motivation, control and mind wandering. David R. Thomson, Derek Besner and Daniel Smilek provide the resource-control account, used here as one explanation rather than consensus. The claim that performance fluctuates rather than falling at one uniform rate also follows sustained-attention work reviewed by Esterman and Rothlein.

Mind wandering. Jonathan Smallwood and Jonathan Schooler review the costs and possible functions of task-unrelated thought. The body distinguishes a thought’s existence from its fit with the present task. Planning and incubation are plausible benefits in some settings; poorer comprehension and vigilance are common costs when thought conflicts with the task.

Breaks. Atsunori Ariga and Alejandro Lleras found that brief, rare disengagement from a repetitive goal prevented the usual decline in one vigilance paradigm. This supports the possibility that a well-placed break can help. It does not validate a universal twenty-five, fifty-two or ninety-minute interval, so no fixed focus cycle is recommended.

Sleep and state. Julian Lim and David Dinges’s meta-analysis supports the broad claim that short-term sleep deprivation impairs several cognitive functions, including vigilance. The book mentions pain, anxiety, low mood, medication and illness as possible influences on attention without offering diagnosis or treatment. Those factors have many pathways and belong primarily to medical and mental-health guidance.

Task switching. Joshua Rubinstein, David Meyer and Jeffrey Evans, and Stephen Monsell, support slower or less accurate performance around switches, task-set reconfiguration and persisting interference. The qualification that some activities can overlap is essential: dual-task costs depend on shared operations, practice, timing and task structure. The text rejects the slogan that humans can never perform two actions at once.

Interruption and resumption. Erik Altmann and Gregory Trafton model goals as varying in activation during interruption. Gregory Trafton and colleagues found that preparation before an interruption can aid later resumption. Sophie Leroy’s experiments support lingering competition from unfinished prior work. The book treats “attention residue” as a setting-specific goal-interference account, not a universal substance or fixed recovery interval.

The twenty-three-minute claim. No universal resumption time is used. Widely repeated figures usually derive from particular workplace observations with particular definitions of return. A complex reasoning task, routine checklist and natural boundary cannot share one defensible penalty. The manuscript therefore describes cost through lost task state, errors and reconstruction rather than a headline duration.

Notifications and acute disruption. Cary Stothart, Ainsley Mitchum and Courtney Yehnert found that call and text notifications disrupted performance on an attention-demanding laboratory task even when participants did not interact with the phone. Kostadin Kushlev, Jason Proulx and Elizabeth Dunn later varied both notification settings and phone proximity in daily life. Participants reported more inattention and hyperactivity symptoms when alerts were on and phones were within reach or sight than when alerts were off and phones were put away. The design therefore does not isolate notification settings from proximity. The latter outcomes were symptoms, not diagnoses. Hippolyte Fournier and colleagues’ 2026 experiment randomly assigned 180 University of Lyon students, including 140 women and seven agender participants, with a mean age of 20.63. All owned iPhones with Screen Time enabled, kept notifications active and regularly used at least four named social-media applications. Smartphone-style notifications during a Stroop task produced an approximately seven-second transient slowing. This is acute laboratory disruption in a selected student sample, not a universal notification cost, resumption time or long-term attention effect. It is the most setting-specific evidence retained in the body.

Mere presence. Douglas Parry’s 2024 review reports six meta-analyses covering fifty-six study samples and 7,093 participants. Only working-memory capacity showed a statistically significant negative pooled effect. Andree Hartanto and colleagues’ 2024 four-level meta-analysis combined 166 effect sizes from thirty-three studies and fifty-three samples, finding no significant overall cognitive effect (d = -0.02). These are different aggregations, not interchangeable estimates: an overall null can coexist with a small domain-specific result. Heterogeneous methods and low power constrain interpretation. The manuscript treats distance as a personal experiment rather than a universal consequence of proximity.

Measuring phone use. Douglas Parry, Brittany Davidson, Craig Sewall, Jacob Fisher, Hannah Mieczkowski and Daniel Quintana synthesised 106 effect sizes comparing self-reported media use with logged measures. Correlations were moderate and reports were rarely accurate in absolute terms. Logs solve part of the duration problem but do not make map use, work, reading and passive feeds equivalent exposures.

Blocking mobile internet. Noah Castelo and colleagues ran a pre-registered randomised waitlist cross-over trial published on 18 February 2025. Participants retained calls, texts and internet on non-phone devices while an application blocked Wi-Fi and mobile data on their iPhones for two weeks. Of 467 who agreed, 266 installed the application and 119 met the pre-registered adherence threshold of at least ten blocked days. The authors’ pooled intention-to-treat change in objective sustained attention was modest, dz = 0.23. Their analyses used available outcome data, not complete follow-up from all 467 volunteers; “intention-to-treat” here means adherence was not required for inclusion. Eighty-three per cent reported high motivation to reduce use, and none of the tested time-use, social-connection, self-control or sleep variables mediated the sustained-attention change. Attrition, low adherence, expectancy effects, the waitlist control and a relatively motivated North American sample constrain transportability. The authors’ comparisons with antidepressant effects and ten years of age-related change are omitted because the populations, outcomes and designs are not measurement-compatible.

Notification batching. Nicholas Fitz and colleagues compared notification schedules in a field experiment and found benefits from batching in some well-being outcomes. The result supports scheduled delivery as a reasonable experiment, not a rule that fits emergency work, caregiving or every social system.

Illustrative sequence and anecdote provenance. The interrupted-writing sequence in the operating section and its redesigned version are labelled illustrative. They combine established mechanisms of task state, alerting, switching, prospective memory, resumption and learned checking. They do not report an observed individual. No invented dialogue or private motive is presented as fact, and example durations are not measurements. Other unnamed examples are functional illustrations unless a study or named source is specified.

Current phone controls. Apple’s current iPhone guidance states that Focus modes can allow or silence selected people and applications and can be scheduled by time, location or application. Google’s current Android guidance describes modes that can allow selected people and applications, plus scheduled modes and a Digital Wellbeing Focus mode that can pause chosen applications and their notifications. Google warns that settings vary by phone. The body therefore describes capabilities shared by major current systems without promising identical menus or behaviour on every device. Guidance was checked on 5 September 2026.

Dopamine and wanting. Wolfram Schultz reviews dopamine reward-prediction-error coding. Kent Berridge and Terry Robinson distinguish hedonic liking from incentive wanting in addiction theory. Together they rule out the popular equation dopamine equals pleasure. Applying these mechanisms to everyday phone checking requires caution: direct evidence differs by behaviour, and social duty, anxiety, information seeking and habit can produce similar checking without one shared cause.

Habit and self-control. Wendy Wood and Dennis Rünger support context-response learning through repetition. Angela Duckworth, Tamar Gendler and James Gross distinguish situational self-control strategies from later response control. Duckworth, Katherine Milkman and David Laibson review strategies across the self-control process. Peter Gollwitzer and Paschal Sheeran support implementation intentions as a class, though effects vary by setting and outcome. These sources justify cue control, friction and if-then planning without claiming that willpower is useless.

Addiction language. The World Health Organization recognises gaming disorder in ICD-11 when impaired control, increasing priority and persistence despite harm produce marked functional impairment. It does not follow that frequent smartphone use is itself a recognised disorder. The manuscript uses this boundary to resist both casual diagnosis and denial of serious digital-behaviour problems.

The goldfish story. Microsoft Canada’s 2015 report, Attention Spans: Consumer Insights, contains the widely circulated graphic listing a human span of eight seconds and a goldfish span of nine. The graphic cites “Statistic Brain” and supplies no compatible tasks, populations or methods. Elsewhere the report says attention cannot be characterised by one duration and distinguishes sustained, selective and alternating attention. The book’s correction follows the missing measurement definition rather than using the graphic as evidence of decline.

Clinical boundary. NICE guideline NG87 covers recognition, specialist diagnosis and management of ADHD in children, young people and adults and reserves diagnosis for appropriately trained specialists. The practical section advises qualified assessment only when problems are broad, persistent or damaging despite sensible support. It does not provide a symptom checklist, suggest a diagnosis or recommend treatment. Current NICE and World Health Organization guidance was checked on 5 September 2026.

Go Deeper editions. Adam Gazzaley and Larry Rosen’s The Distracted Mind was first published by MIT Press in 2016. Harold Pashler’s The Psychology of Attention was published by MIT Press in 1998. James Williams’s Stand Out of Our Light was published by Cambridge University Press in 2018. William James’s The Principles of Psychology appeared in two volumes from Henry Holt in 1890; the attention chapter is Chapter XI of Volume I.

Bibliography

Original research, reports and guidance

Anderson, Brian A., Patryk A. Laurent and Steven Yantis. “Value-driven Attentional Capture.” Proceedings of the National Academy of Sciences 108, no. 25 (2011): 10367-10371.

Ariga, Atsunori, and Alejandro Lleras. “Brief and Rare Mental ‘Breaks’ Keep You Focused: Deactivation and Reactivation of Task Goals Preempt Vigilance Decrements.” Cognition 118, no. 3 (2011): 439-443.

Castelo, Noah, Kostadin Kushlev, Adrian F. Ward, Michael Esterman and Peter B. Reiner. “Blocking Mobile Internet on Smartphones Improves Sustained Attention, Mental Health, and Subjective Well-being.” PNAS Nexus 4, no. 2 (2025): pgaf017. DOI 10.1093/pnasnexus/pgaf017.

Chase, William G., and Herbert A. Simon. “Perception in Chess.” Cognitive Psychology 4, no. 1 (1973): 55-81.

Fan, Jin, Bruce D. McCandliss, Tobias Sommer, Amir Raz and Michael I. Posner. “Testing the Efficiency and Independence of Attentional Networks.” Journal of Cognitive Neuroscience 14, no. 3 (2002): 340-347.

Fitz, Nicholas, Kostadin Kushlev, Ranjan Jagannathan, Terrel Lewis, Devang Paliwal and Dan Ariely. “Batching Smartphone Notifications Can Improve Well-being.” Computers in Human Behavior 101 (2019): 84-94.

Fournier, Hippolyte, Arnaud Fournel, François Osiurak, Olivier Koenig, Flora Pâris, Vivien Gaujoux and Fabien Ringeval. “Attention Hijacked: How Social Media Notifications Disrupt Cognitive Processing.” Computers in Human Behavior 179 (2026): 108926. DOI 10.1016/j.chb.2026.108926.

Kushlev, Kostadin, Jason Proulx and Elizabeth W. Dunn. “‘Silence Your Phones’: Smartphone Notifications Increase Inattention and Hyperactivity Symptoms.” In Proceedings of the 2016 CHI Conference on Human Factors in Computing Systems, 1011-1020. New York: Association for Computing Machinery, 2016.

Lavie, Nilli, Aleksandra Hirst, Jan W. de Fockert and Essi Viding. “Load Theory of Selective Attention and Cognitive Control.” Journal of Experimental Psychology: General 133, no. 3 (2004): 339-354.

Leroy, Sophie. “Why Is It So Hard to Do My Work? The Challenge of Attention Residue When Switching between Work Tasks.” Organizational Behavior and Human Decision Processes 109, no. 2 (2009): 168-181.

Microsoft Canada. Attention Spans: Consumer Insights. Spring 2015.

Parry, Douglas A. “Does the Mere Presence of a Smartphone Impact Cognitive Performance? A Meta-analysis of the ‘Brain Drain’ Effect.” Media Psychology 27, no. 5 (2024): 737-762. DOI 10.1080/15213269.2023.2286647.

Parry, Douglas A., Brittany I. Davidson, Craig J. R. Sewall, Jacob T. Fisher, Hannah Mieczkowski and Daniel S. Quintana. “A Systematic Review and Meta-analysis of Discrepancies between Logged and Self-reported Digital Media Use.” Nature Human Behaviour 5, no. 11 (2021): 1535-1547.

Rubinstein, Joshua S., David E. Meyer and Jeffrey E. Evans. “Executive Control of Cognitive Processes in Task Switching.” Journal of Experimental Psychology: Human Perception and Performance 27, no. 4 (2001): 763-797.

Simons, Daniel J., and Christopher F. Chabris. “Gorillas in Our Midst: Sustained Inattentional Blindness for Dynamic Events.” Perception 28, no. 9 (1999): 1059-1074.

Stothart, Cary, Ainsley Mitchum and Courtney Yehnert. “The Attentional Cost of Receiving a Cell Phone Notification.” Journal of Experimental Psychology: Human Perception and Performance 41, no. 4 (2015): 893-897.

Trafton, J. Gregory, Erik M. Altmann, Derek P. Brock and Farilee E. Mintz. “Preparing to Resume an Interrupted Task: Effects of Prospective Goal Encoding and Retrospective Rehearsal.” International Journal of Human-Computer Studies 58, no. 5 (2003): 583-603.

World Health Organization. Clinical Descriptions and Diagnostic Requirements for ICD-11 Mental, Behavioural and Neurodevelopmental Disorders. Geneva: World Health Organization, 2024.

Reviews, syntheses and major interpretations

Altmann, Erik M., and J. Gregory Trafton. “Memory for Goals: An Activation-based Model.” Cognitive Science 26, no. 1 (2002): 39-83.

Awh, Edward, Artem V. Belopolsky and Jan Theeuwes. “Top-down versus Bottom-up Attentional Control: A Failed Theoretical Dichotomy.” Trends in Cognitive Sciences 16, no. 8 (2012): 437-443.

Berridge, Kent C., and Terry E. Robinson. “Liking, Wanting, and the Incentive-sensitization Theory of Addiction.” American Psychologist 71, no. 8 (2016): 670-679.

Duckworth, Angela L., Tamar Szabó Gendler and James J. Gross. “Situational Strategies for Self-control.” Perspectives on Psychological Science 11, no. 1 (2016): 35-55.

Duckworth, Angela L., Katherine L. Milkman and David Laibson. “Beyond Willpower: Strategies for Reducing Failures of Self-control.” Psychological Science in the Public Interest 19, no. 3 (2018): 102-129.

Esterman, Michael, and David Rothlein. “Models of Sustained Attention.” Current Opinion in Psychology 29 (2019): 174-180.

Gollwitzer, Peter M., and Paschal Sheeran. “Implementation Intentions and Goal Achievement: A Meta-analysis of Effects and Processes.” Advances in Experimental Social Psychology 38 (2006): 69-119.

Hartanto, Andree, Verity Y. Q. Lua, K. T. A. Sandeeshwara Kasturiratna, Paye Shin Koh, Germaine Y. Q. Tng, Manmeet Kaur, Frosch Y. X. Quek, Jonathan L. Chia and Nadyanna M. Majeed. “The Effect of Mere Presence of Smartphone on Cognitive Functions: A Four-level Meta-analysis.” Technology, Mind, and Behavior 5, no. 1 (2024): 1-14. DOI 10.1037/tmb0000123.

Kane, Michael J., and Randall W. Engle. “The Role of Prefrontal Cortex in Working-memory Capacity, Executive Attention, and General Fluid Intelligence.” Psychonomic Bulletin & Review 9, no. 4 (2002): 637-671.

Lim, Julian, and David F. Dinges. “A Meta-analysis of the Impact of Short-term Sleep Deprivation on Cognitive Variables.” Psychological Bulletin 136, no. 3 (2010): 375-389.

Mack, Arien, and Irvin Rock. Inattentional Blindness. Cambridge, MA: MIT Press, 1998.

Monsell, Stephen. “Task Switching.” Trends in Cognitive Sciences 7, no. 3 (2003): 134-140.

Oberauer, Klaus. “Working Memory and Attention: A Conceptual Analysis and Review.” Journal of Cognition 2, no. 1 (2019): 36.

Petersen, Steven E., and Michael I. Posner. “The Attention System of the Human Brain: 20 Years After.” Annual Review of Neuroscience 35 (2012): 73-89.

Posner, Michael I., and Steven E. Petersen. “The Attention System of the Human Brain.” Annual Review of Neuroscience 13 (1990): 25-42.

Schultz, Wolfram. “Dopamine Reward Prediction Error Coding.” Dialogues in Clinical Neuroscience 18, no. 1 (2016): 23-32.

Smallwood, Jonathan, and Jonathan W. Schooler. “The Science of Mind Wandering: Empirically Navigating the Stream of Consciousness.” Annual Review of Psychology 66 (2015): 487-518.

Thomson, David R., Derek Besner and Daniel Smilek. “A Resource-control Account of Sustained Attention: Evidence from Mind-wandering and Vigilance Paradigms.” Perspectives on Psychological Science 10, no. 1 (2015): 82-96.

Wood, Wendy, and Dennis Rünger. “Psychology of Habit.” Annual Review of Psychology 67 (2016): 289-314.

Further reading used in Go Deeper

Gazzaley, Adam, and Larry D. Rosen. The Distracted Mind: Ancient Brains in a High-Tech World. Cambridge, MA: MIT Press, 2016.

James, William. The Principles of Psychology. Vol. 1. New York: Henry Holt, 1890.

Pashler, Harold E. The Psychology of Attention. Cambridge, MA: MIT Press, 1998.

Williams, James. Stand Out of Our Light: Freedom and Resistance in the Attention Economy. Cambridge: Cambridge University Press, 2018.

Current platform and clinical guidance

Apple. “Allow or Silence Notifications for a Focus on iPhone.” Apple Support. Accessed 5 September 2026.

Apple. “Turn On or Schedule a Focus on iPhone.” Apple Support. Accessed 5 September 2026.

Google. “Limit Interruptions with Modes and Do Not Disturb on Android.” Android Help. Accessed 5 September 2026.

Google. “Manage How You Spend Time on Your Android Phone with Digital Wellbeing.” Android Help. Accessed 5 September 2026.

National Institute for Health and Care Excellence. Attention Deficit Hyperactivity Disorder: Diagnosis and Management. NICE Guideline NG87. London: NICE, 2018, updated 2019. Accessed 5 September 2026.

World Health Organization. “Addictive Behaviours: Gaming Disorder.” Questions and Answers. 22 October 2020. Accessed 5 September 2026.