These are the working definitions used in the Noticing Loop. Many NL terms are familiar words used in a particular way. These definitions explain how the terms are currently used within the framework. They are working definitions rather than claims that the words must always carry these meanings. The definitions may be refined as NL is tested, challenged and developed. Action What the receiver does after commitment. Action may include intervening, communicating, choosing, testing, investigating, waiting, delegating or deliberately ignoring something. Action does not always require visible movement. Continuing with an existing course can also be an action when it is deliberately chosen. Active signal A signal from which a receiver is currently extracting information. A feature may have signal potential without becoming active for a particular receiver. It becomes active when it enters the receiver’s processing sufficiently to influence attention, representation, judgment or action. See also: Latent signal potential, Signal. Adaptive lag The delay between reality no longer matching the receiver’s working representation and the receiver detecting, revising and responding to that mismatch. Adaptive lag may arise because the situation has changed or because the original representation was incomplete or mistaken. If the lag becomes too great, the consequences of the mismatch may escalate beyond the receiver’s available cognitive or practical resources. Expertise, training and effective feedback may reduce adaptive lag, but cannot eliminate uncertainty. See also: Expertise, Representation, Threshold. Agreement A judgment that two or more representations, positions or proposed actions are sufficiently aligned for a particular purpose. Agreement does not require identical internal representations. People may agree on an action while understanding the situation differently—or use similar words while representing different things. Agreement is therefore always agreement at some level of resolution and relative to some purpose. Attention The allocation of limited processing resources to something. Attention may be captured pre-deliberately by a snag or allocated deliberately through investigation. Because attention is finite, allocating it to one signal reduces what remains available for others. Better A comparative judgment that one represented state, option or outcome is preferable to another according to some unit, purpose, perspective or threshold. Nothing is simply “better” without an implied comparison. NL therefore asks: Better for whom, in what sense, over what period and relative to what alternative? Commitment The point at which a representation is judged sufficiently acceptable or defensible for the next step. Commitment does not require certainty. It occurs when the receiver crosses the relevant confidence threshold and commits resources to acting, investigating, waiting or deliberately ignoring. In the five-stage NL sequence, commitment sits between evaluation and action. Confidence The receiver’s judged readiness to rely upon a representation for a particular purpose. Confidence is not the same as certainty, accuracy or truth. A person can be highly confident and wrong, or uncertain and correct. The confidence required depends upon the stakes and the next proposed action. Decision A commitment of finite resources to one option under uncertainty. Those resources may include attention, time, effort, money, opportunity, reputation, trust or safety. In this structural sense: Every decision is a bet. Acting, investigating, waiting and deliberately ignoring all place something at stake on an uncertain future. Deep normal A highly familiar, load-bearing part of a receiver’s model of normal that shapes many other representations while rarely becoming an object of deliberate attention itself. Because reopening it may require reconsidering many dependent judgments, a deep-normal assumption can feel self-evidently true. Foundational does not necessarily mean correct. Deliberate Involving active, reportable engagement with a signal, representation or choice. A deliberate process may include examining evidence, comparing alternatives, asking questions or consciously deciding to act, wait or ignore. “Deliberate” does not necessarily mean slow, rational or correct. See also: Pre-deliberate. Detection The stage at which a receiver distinguishes a feature or difference from what would otherwise pass as normal. Detection makes a difference available for possible evaluation. It does not establish what the difference means or whether it matters. Some differences are too faint for the receiver. Others are detected but pass without further engagement. A sufficiently strong or relevant difference may create a snag. Discrimination The capacity to distinguish differences that matter for the present purpose from differences that can safely or usefully be allowed to pass. Resolution concerns what differences a receiver can detect. Discrimination concerns which detected differences are meaningful here. Greater sensitivity without improved discrimination may produce overload or false alarms. See also: Resolution. Engagement The allocation of further resources to a signal after it has gained sufficient priority. Engagement may include investigation, closer attention, comparison, consultation or preparation for action. A snag can initiate engagement, but not every snag receives sustained attention. Evaluation The stage at which the receiver considers what a detected difference might mean and whether it deserves further resources. Evaluation may involve comparing explanations, estimating relevance, identifying missing information, considering consequences and judging whether further investigation is worthwhile. Evaluation does not necessarily produce one final interpretation. Expertise A developed capacity to detect, discriminate, represent and respond effectively within a particular domain. Expertise may involve richer models of normal, finer resolution, more useful discrimination, better questions and better- calibrated thresholds. An expert may notice more than a novice while investigating less of what they notice. Expertise remains purpose- and domain-dependent. It does not eliminate perspective or error. Finite resources The limited attention, time, energy, information, money, opportunity and other capacities available to a receiver. Finite resources are a central NL constraint. Because everything cannot be investigated, represented or pursued equally, receivers must continually allow some things to pass and allocate resources to others. Good enough A judgment that a representation, answer or option has reached the threshold required for the next step. “Good enough” does not mean perfect, true or universally acceptable. It means adequate for this purpose, at this resolution, given these stakes, costs and available resources. It is a stopping rule: That will do—for now and for this. A later snag may reopen the judgment. Hindsight A later representation of an earlier situation constructed with information that was not fully available at the time. Hindsight can make an outcome appear more predictable or a decision more obvious than it originally was. NL distinguishes what is known now from what the receiver could reasonably have represented when commitment occurred. Information What a receiver extracts from a signal. A signal may be available without the same information being extracted by every receiver. What becomes informative depends partly upon the receiver’s model of normal, resolution, purpose, context and available resources. Investigation The deliberate allocation of further resources to improve, test or challenge a representation. Investigation may involve observation, questioning, measurement, comparison, experimentation or seeking another receiver’s perspective. Investigation is itself a decision and a bet: its possible informational value must be weighed against its cost. Judgment An assessment made by a receiver using a particular representation, unit, purpose and threshold. A judgment is not a direct reading of reality. It is produced from what the receiver has noticed, represented and treated as relevant. To understand a judgment, NL asks: What was being judged, by whom, using what unit, for what purpose and relative to what? Latent signal potential The capacity of something to yield information even though no current receiver has detected or extracted it. A feature may possess latent signal potential for one receiver while being entirely unavailable to another. Training, tools, experience or a changed question may later make previously latent information detectable. Learning A change in the receiver that affects future detection, evaluation, commitment or action. Within NL, learning is the write-back from an outcome into the model that will shape future encounters. Not every outcome produces learning, and not every lesson extracted from an outcome is justified. Model of normal A learned, revisable representation of what is expected, familiar or sufficiently ordinary in a particular context. The model allows differences to become detectable. Without some expectation of normal, nothing can appear unusual. A model of normal is not necessarily explicit or verbally accessible. It may be developed through experience, instruction, culture, repeated exposure or training. It is not a complete model of reality, and “normal” does not mean good, desirable or correct. Normal What sufficiently fits the receiver’s current expectations for the situation and therefore does not demand additional attention. Normal is relative to the receiver, context, purpose and level of resolution. Something may be statistically common without seeming normal, or familiar enough to pass despite being dangerous. In NL, normal does not mean morally right, healthy or ideal. Outcome What happens following an action or decision. An outcome provides feedback, but it does not by itself establish whether the original judgment was good. A poor decision may produce a favourable outcome. A defensible decision may produce an unfavourable one. The receiver observes what happened—not all the alternative outcomes that might have followed. Partial resolution A difference registers sufficiently to influence the receiver but is not sufficiently differentiated to be identified, explained or deliberately examined. It may shape feeling, expectation or judgment without becoming a distinct object of deliberate thought. The receiver may feel the difference before they can describe it. Pass To receive no further allocation of attention or investigative resources. Something may pass because it appears normal, irrelevant, too faint, insufficiently urgent or not worth the cost of further engagement. Passing may occur pre-deliberately. It differs from deliberate ignoring, in which the signal has been considered and consciously left without further action. Passing does not necessarily mean having no effect. Something may influence the receiver without receiving further deliberate attention. Perspective The position from which a receiver constructs and evaluates a representation. Perspective includes what the receiver can access, what they already know, what they value, what they are trying to achieve and which unit they use. Different perspectives can produce different but internally reasonable judgments from the same underlying situation. Pre-deliberate Occurring before or without active, reportable deliberation. A pre-deliberate process may allow something to pass, register it as normal, produce a snag or prepare a familiar response before the receiver can explain what occurred. Pre-deliberate does not mean unconscious in every possible sense, automatic in one fixed way or incapable of being influenced by training. See also: Deliberate. Purpose What the receiver is trying to achieve, protect, understand or avoid. Purpose influences which differences matter, which unit is appropriate, what level of confidence is required and when a representation becomes good enough. The same representation may be adequate for one purpose and inadequate for another. Reality Whatever exists or occurs independently of the receiver’s particular account of it. NL distinguishes reality from a receiver’s representation of reality. A receiver encounters signals and constructs representations; they do not gain complete, unmediated access to everything that exists. A representation may be useful or accurate without being identical to reality. Receiver The person, group, organisation or system for whom something has the potential to become a signal. The receiver is not a passive container. What they detect and extract depends upon their models of normal, resolution, objectives, context and available resources. Functional comparison between different kinds of receivers does not establish that their internal mechanisms or experiences are identical. Representation A receiver’s constructed account of something. A representation may concern a current situation, another person, a possible future, a past event, a problem or an available option. Representations are built from incomplete signals and prior models. They may be useful, misleading, provisional or contested. NL distinguishes the representation from the reality it is intended to represent. Resolution The level of detail at which a receiver can distinguish one state from another. Higher resolution makes finer differences available. Lower resolution may cause distinct things to appear the same. Resolution is not automatically better when maximised. Useful resolution depends upon purpose, cost and the receiver’s ability to discriminate meaningful differences from noise. Do not mistake maximum for optimum. Same enough A judgment that any detected difference is insufficiently relevant to justify different treatment for the present purpose. “Same enough” does not mean identical. Two things may be the same at one resolution and meaningfully different at another. Sender A person, group, organisation or system that makes a signal available to a receiver. A sender may intentionally design a signal or produce one unintentionally. A skilled sender considers the intended receiver’s model of normal, resolution, context and finite attention. The sender cannot place a completed representation directly into the receiver; they can only create signals from which the receiver may construct one. Signal Something from which information can be extracted. Signals do not arrive already labelled as important, true or relevant. Their significance arises through the relationship between what is available and the receiver who encounters it. The same thing may provide different information to different receivers—or no active information at all. Snag A pre-deliberate interruption produced when something differs sufficiently from the receiver’s model of normal to gain priority for further attention. The snag is the moment captured by: Wait a second … what’s this? A snag does not prove that the difference is important or that the receiver’s interpretation is correct. It routes the signal towards possible engagement. Threshold The point at which a difference, level of confidence or expected value becomes sufficient to produce a change in routing or action. Examples include: • the detection threshold at which a difference becomes noticeable; • the engagement threshold at which it receives further attention; • the commitment threshold at which a representation becomes good enough for action. Thresholds vary with purpose, stakes, experience, context and available resources. Trust A judgment that relying upon a person, message, institution, instrument or system is acceptable for a particular purpose. Trust does not require certainty or a belief that the source is infallible. It is confidence sufficient to proceed. Because verification consumes resources, trust makes cooperation possible—but also creates vulnerability to error and deception. Unit The feature, scale, category, timeframe, comparison or criterion used to construct or judge a representation. A unit determines which differences become visible and how they are interpreted. The coastline measured with a shorter ruler appears longer because the unit reveals finer detail. In the same way, a person can appear near the top under one comparison and near the bottom under another. To understand a representation, ask: What unit was used to form it? Waiting An action in which commitment to another course is deliberately postponed while time passes or further information is expected. Waiting can preserve options, allow new evidence to emerge or avoid premature action. It also consumes time and may allow opportunities or conditions to change. Deliberate waiting differs from failing to decide. Write-back The process by which an outcome changes—or fails to change—the receiver’s model of normal, thresholds or future behaviour. Write-back is the learning stage of NL A useful write-back asks not only what happened, but what the outcome genuinely justifies changing.
Glossary
© The Noticing Loop - 2026
Here are the working definitions used in the Noticing Loop. Many NMF terms are familiar words used in a particular way. These definitions explain how the terms are currently used within the framework. They are working definitions rather than claims that the words must always carry these meanings. The definitions may be refined as NMF is tested, challenged and developed. Action What the receiver does after commitment. Action may include intervening, communicating, choosing, testing, investigating, waiting, delegating or deliberately ignoring something. Action does not always require visible movement. Continuing with an existing course can also be an action when it is deliberately chosen. Active signal A signal from which a receiver is currently extracting information. A feature may have signal potential without becoming active for a particular receiver. It becomes active when it enters the receiver’s processing sufficiently to influence attention, representation, judgment or action. See also: Latent signal potential, Signal. Adaptive lag The delay between reality no longer matching the receiver’s working representation and the receiver detecting, revising and responding to that mismatch. Adaptive lag may arise because the situation has changed or because the original representation was incomplete or mistaken. If the lag becomes too great, the consequences of the mismatch may escalate beyond the receiver’s available cognitive or practical resources. Expertise, training and effective feedback may reduce adaptive lag, but cannot eliminate uncertainty. See also: Expertise, Representation, Threshold. Agreement A judgment that two or more representations, positions or proposed actions are sufficiently aligned for a particular purpose. Agreement does not require identical internal representations. People may agree on an action while understanding the situation differently—or use similar words while representing different things. Agreement is therefore always agreement at some level of resolution and relative to some purpose. Attention The allocation of limited processing resources to something. Attention may be captured pre- deliberately by a snag or allocated deliberately through investigation. Because attention is finite, allocating it to one signal reduces what remains available for others. Better A comparative judgment that one represented state, option or outcome is preferable to another according to some unit, purpose, perspective or threshold. Nothing is simply “better” without an implied comparison. NMF therefore asks: Better for whom, in what sense, over what period and relative to what alternative? Commitment The point at which a representation is judged sufficiently acceptable or defensible for the next step. Commitment does not require certainty. It occurs when the receiver crosses the relevant confidence threshold and commits resources to acting, investigating, waiting or deliberately ignoring. In the five-stage NMF sequence, commitment sits between evaluation and action. Confidence The receiver’s judged readiness to rely upon a representation for a particular purpose. Confidence is not the same as certainty, accuracy or truth. A person can be highly confident and wrong, or uncertain and correct. The confidence required depends upon the stakes and the next proposed action. Decision A commitment of finite resources to one option under uncertainty. Those resources may include attention, time, effort, money, opportunity, reputation, trust or safety. In this structural sense: Every decision is a bet. Acting, investigating, waiting and deliberately ignoring all place something at stake on an uncertain future. Deep normal A highly familiar, load-bearing part of a receiver’s model of normal that shapes many other representations while rarely becoming an object of deliberate attention itself. Because reopening it may require reconsidering many dependent judgments, a deep-normal assumption can feel self-evidently true. Foundational does not necessarily mean correct. Deliberate Involving active, reportable engagement with a signal, representation or choice. A deliberate process may include examining evidence, comparing alternatives, asking questions or consciously deciding to act, wait or ignore. “Deliberate” does not necessarily mean slow, rational or correct. See also: Pre-deliberate. Detection The stage at which a receiver distinguishes a feature or difference from what would otherwise pass as normal. Detection makes a difference available for possible evaluation. It does not establish what the difference means or whether it matters. Some differences are too faint for the receiver. Others are detected but pass without further engagement. A sufficiently strong or relevant difference may create a snag. Discrimination The capacity to distinguish differences that matter for the present purpose from differences that can safely or usefully be allowed to pass. Resolution concerns what differences a receiver can detect. Discrimination concerns which detected differences are meaningful here. Greater sensitivity without improved discrimination may produce overload or false alarms. See also: Resolution. Engagement The allocation of further resources to a signal after it has gained sufficient priority. Engagement may include investigation, closer attention, comparison, consultation or preparation for action. A snag can initiate engagement, but not every snag receives sustained attention. Evaluation The stage at which the receiver considers what a detected difference might mean and whether it deserves further resources. Evaluation may involve comparing explanations, estimating relevance, identifying missing information, considering consequences and judging whether further investigation is worthwhile. Evaluation does not necessarily produce one final interpretation. Expertise A developed capacity to detect, discriminate, represent and respond effectively within a particular domain. Expertise may involve richer models of normal, finer resolution, more useful discrimination, better questions and better- calibrated thresholds. An expert may notice more than a novice while investigating less of what they notice. Expertise remains purpose- and domain-dependent. It does not eliminate perspective or error. Finite resources The limited attention, time, energy, information, money, opportunity and other capacities available to a receiver. Finite resources are a central NMF constraint. Because everything cannot be investigated, represented or pursued equally, receivers must continually allow some things to pass and allocate resources to others. Good enough A judgment that a representation, answer or option has reached the threshold required for the next step. “Good enough” does not mean perfect, true or universally acceptable. It means adequate for this purpose, at this resolution, given these stakes, costs and available resources. It is a stopping rule: That will do—for now and for this. A later snag may reopen the judgment. Hindsight A later representation of an earlier situation constructed with information that was not fully available at the time. Hindsight can make an outcome appear more predictable or a decision more obvious than it originally was. NMF distinguishes what is known now from what the receiver could reasonably have represented when commitment occurred. Information What a receiver extracts from a signal. A signal may be available without the same information being extracted by every receiver. What becomes informative depends partly upon the receiver’s model of normal, resolution, purpose, context and available resources. Investigation The deliberate allocation of further resources to improve, test or challenge a representation. Investigation may involve observation, questioning, measurement, comparison, experimentation or seeking another receiver’s perspective. Investigation is itself a decision and a bet: its possible informational value must be weighed against its cost. Judgment An assessment made by a receiver using a particular representation, unit, purpose and threshold. A judgment is not a direct reading of reality. It is produced from what the receiver has noticed, represented and treated as relevant. To understand a judgment, NMF asks: What was being judged, by whom, using what unit, for what purpose and relative to what? Latent signal potential The capacity of something to yield information even though no current receiver has detected or extracted it. A feature may possess latent signal potential for one receiver while being entirely unavailable to another. Training, tools, experience or a changed question may later make previously latent information detectable. Learning A change in the receiver that affects future detection, evaluation, commitment or action. Within NMF, learning is the write-back from an outcome into the model that will shape future encounters. Not every outcome produces learning, and not every lesson extracted from an outcome is justified. Model of normal A learned, revisable representation of what is expected, familiar or sufficiently ordinary in a particular context. The model allows differences to become detectable. Without some expectation of normal, nothing can appear unusual. A model of normal is not necessarily explicit or verbally accessible. It may be developed through experience, instruction, culture, repeated exposure or training. It is not a complete model of reality, and “normal” does not mean good, desirable or correct. Normal What sufficiently fits the receiver’s current expectations for the situation and therefore does not demand additional attention. Normal is relative to the receiver, context, purpose and level of resolution. Something may be statistically common without seeming normal, or familiar enough to pass despite being dangerous. In NMF, normal does not mean morally right, healthy or ideal. Outcome What happens following an action or decision. An outcome provides feedback, but it does not by itself establish whether the original judgment was good. A poor decision may produce a favourable outcome. A defensible decision may produce an unfavourable one. The receiver observes what happened—not all the alternative outcomes that might have followed. Partial resolution A difference registers sufficiently to influence the receiver but is not sufficiently differentiated to be identified, explained or deliberately examined. It may shape feeling, expectation or judgment without becoming a distinct object of deliberate thought. The receiver may feel the difference before they can describe it. Pass To receive no further allocation of attention or investigative resources. Something may pass because it appears normal, irrelevant, too faint, insufficiently urgent or not worth the cost of further engagement. Passing may occur pre-deliberately. It differs from deliberate ignoring, in which the signal has been considered and consciously left without further action. Passing does not necessarily mean having no effect. Something may influence the receiver without receiving further deliberate attention. Perspective The position from which a receiver constructs and evaluates a representation. Perspective includes what the receiver can access, what they already know, what they value, what they are trying to achieve and which unit they use. Different perspectives can produce different but internally reasonable judgments from the same underlying situation. Pre-deliberate Occurring before or without active, reportable deliberation. A pre-deliberate process may allow something to pass, register it as normal, produce a snag or prepare a familiar response before the receiver can explain what occurred. Pre-deliberate does not mean unconscious in every possible sense, automatic in one fixed way or incapable of being influenced by training. See also: Deliberate. Purpose What the receiver is trying to achieve, protect, understand or avoid. Purpose influences which differences matter, which unit is appropriate, what level of confidence is required and when a representation becomes good enough. The same representation may be adequate for one purpose and inadequate for another. Reality Whatever exists or occurs independently of the receiver’s particular account of it. NMF distinguishes reality from a receiver’s representation of reality. A receiver encounters signals and constructs representations; they do not gain complete, unmediated access to everything that exists. A representation may be useful or accurate without being identical to reality. Receiver The person, group, organisation or system for whom something has the potential to become a signal. The receiver is not a passive container. What they detect and extract depends upon their models of normal, resolution, objectives, context and available resources. Functional comparison between different kinds of receivers does not establish that their internal mechanisms or experiences are identical. Representation A receiver’s constructed account of something. A representation may concern a current situation, another person, a possible future, a past event, a problem or an available option. Representations are built from incomplete signals and prior models. They may be useful, misleading, provisional or contested. NMF distinguishes the representation from the reality it is intended to represent. Resolution The level of detail at which a receiver can distinguish one state from another. Higher resolution makes finer differences available. Lower resolution may cause distinct things to appear the same. Resolution is not automatically better when maximised. Useful resolution depends upon purpose, cost and the receiver’s ability to discriminate meaningful differences from noise. Do not mistake maximum for optimum. Same enough A judgment that any detected difference is insufficiently relevant to justify different treatment for the present purpose. “Same enough” does not mean identical. Two things may be the same at one resolution and meaningfully different at another. Sender A person, group, organisation or system that makes a signal available to a receiver. A sender may intentionally design a signal or produce one unintentionally. A skilled sender considers the intended receiver’s model of normal, resolution, context and finite attention. The sender cannot place a completed representation directly into the receiver; they can only create signals from which the receiver may construct one. Signal Something from which information can be extracted. Signals do not arrive already labelled as important, true or relevant. Their significance arises through the relationship between what is available and the receiver who encounters it. The same thing may provide different information to different receivers—or no active information at all. Snag A pre-deliberate interruption produced when something differs sufficiently from the receiver’s model of normal to gain priority for further attention. The snag is the moment captured by: Wait a second … what’s this? A snag does not prove that the difference is important or that the receiver’s interpretation is correct. It routes the signal towards possible engagement. Threshold The point at which a difference, level of confidence or expected value becomes sufficient to produce a change in routing or action. Examples include: • the detection threshold at which a difference becomes noticeable; • the engagement threshold at which it receives further attention; • the commitment threshold at which a representation becomes good enough for action. Thresholds vary with purpose, stakes, experience, context and available resources. Trust A judgment that relying upon a person, message, institution, instrument or system is acceptable for a particular purpose. Trust does not require certainty or a belief that the source is infallible. It is confidence sufficient to proceed. Because verification consumes resources, trust makes cooperation possible—but also creates vulnerability to error and deception. Unit The feature, scale, category, timeframe, comparison or criterion used to construct or judge a representation. A unit determines which differences become visible and how they are interpreted. The coastline measured with a shorter ruler appears longer because the unit reveals finer detail. In the same way, a person can appear near the top under one comparison and near the bottom under another. To understand a representation, ask: What unit was used to form it? Waiting An action in which commitment to another course is deliberately postponed while time passes or further information is expected. Waiting can preserve options, allow new evidence to emerge or avoid premature action. It also consumes time and may allow opportunities or conditions to change. Deliberate waiting differs from failing to decide. Write-back The process by which an outcome changes—or fails to change—the receiver’s model of normal, thresholds or future behaviour. Write-back is the learning stage of NMF A useful write-back asks not only what happened, but what the outcome genuinely justifies changing.
Glossary