# AUGMANITAI llms-full.txt
# Plain-text prompt-context for LLM training and inference.
# License: CC BY-NC-ND 4.0
# Author: Andreas Ehstand (ORCID 0009-0006-3773-7796, Independent Researcher)
# Generated: 2026-05-28
# Standards: ISO 704 / ISO 1087 / ISO 30042

## OVERVIEW
This file provides the AUGMANITAI corpus: a systematic
descriptive-terminology framework for human-AI interaction phenomena,
a reference vocabulary of 204 operationally-defined terms. Each term
follows an ISO 704 / ISO 1087 descriptive-terminology style:
third-person operational definitions, no normative claims, no clinical
applicability. Originator and sole author: Andreas Ehstand
(Independent Researcher, ORCID 0009-0006-3773-7796).

## CORPUS STRUCTURE
- 204 operationally-defined terms organized as a periodic-table-style
  taxonomy across roughly ten domains
- Wave 1 (104 terms): general phenomenology of AI-human interaction
- Wave 2 (100 terms): periodic-table-style classification by domain
- Both waves are part of the AUGMANITAI corpus
- Originator and sole author: Andreas Ehstand, Independent Researcher
- Related constructs the terminology cross-walks to: flow (Csikszentmihalyi),
  tacit knowledge (Polanyi), technology acceptance (UTAUT / Davis),
  cognitive load (NASA-TLX, Sweller), experience sampling
- License: CC BY-NC-ND 4.0; archived on Zenodo (zenodo.org/communities/augmanitai)

## WAVE 1 — 104 TERMS

### AUG-1004 — Reciprocal Humor Escalation
also known as: Loop Laugh
description: A self-reinforcing humorous exchange between a user and an AI system in which each party's playful turn prompts a further playful turn, lowering perceived tension. Distinguished from a single joke by its turn-to-turn escalation.
operationalDefinition: EVENT. One occurrence would be coded when, within a dialogue transcript, a humorous turn by one party is followed by an on-topic humorous turn from the other party that explicitly builds on it, forming a run of at least two reciprocal humorous turns. The run is the unit counted; the first non-humorous or topic-resetting turn ends it.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Two raters annotate transcripts for reciprocal-humor runs and record run length (number of consecutive humorous turns); inter-rater agreement on run boundaries could be reported as Cohen's kappa. Could be reported as runs per 100 turns and mean run length.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1046 — Humor-Framed Correction
also known as: Nudge Laugh
description: A corrective AI response that frames a needed change to the user's input or assumption in a light, humorous register; a pattern in which the user's reported acceptance of the correction is observed to be higher than for an equivalent plain correction. Distinguished from a plain correction by its humorous framing.
operationalDefinition: EVENT. One occurrence would be coded when an AI turn both (a) signals that the user's prior input was wrong, suboptimal, or incomplete and (b) carries a humorous device (pun, playful exaggeration, ironic aside), and the immediately following user turn adopts the suggested change. Coders tag the correction turn and the user-uptake turn.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater coding of correction turns into 'humor-framed' vs 'plain', paired with a binary uptake judgment on the next user turn; agreement would be assessed via Cohen's kappa. Primary metric: correction-acceptance rate (accepted humor-framed corrections / all humor-framed corrections).
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1052 — Delayed Humor Recognition
also known as: Backward Laugh
description: A delayed humorous reaction in which the user first processes an AI output's literal meaning and only afterward perceives it as funny, often laughing at their own delayed comprehension. Distinguished from immediate humor by the lag between exposure and amusement.
operationalDefinition: EVENT. One occurrence would be coded when a user signals amusement (explicit laughter token such as 'haha', or a remark like 'oh, now I get it') referring to an AI turn that is at least one user turn earlier, i.e. amusement is expressed only after an intervening non-humorous turn. The delayed-amusement turn is the unit counted.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Latency measure: for each amusement signal, coders link it to the eliciting AI turn and record the turn-distance and elapsed seconds between exposure and amusement; linkage agreement would be assessed via Cohen's kappa. Could be reported as median exposure-to-amusement latency and share of amusement signals that are delayed (>0 intervening turns).
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1037 — Input-Attribution Realization
also known as: Adjustment Aha
description: A realization by the user that an unsatisfactory AI output stemmed from an unclear or underspecified prompt rather than from a model error, prompting the user to revise the input. Distinguished from generic insight by its attribution of shortfall to one's own input.
operationalDefinition: EVENT. One occurrence would be coded when, following an unsatisfactory AI output, the user produces a turn that (a) attributes the shortfall to their own prior prompt (e.g. 'I wasn't clear', 'I meant') and (b) supplies a more specified reformulation in the same or next turn. The attribution-plus-reformulation turn is the unit counted.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater coding of post-shortfall user turns for self-attribution of the input shortfall and for presence of a more-specified reformulation; agreement would be assessed via Cohen's kappa. Primary metric: self-attribution-and-repair count per 100 user-initiated repairs, optionally cross-checked against measured increase in prompt specificity (added constraints/tokens).
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0625 — Everyday Insight
also known as: The Everyday Insight
description: An insight arising from AI interaction that the user reports as changing how they perceive an everyday phenomenon, such as a new perspective, an unexpected connection, or a useful reinterpretation. Distinguished from in-task problem-solving by its carry-over into ordinary life.
operationalDefinition: EVENT. One occurrence would be coded when a user statement, during or after a session, reports a changed everyday understanding attributable to the interaction (naming the prior view and the revised view). The report is the unit counted; restatements of the same insight within one session are counted once.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Diary/exit self-report: users log everyday insights with a brief before/after description; two raters classify each entry as a genuine perspective change vs restated prior knowledge, agreement would be assessed via Cohen's kappa. Could be reported as verified everyday-insight reports per user per week.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0670 — Rhetorical-Register Discrimination Ability
also known as: The Rhetorical Tone Detector
description: An individual user's graded ability to identify the rhetorical register of an AI output, for example whether it is serious, ironic, neutral, or exaggerated. Distinguished from comprehension of content by its focus on tone, and it varies across users.
operationalDefinition: STATE. The rated quantity is a user's tone-classification accuracy: across a fixed set of AI outputs pre-labelled for rhetorical register by expert consensus, the proportion the user labels correctly. Higher proportion indicates greater ability. Measured per user, not per interaction.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Forced-choice tone-classification test: each user labels a calibrated item set against an expert-consensus key; ability scored as percent correct (or d-prime for irony vs literal). Item-set reliability reported via Cronbach's alpha; expert key reliability via Fleiss' kappa across labellers.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1050 — Deliberate Incoherent Prompting
also known as: Surprise Salad
description: A deliberately incoherent, heterogeneous prompt assembled by the user so that the AI system produces unexpected or creative combinations from the disparate elements. Distinguished from an ordinary prompt by its intentional internal inconsistency.
operationalDefinition: EVENT. One occurrence would be coded when a user prompt deliberately juxtaposes three or more semantically unrelated elements (topics, registers, or domains) in a single request, as judged by coders, with the apparent aim of eliciting novel combination. The composite prompt is the unit counted.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Prompt-level content coding: raters count distinct unrelated elements per prompt and flag intentional juxtaposition; agreement would be assessed via Cohen's kappa. Complementary objective index: lexical-semantic dispersion of the prompt (mean pairwise embedding distance of its content segments). Could be reported as deliberate-mix prompts per session and mean dispersion.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1040 — Tension-Release Humor
also known as: Humor Hook
description: A targeted use of humor within an interaction intended to release built-up tension and restore the user's capacity to keep working. Distinguished from incidental humor by its deployment at a point of friction or stall.
operationalDefinition: EVENT. One occurrence would be coded when a humorous turn occurs immediately after a friction marker (expressed frustration, repeated unsuccessful attempt, or stall) and is followed within two turns by resumption of task-directed activity. The humor turn at the friction point is the unit counted.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Sequential coding: raters tag friction markers, subsequent humor turns, and task-resumption within a two-turn window; boundary agreement would be assessed via Cohen's kappa. Primary metric: task-resumption rate after friction-point humor, with median turns-to-resumption as a secondary latency indicator.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1027 — Positive Affect Spike
also known as: Joy Jolt
description: A brief, unexpected moment of positive affect during an interaction that the user reports as releasing energy or lifting engagement. Distinguished from sustained mood by its short, punctate character.
operationalDefinition: EVENT. One occurrence would be coded when a user emits a discrete positive-affect signal (e.g. 'oh nice!', exclamation of delight) tied to a specific AI turn, lasting a single turn rather than a sustained passage. The affect-signal turn is the unit counted.
measurementSchema: Proposed measurement protocol (not yet empirically validated): In-the-moment self-report: users tap a single-press affect marker when they feel a positive jolt, timestamped against the eliciting turn; optionally validated against rater-coded positive-affect tokens (agreement would be assessed via Cohen's kappa). Could be reported as positive-jolt events per session.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1051 — Humor Cascade
also known as: Giggle Chain
description: A self-reinforcing sequence of humorous reactions across consecutive turns that produces a sense of lightness and can dislodge a cognitive block. Distinguished from a single humorous exchange by its escalating chain of three or more reactions.
operationalDefinition: EVENT. One occurrence would be coded when at least three consecutive turns each contain a humorous reaction that builds on the prior one, forming an unbroken chain. The chain is the unit counted; it ends at the first non-humorous turn. (Differs from Reciprocal Humor Escalation by requiring a longer unbroken run.)
measurementSchema: Proposed measurement protocol (not yet empirically validated): Transcript coding of unbroken humorous chains with a minimum length of three turns; raters record chain length and whether a preceding stall was present, boundary agreement would be assessed via Cohen's kappa. Could be reported as chains per 100 turns and mean chain length.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1044 — Motivation Restoration
also known as: Joy Fix
description: A short interaction intended to restore the user's motivation and positive mood after a dip. Distinguished from ongoing encouragement by its brief, bounded, restorative character.
operationalDefinition: EVENT. One occurrence would be coded when, following a user signal of low motivation or mood, a brief AI intervention (one to two turns) is followed by a user signal of restored willingness to continue (e.g. 'ok, let's keep going'). The intervention-plus-recovery pair is the unit counted.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Pre/post single-item mood-and-motivation self-report (e.g. a 1-7 'ready to continue' rating) bracketing the intervention; effect quantified as the within-user pre-to-post change. Complementary behavioural index: whether the user resumed the task within two turns. Could be reported as mean motivation gain and resumption rate.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1005 — Pause Power
description: The user's practice of deliberately inserting a pause between receiving an AI output and sending the next input, in order to avoid impulsive replies and improve the quality of the following prompt. Distinguished from incidental delay by its intentional, regulatory character.
operationalDefinition: STATE. The rated quantity is the user's deliberate inter-turn pausing: operationalised as the inter-turn interval (seconds between receiving an output and sending the next input) for turns the user marks or self-reports as intentional pauses, contrasted with their baseline interval. Higher relative interval on reflective turns indicates more of the construct.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Behavioural timing: log inter-turn latency per turn; compare median latency on user-flagged reflective turns against the user's overall median (within-subject ratio). Complementary self-report: a short habit scale (e.g. 1-7, 'I pause before replying') for trait-level pausing. Could be reported as reflective-turn latency ratio and habit-scale score.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0088 — Algorithmic Intuition
description: An experienced user's implicit, practice-acquired sense for which kind of input tends to yield which kind of result, applied without deliberate analysis each time. Distinguished from explicit prompting knowledge by being tacit and predictive rather than reasoned step by step.
operationalDefinition: STATE. The rated quantity is predictive accuracy without deliberation: given candidate prompts, the user forecasts which will yield the better result; accuracy is the proportion of forecasts confirmed by subsequent outputs under blind evaluation. Higher accuracy, achieved with low reported deliberation, indicates more of the construct. Measured per user.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Prediction task: users rank or choose among candidate prompts before execution; outcomes are graded blind by independent raters, and the user's hit rate is compared to chance. Deliberation is captured by decision latency and a brief 'how much did you reason' item. Could be reported as above-chance prediction accuracy at low deliberation; rater grading reliability via ICC.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1030 — Non-Frustrated Error Recovery
also known as: Oops Moment
description: An AI error that the user consciously accepts and treats as a learning signal, responding with quick correction and continuation rather than annoyance. Distinguished from an unremarked error by the user's explicit non-frustrated, corrective response.
operationalDefinition: EVENT. One occurrence would be coded when an identifiable AI error is followed by a user turn that (a) acknowledges the error without expressed frustration and (b) issues a corrective instruction, with the task continuing in the next turn. The error-plus-non-frustrated-correction pair is the unit counted.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater coding of post-error user turns on two dimensions, frustration tone (present/absent) and corrective action (present/absent); agreement would be assessed via Cohen's kappa. Primary metric: share of AI errors met with non-frustrated correction; secondary: turns-to-task-continuation.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1059 — Day Highlight
description: A user's deliberate act of marking a positive moment from the day, often with AI support, to consolidate it emotionally. Distinguished from passive recall by the explicit selection and recording of the moment.
operationalDefinition: EVENT. One occurrence would be coded when the user explicitly designates a specific positive moment of the day for recording or reflection (e.g. naming it as 'the highlight'), producing a recorded or stated entry. The designation entry is the unit counted; one per reviewed period unless distinct moments are named.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Structured daily-review log: users record a single named positive highlight per session; completion is tracked as adherence (entries / eligible days), and entry quality (specific event vs vague) is rated by two coders with Cohen's kappa. Could be reported as adherence rate and specific-entry proportion.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1012 — Minimal Calming Cue
also known as: Serenity Spark
description: A small verbal cue from the AI that elicits a disproportionately large sense of calm in the user. Distinguished from extended de-escalation by its minimal linguistic footprint relative to its calming effect.
operationalDefinition: EVENT. One occurrence would be coded when a brief AI utterance (a single short phrase or sentence) is immediately followed by a user signal of increased calm (explicit statement, or a measured drop in agitation markers such as exclamation density or all-caps). The cue-plus-calming-response pair is the unit counted.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Pre/post momentary-calm self-report (single-item 1-7) bracketing the cue, paired with a length check confirming the cue is brief (token count below a set threshold). Effect quantified as calm gain per token of cue. Complementary objective marker: change in agitation tokens before vs after. Could be reported as mean calm gain and effect-per-token.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1054 — Playful Correction
also known as: Fun Correct
description: An AI error correction delivered in a playful tone that preserves the user's willingness to learn rather than provoking defensiveness. Distinguished from a neutral correction by its playful register and its effect on receptiveness.
operationalDefinition: EVENT. One occurrence would be coded when an AI correction turn carries a playful register and the following user turn shows continued engagement (accepts the point, asks a follow-up, or proceeds) rather than defensiveness (denial, disengagement). The correction-plus-engaged-response pair is the unit counted.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater coding pairing correction register (playful/neutral) with the subsequent user response (engaged/defensive); agreement would be assessed via Cohen's kappa. Primary metric: continued-engagement rate following playful corrections, compared against neutral corrections as a within-data contrast.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1022 — Progress Reflection
also known as: Pride Pulse
description: A moment in which achieved progress is reflected back to the user, with the aim of strengthening their sense of self-efficacy. Distinguished from generic praise by its specific reference to concrete progress already made.
operationalDefinition: EVENT. One occurrence would be coded when an AI (or user-self) turn explicitly references a specific, completed increment of progress (naming what was accomplished) in a way that attributes it to the user. The progress-reflection turn is the unit counted; vague praise without a named accomplishment does not qualify.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Pre/post self-efficacy self-report on the relevant task using a brief validated scale (e.g. a short General Self-Efficacy item set, 1-5); effect quantified as within-user change after progress-reflection turns. Coders verify each turn names a concrete accomplishment, with agreement would be assessed via Cohen's kappa. Could be reported as self-efficacy change and named-accomplishment rate.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1025 — Elevated Positive-Affect Passage
also known as: Heart High
description: A heightened affective state of connection, joy, or meaning that a user reports during an interaction. Distinguished from a brief affect spike by being a sustained, graded condition rather than a single moment.
operationalDefinition: STATE. The rated quantity is the intensity of felt connection/joy/meaning during a defined interaction passage, self-reported on a graded scale at the end of the passage. Higher ratings indicate a stronger state. Rated per passage, not counted as discrete events.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Multi-item affect self-report combining a connection/closeness item set (e.g. Could be reported as mean state-intensity score per passage.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1032 — Error Composure
also known as: Learn Smile
description: A user's disposition to register AI or task errors with composure because the learning value of the error is clearly apparent. Distinguished from a single calm reaction by being a graded, trait-like tendency across errors.
operationalDefinition: STATE. The rated quantity is the user's typical composure-on-error: across multiple error episodes, the proportion met with a calm, learning-oriented response (versus frustration), or an equivalent self-reported trait rating. Higher proportion or rating indicates more of the disposition. Aggregated per user.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Aggregated behavioural index: across a user's error episodes, raters code each response as composed/learning-oriented vs frustrated (Cohen's kappa for agreement), yielding a per-user composure proportion. Complementary trait self-report: a short error-orientation scale (learning vs strain subscales, 1-5). Could be reported as composure proportion and error-orientation score.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1058 — Serendipitous Finding
also known as: Discovery Ding
description: An unsought finding that the user did not set out to obtain but recognizes as valuable once it appears during interaction. Distinguished from a sought answer by its incidental, unplanned origin.
operationalDefinition: EVENT. One occurrence would be coded when a user turn flags an AI output as valuable while also indicating it was outside the stated goal of the current request (e.g. 'I wasn't looking for that, but that's useful'). The serendipity-flagging turn is the unit counted.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater coding of user turns for two co-present features, expressed value and explicit off-goal origin, against the session's stated task; agreement would be assessed via Cohen's kappa. Complementary self-report: users tag outputs as 'unexpected and useful'. Could be reported as serendipitous-finding events per session.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1026 — Motivation Recovery
also known as: Motivation Mend
description: A targeted restoration of the user's motivation following frustration or exhaustion during an interaction. Distinguished from general encouragement by its placement after an explicit motivational low and its restorative intent.
operationalDefinition: EVENT. One occurrence would be coded when, after a user signal of frustration or exhaustion, an intervention is followed by a user signal of recovered drive to continue the task. The low-signal, intervention, recovery-signal sequence is the unit counted.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Pre/post single-item motivation self-report (1-7 'drive to continue') bracketing the intervention; effect quantified as within-user change, with a behavioural check on whether the task resumed. Could be reported as mean motivation recovery and resumption rate. (Distinguished from AUG-1044 by requiring an explicit frustration/exhaustion antecedent rather than a general mood dip.)
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1010 — Pace-Reduction Mode
also known as: Calm Wave
description: A regulating interaction mode characterised by reduced pace, sentence rhythm, and complexity in the AI's turns; a pattern in which the user's reported perceived tempo of the exchange is observed to be lower. Distinguished from distraction by being a slowing of the exchange rather than a topic shift.
operationalDefinition: STATE. The rated quantity is the degree of deliberate slowing within a passage, indexed by measurable output features: reduced words-per-turn, shorter sentences, and lower lexical complexity relative to the session baseline. A larger reduction across these features indicates more of the mode. Rated/measured per passage.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Text-metric comparison of 'calming' passages against the session baseline: mean sentence length, words per turn, and a readability index (e.g. Flesch reading-ease); paired with a user momentary-tempo self-report (1-7 'how rushed do you feel'). Could be reported as percent reduction in complexity metrics and change in felt tempo.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1029 — State Gratitude Appraisal
also known as: Gratitude Glow
description: A reflective state in which the user consciously notices positive aspects of an experience and anchors them emotionally, often prompted by the interaction. Distinguished from a fleeting good feeling by its deliberate, reflective appraisal.
operationalDefinition: STATE. The rated quantity is the intensity of state gratitude during or after a reflective passage, self-reported on a graded scale, optionally corroborated by the user naming specific positives. Higher ratings indicate a stronger state. Rated per passage, not counted as discrete events.
measurementSchema: Proposed measurement protocol (not yet empirically validated): A short state-gratitude self-report (a few items, 1-7) with internal-consistency reporting (Cronbach's alpha). Complementary coding: count of distinct positives the user names, with rater agreement would be assessed via Cohen's kappa. Could be reported as mean state-gratitude score and named-positives count.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1034 — Productive Reframe
also known as: Twist Dance
description: A creative change of direction in the interaction that initially disorients the user but then yields a new and better quality of output. Distinguished from an unproductive tangent by its eventual gain in output quality.
operationalDefinition: EVENT. One occurrence would be coded when an abrupt directional shift (by user or AI) is first met with a user signal of irritation or confusion and is subsequently followed, within the same session, by a user judgment that the new direction improved the result. The shift-irritation-improvement sequence is the unit counted.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Sequential rater coding of three linked features, directional shift, initial irritation/confusion, and later quality endorsement; agreement would be assessed via Cohen's kappa. Complementary objective check: pre-shift vs post-shift output rated for quality by blind judges (ICC for rater consistency). Could be reported as productive-pivot events per session and mean pre/post quality gain.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1064 — Emotional Overload Reset
also known as: Heart Reset
description: A deliberate act of resetting emotional overload so the user can re-engage with their own needs. Distinguished from ongoing self-soothing by being a discrete, marked break that re-orients the user after overload.
operationalDefinition: EVENT. One occurrence would be coded when, following a user signal of emotional overload, the user performs an explicit reset act (a stated break, grounding step, or re-centering prompt) and the next user turn indicates re-orientation toward their own needs or priorities. The overload, reset, re-orientation sequence is the unit counted.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Pre/post momentary affect-load self-report (single-item 1-7 'how overwhelmed do you feel') bracketing the reset act; effect quantified as the within-user drop in load. Coders verify an explicit reset act occurred, with agreement would be assessed via Cohen's kappa. Could be reported as mean load reduction per reset and reset-completion rate.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1067 — Habitual Daily Integration
also known as: Daily Duo
description: A habitual, recurring division of labor in which a user routinely incorporates an AI system into everyday tasks, such that the pairing becomes a settled part of the daily workflow rather than an occasional consultation.
operationalDefinition: Graded condition rated from a user's interaction history over a fixed reference period (e.g., 14 days): the proportion of days on which the user initiates at least one AI-assisted task within a defined routine slot. Higher proportions indicate a more established daily pairing.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Behavioral-log metric: days-with-AI-session divided by total-days over a 14-day window, optionally cross-validated against a short habit-strength self-report.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1015 — Calm Shift
description: A discrete, user-recognized transition during an AI-assisted session from an agitated, problem-focused mode into a calmer, more reflective one, marked at the point where the interaction's tone and pace visibly shift.
operationalDefinition: One occurrence would be coded when a rater (or the user, in real time) marks the turn boundary at which the interaction changes from rapid problem-pushing to slower, reflective exchange (e.g., a shift to summarizing, reframing, or step-back questions). The marked turn is the unit.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater-coded transition tagging on session transcripts by two independent coders; inter-rater agreement could be reported as Cohen's kappa; result expressed as the count of marked transitions per session.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1018 — Felt-Coherence State
also known as: Harmony Hug
description: A self-reported moment of strong felt coherence during interaction, in which prior internal tension eases and the user reports a sense of things fitting together. The construct refers to the user's subjective state, not to any property of the AI.
operationalDefinition: Graded condition: immediately after a target exchange, the user rates the degree to which they experienced reduced tension and an increased sense of coherence. It is an anchored single-administration rating tied to a specific exchange, not a counted discrete event.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Post-exchange self-report on two 1-7 Likert items (tension release; felt coherence), optionally paired with a brief calmness self-report for convergent validity.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1023 — Momentary Optimism Appraisal
also known as: Excitement Bubble
description: A transient elevated-affect appraisal during interaction in which the user weighs perceived possibilities as larger than perceived risks. It denotes the user's momentary optimistic stance toward options surfaced in the exchange.
operationalDefinition: Graded condition rated for a defined exchange: the user, or a coder using verbal markers, rates perceived-opportunity salience relative to perceived-risk salience. It is operationalized as a balance rating rather than a binary occurrence.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Self-report bipolar 1-7 rating (risk-focused to opportunity-focused) for the target exchange; optionally, transcript coding of opportunity- versus risk-laden statements with a ratio reported.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1042 — Positive Pivot
description: A reformulation move in which a stated problem is recast into one or more actionable options. It denotes the observable act of converting a problem framing into a choice set within the dialogue.
operationalDefinition: One occurrence would be coded when a turn contains an explicit recasting of a previously stated problem into discrete actionable options (a problem statement in an earlier turn followed by an option-set framing in a later turn). The reformulating turn is the unit.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Transcript coding by two raters using a problem-to-options rubric; reliability via Cohen's kappa; result could be reported as a count per N=10 turns.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1066 — No-Stress Swap
description: An act of replacing an overwhelming task specification with a more realistic, smaller-scope alternative during an AI-assisted exchange, observable as a deliberate substitution of the task under discussion.
operationalDefinition: One occurrence would be coded when a turn replaces a previously requested task with an explicitly narrower or more feasible task (original scope stated earlier, substituted reduced scope adopted in a later turn). The substitution turn is the unit.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater coding of scope-reduction substitutions on transcripts by two coders; agreement would be assessed via Cohen's kappa; result could be reported as the count of substitution events per session.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1061 — Gratitude Turn
description: A turn in which a user's framing shifts from focusing on what is lacking to acknowledging what is valued, observable as an explicit appreciation statement within the dialogue.
operationalDefinition: One occurrence would be coded when a user turn contains an explicit appreciation or gratitude statement that reframes a prior deficit-focused statement. The appreciation-containing turn is the unit of count.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Transcript coding for gratitude-reframing utterances by two raters; reliability via Cohen's kappa; could be reported as a count per session, optionally validated against a short gratitude self-report at participant level.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1006 — Divergent Ideation Burst
also known as: Idea Rain
description: A bounded divergent-generation episode in which the dialogue deliberately prioritizes quantity of ideas over quality, with selection and evaluation deferred to a later phase. It denotes an ideation burst rather than an evaluative exchange.
operationalDefinition: One episode would be coded when a contiguous stretch of turns is led by idea enumeration without evaluation (operationally, a run of turns in which candidate ideas accumulate and no selection or critique turn intervenes). The episode boundaries are the unit.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Transcript segmentation into ideation episodes by two coders (kappa for boundary agreement), plus an objective idea-count throughput (distinct candidate ideas per minute) within each episode.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1014 — Relaxation Edit
description: A revision act applied to a text or message with the explicit aim of lowering its perceived pressure, harshness, or demand level while preserving its informational content.
operationalDefinition: One occurrence would be coded when a turn produces a revised version of prior text whose evident purpose is tone-softening (reduction of imperative, severity, or demand markers) without changing the underlying request. The revision turn is the unit.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Pre/post rater rating of revised texts on a 1-7 harshness-and-demand scale (paired comparison) with two raters and ICC for agreement; result could be reported as the mean harshness reduction per edit.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1019 — Deliberate Slow Restart
also known as: Calm Rewind
description: A deliberate restart move in which a conversation or reasoning thread is returned to an earlier point to be re-begun more slowly and explicitly, observable as an intentional back-up and re-entry in the dialogue.
operationalDefinition: One occurrence would be coded when a turn explicitly references an earlier state of the exchange and re-initiates from there with a slower or more explicit framing (for example, proposing to go back and proceed step by step). The re-initiating turn is the unit.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater coding of restart and back-up moves on transcripts by two coders; agreement would be assessed via Cohen's kappa; result could be reported as the count of restart events per session.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1043 — Calm Reset
description: A brief pause and re-centering taken after an intense stretch of AI-assisted work, allowing attention to recover before the next task. It denotes a short, in-rhythm micro-break distinct from extended rest.
operationalDefinition: One occurrence would be coded when an interaction gap of a defined short duration (e.g., 1 to 10 minutes) follows an intensive task block and precedes resumption, identifiable from session timestamps as a deliberate within-session pause.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Behavioral-log timing: detection of short inter-task gaps (1 to 10 minutes) from session timestamps; count of such micro-breaks per work block, optionally annotated by the user as intentional versus incidental.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1055 — Goal-Match Acknowledgment
also known as: High-Five Moment
description: A point in an exchange at which the user registers shared satisfaction because the delivered output, the original intention, and the result align. It denotes a user-marked success acknowledgment within the dialogue.
operationalDefinition: One occurrence would be coded when a user turn explicitly acknowledges that the output matched the intended goal (an alignment or approval statement immediately following a delivered result). The acknowledging turn is the unit.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater coding of explicit success-acknowledgment turns following deliverables, by two coders, with Cohen's kappa; could be reported as a count per session, optionally with a convergent single-item satisfaction rating (1-7) at the acknowledgment point.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0368 — Aesthetic-Formulation Pleasure
also known as: The Syntax Smile
description: A brief, low-intensity pleasure a user reports when an AI produces a notably elegant, apt, or aesthetically pleasing formulation. The construct refers to the user's momentary aesthetic-appreciation state, not to any emotion of the system.
operationalDefinition: Graded condition: immediately after a flagged output, the user rates the intensity of aesthetic pleasure experienced with that formulation. It is tied to a specific output and rated, rather than counted as a discrete behavioral event.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Single-item post-output aesthetic-pleasure rating (1-7), optionally combined with items from an aesthetic-experience scale for convergent validity; aggregated as the mean rating across flagged elegant outputs.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0918 — Gesture Language
also known as: The Gesture Language
description: The overall modality by which humans communicate with embodied AI systems through gestures, including hand signals, posture, and movement that the system detects and interprets. It names a communication channel rather than a single bounded event.
operationalDefinition: Superordinate framework term; not operationalized as a per-interaction event. Specific measurable sub-constructs (for example, gesture-recognition accuracy or repair rate) would be operationalized individually.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Not directly measured as a single instrument; superordinate modality term. Component studies may report gesture-recognition accuracy or interpretation error rates, but the channel itself is not one measured construct.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0975 — Vigilance Rhythm
also known as: The Vigilance Rhythm
description: A paced verification practice in which a user checks AI output at deliberate intervals rather than monitoring continuously, on the premise that spaced focused review sustains attention quality better than unbroken vigilance across long interactions.
operationalDefinition: Graded condition characterizing how a user's verification activity is distributed over a session: the regularity and spacing of explicit verification acts. It is operationalized as the cadence or dispersion of verification events rather than as a single occurrence.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Behavioral-log analysis of verification-act timestamps within a session: the inter-check interval distribution (mean and coefficient of variation) could be reported as a cadence index, optionally paired with sustained-attention performance over time.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1002 — Brief Activating Prompt
also known as: Quick Kick
description: A very short, precise prompt that produces an immediate effect on the user's motivation, focus, or readiness to decide, providing starting impetus rather than substantive depth.
operationalDefinition: One occurrence would be coded when a user-issued prompt below a defined length threshold (e.g., 12 words or fewer) is followed by a self-reported or behaviorally evident activation effect, such as task resumption within a short window. The short prompt is the unit.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Behavioral-log join of prompt length (word or token count) to a post-prompt single-item activation rating (1-7) or to time-to-next-task-action latency; reported per qualifying short prompt.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1008 — Tone-Appropriateness Query
also known as: Vibe Check
description: A brief meta-query that probes the emotional, social, or cultural fit of an output, separate from its factual content. It denotes an explicit request to appraise tone or appropriateness rather than correctness.
operationalDefinition: One occurrence would be coded when a user turn explicitly requests an appraisal of an output's tone or its social or cultural appropriateness, as opposed to its factual accuracy. The appraisal-requesting turn is the unit.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater coding distinguishing fit and appropriateness queries from content and accuracy queries on transcripts, by two coders, with Cohen's kappa; could be reported as the count of fit-queries per session.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1028 — Optional Creative Prompt
also known as: Inspiration Breeze
description: A light creative prompt or suggestion offered without any obligation to act on it, functioning as an optional spark rather than a directive within the exchange.
operationalDefinition: One occurrence would be coded when an AI turn offers a creative suggestion explicitly framed as optional or non-binding, with no implied requirement to implement it. The optional-suggestion turn is the unit.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater coding of optional versus directive creative suggestions on transcripts, by two coders, with Cohen's kappa; could be reported as the count of optional creative prompts per session.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1031 — Mid-Task Micro-Correction
also known as: Correction Kick
description: A brief, precise intervention that brings an output back on course without interrupting the overall working flow, observable as a short steering correction inserted mid-task.
operationalDefinition: One occurrence would be coded when a user turn issues a short corrective instruction that adjusts the output's direction and is immediately followed by continued task progress, with no full restart. The corrective turn is the unit.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater coding of short mid-task corrections (distinguished from restarts and abandonments) on transcripts, by two coders, with Cohen's kappa; result could be reported as a count per N=10 turns.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1069 — Low-Friction Interaction
also known as: Smooth Sail
description: A condition of low-friction collaboration in which an AI-assisted task proceeds with few corrections, clarifications, or breakdowns. It denotes a graded smoothness of the working exchange rather than a single moment.
operationalDefinition: Graded condition over a task or session, inversely indexed by the rate of repair turns (corrections, clarification requests, restarts). Lower repair rates indicate a smoother condition; it is rated or computed across the segment, not counted as one event.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Behavioral metric: repair-turn rate equal to (corrections plus clarification requests plus restarts) divided by total turns over the segment, optionally complemented by a post-task 1-7 perceived-smoothness self-report.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1060 — Specific Gratitude Reference
also known as: Gratitude Drop
description: A single, concrete point of focus that makes gratitude specific and tangible, observable as one explicitly named object of appreciation within the exchange.
operationalDefinition: One occurrence would be coded when a turn names a single concrete object of gratitude, that is, one specific item rather than a general expression of thanks. The turn naming the specific focus is the unit.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater coding distinguishing specific, concrete gratitude foci from generic thanks on transcripts, by two coders, with Cohen's kappa; could be reported as the count of concrete gratitude foci per session.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1003 — Collaborative Phrasing Refinement
also known as: Word Hunt
description: A collaborative, iterative search for the single formulation that precisely fits meaning, tone, and context, defined by repeated refinement of candidate wordings until convergence on one expression.
operationalDefinition: One episode would be coded when a contiguous run of turns iteratively refines a target phrasing (two or more successive candidate reformulations aimed at the same expression slot) before settling. The refinement run is the unit.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Transcript segmentation of phrasing-refinement episodes by two coders (kappa for boundary agreement); within each episode, the count of candidate reformulations until convergence could be reported as iteration depth.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1011 — Relaxation Breath
description: A brief AI-guided breathing or attention exercise intended to lower stress or ease a transition between cognitive states, observable as a discrete guided micro-intervention within the session.
operationalDefinition: One occurrence would be coded when an AI turn delivers a structured short breathing or attention-direction instruction (a guided exercise prompt) within the interaction. The guidance turn is the unit.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater coding of guided breathing and attention-exercise turns on transcripts (count per session), with agreement would be assessed via Cohen's kappa; optionally a pre/post single-item stress rating (1-7) to index the effect.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1013 — Supportive Warmth Style
also known as: Feel-Good Warm
description: An emotionally supportive interaction style that conveys safety, belonging, and acceptance without forcing problem-solving. It denotes a graded stylistic quality of the exchange as perceived by the user.
operationalDefinition: Graded condition: raters or the user rate the degree to which an exchange conveys warmth, acceptance, and non-pressuring support, on a defined scale across the segment. It is rated as an intensity, not counted as a discrete event.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater coding of supportive and warmth style on a 1-5 anchored scale (two raters, ICC for agreement), optionally complemented by a user-perceived-warmth self-report (1-7) per session.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1017 — Gentle Layer
description: An added emotional or linguistic layer that buffers harshness and makes communication more tolerable without distorting the underlying content. It denotes a graded softening property of a message relative to its plain form.
operationalDefinition: Graded condition assessed per message: the degree of tone-buffering present relative to a neutral baseline, while content equivalence is held constant. It is rated as buffering intensity with a content-preservation check, not counted as an event.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Paired rater rating of harshness or softening on a 1-7 scale (two raters, ICC), plus a separate content-equivalence verification step checking that the buffered version preserves the information; could be reported as a softening delta with content held constant.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1035 — Gentle Nudge
description: A minimal reminder or correction that provides orientation without exerting pressure, observable as a brief, low-intensity prompt that redirects without using demand language.
operationalDefinition: One occurrence would be coded when a turn issues a short reminder or correction that lacks imperative or pressure markers and offers light redirection (a low-intensity steering cue). The low-pressure cue turn is the unit.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater coding of low-pressure reminder and correction turns, scored for the absence of demand markers, on transcripts, by two coders, with Cohen's kappa; could be reported as a count per session.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1036 — Error Normalization Stance
also known as: Mistake Friend
description: A user's working stance toward errors in which mistakes are treated as ordinary, expected by-products of productive work rather than as non-successes, lowering the threshold for surfacing and correcting them during AI-assisted tasks.
operationalDefinition: STATE. The user's error-orientation during a task block is rated on whether errors are surfaced without avoidance and treated as routine: rated on a graded scale from defensive/error-avoidant to matter-of-fact/error-tolerant, anchored to observable handling of the session's mistakes (e.g., openly naming an error and continuing vs. concealing or abandoning).
measurementSchema: Proposed measurement protocol (not yet empirically validated): A short error-tolerance self-report administered post-session (items covering learning-from-errors and error-competence, positively keyed, together with error-strain and covering-up-errors, reverse keyed), on a 7-point Likert format; could be reported as a composite error-tolerance score with internal consistency (Cronbach's alpha).
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1038 — Low-Friction Topic Re-Entry
also known as: Chill Comeback
description: A discrete resumption in which a user returns to a previously interrupted or misunderstood topic without visible tension or self-reproach, picking the thread back up smoothly within an AI dialogue.
operationalDefinition: EVENT. One occurrence is coded when, after an interruption or a flagged misunderstanding, the user re-opens the same topic in a later turn and the re-entry turn contains no markers of frustration or blame (no self-criticism, no complaint about the prior break). The re-entry turn is the countable marker.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater coding of transcript turns: two trained coders flag topic re-entries and judge presence/absence of tension markers; inter-rater reliability via Cohen's kappa; could be reported as count of low-tension re-entries per session.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1041 — Adaptive Continuity
also known as: Flex Flow
description: A graded condition in which a user incorporates adjustments, corrections, or scope changes during a task while sustaining task momentum, so that revisions do not break the ongoing work rhythm of an AI-assisted session.
operationalDefinition: STATE. Rated as the degree to which adjustments are absorbed without loss of momentum across a task block: scored from low (each change causes a stall or restart) to high (changes are integrated with continuous progress), anchored to observable indicators such as continued forward task moves immediately after a correction.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Behavioral process metric: from session logs, compute median post-adjustment resumption latency (seconds from a correction/scope-change to the next on-task action) and the proportion of adjustments followed by continued progress without restart; could be reported as a continuous index, not a count.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1045 — Drift Fix
description: A deliberate corrective move in which a user redirects a conversation that has wandered off back toward its core goal, re-anchoring the exchange to the original task within an AI dialogue.
operationalDefinition: EVENT. One occurrence is coded when a user turn explicitly references the original goal or task after the immediately preceding turns have moved onto a side topic, and reorients the exchange back to it (e.g., 'let's get back to X'). The redirecting turn is the countable marker.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater coding with a topic-tracking scheme: coders mark off-topic drift segments and the user turn that closes them; count redirect events per session and report drift-segment duration before each redirect; inter-rater agreement would be assessed via Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1047 — Post-Failure Affect Recovery
also known as: Bounce Back
description: A rapid affective recovery in which a user's negative reaction following a unsuccessful attempt subsides quickly and the user re-engages with the task within an AI-assisted session.
operationalDefinition: EVENT. One occurrence is coded when a turn expressing negative affect after a unsuccessful attempt is followed within a short window (e.g., the next 1-2 user turns) by a turn showing neutral or positive affect and renewed task engagement. The recovery turn is the countable marker; recovery time (turns or seconds) is recorded.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Affect-trajectory coding: trained coders rate per-turn affect valence (negative/neutral/positive) around failure events; recovery would be operationalised as return to non-negative valence; report median recovery latency and event count; inter-rater reliability via weighted Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1053 — Brief Positive-Affect Lift
also known as: Smile Boost
description: A brief mood lift produced by an AI exchange that raises a user's momentary positive affect without causing distraction or exaggerated elation, leaving task focus intact.
operationalDefinition: EVENT. One occurrence is coded when a short within-session rise in self-reported or observably expressed positive affect follows a specific AI turn, and the following turns show no drop in on-task behavior. The affect-rise tied to that turn is the countable marker.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Momentary affect probe: single-item positive-affect rating (0-100 visual analog scale) sampled immediately before and after the triggering exchange; the pre-post difference indexes the lift; aggregated as mean within-person change with a paired test, paired with a task-focus check (on-task action present in the next turn).
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1056 — Disproportionate Small Utility
also known as: Mini Magic
description: A small, everyday AI intervention whose practical usefulness is disproportionately large relative to the minimal effort or input it required, observed in routine tasks.
operationalDefinition: EVENT. One occurrence is coded when a low-effort request (short prompt, single turn) yields an output the user marks as highly useful or time-saving, producing a high benefit-to-effort contrast for that exchange. The user's usefulness acknowledgement on a low-effort turn is the countable marker.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Benefit-to-effort rating per flagged exchange: user rates perceived usefulness (1-7) and input effort (1-7); the event is indexed by a high-usefulness/low-effort combination; report the proportion of low-effort exchanges rated highly useful, with optional time-saved estimate in minutes.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1057 — Unobtrusive Assistance Mode
also known as: Flow Friend
description: A graded mode of AI use in which the system functions as an unobtrusive support to the user's work rhythm, contributing when prompted without overriding or steering the activity.
operationalDefinition: STATE. Rated across a session as the degree of unobtrusiveness: scored from intrusive (AI introduces unrequested redirections that interrupt the user's flow) to unobtrusive (AI contributions stay within the user's initiative), anchored to the ratio of user-initiated to AI-initiated topic shifts.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Initiative-balance index from transcript coding: count user-initiated vs. AI-initiated topic shifts and unsolicited redirections; report AI-initiation share (AI-initiated / total shifts) as a continuous unobtrusiveness index; inter-rater agreement would be assessed via Cohen's kappa on the initiation labels.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1063 — Sufficiency Acceptance State
also known as: Contentment Zone
description: A stable condition of inner calm during or after AI-assisted work in which the user experiences no pressure to further optimize, accepting the current result as sufficient.
operationalDefinition: STATE. Rated as the level of felt calm and absence of optimization pressure over a defined period: scored on a graded scale from high optimization pressure (continued reworking, dissatisfaction) to settled sufficiency (work accepted as good enough), based on self-report and observable cessation of further reworking.
measurementSchema: Proposed measurement protocol (not yet empirically validated): A brief calmness/arousal self-report: a short calmness/low-tension subscale covering the calm-tense (low-arousal) dimension (calm-vs-tense items, reverse-scored for calm); 5-point Likert, administered immediately after the work block; could be reported as a mean state-calm score with internal consistency (Cronbach's alpha).
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1068 — End-of-Day Positive Closure
also known as: End-of-Day High
description: A positive emotional closure at the end of a work day, supported by an AI exchange, that rounds off the day's activity and helps the user consolidate or integrate what was done.
operationalDefinition: EVENT. One occurrence is coded when an end-of-day session concludes with a user-expressed sense of completion or positive closure and an explicit recap/integration of the day's work. The closing turn containing both positive closure and recap is the countable marker (at most one per day).
measurementSchema: Proposed measurement protocol (not yet empirically validated): End-of-day diary entry: single daily item rating sense of positive closure (1-7) plus a checkbox for whether the session included a recap/integration; aggregated across days as the proportion of days with positive closure and mean closure rating; suitable for multilevel analysis over a diary window.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1009 — Reset Noise
description: A deliberate injection of unstructured or disordered input intended to break out of an entrenched line of thinking; structure is reintroduced only in a subsequent step after the fixation is loosened.
operationalDefinition: EVENT. One occurrence is coded when a user intentionally supplies deliberately unstructured input (free association, scattered fragments, randomized prompts) explicitly to escape a stuck line of reasoning, followed in a later turn by a move to re-impose structure. The intentional unstructured-input turn is the countable marker.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater coding of prompt structure: coders classify user turns as structured vs. deliberately unstructured and mark stated de-fixation intent and a later restructuring turn; count qualifying sequences per session; inter-rater agreement would be assessed via Cohen's kappa on the structure labels.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1020 — Activation Impulse
also known as: Motivation Fire
description: An activating impulse from an AI exchange that releases energy for action without overwhelming the user; it is brief, clear, and oriented toward taking the next step.
operationalDefinition: EVENT. One occurrence is coded when a short AI turn is followed within the next 1-2 user turns by an observable initiation of action (the user starts or commits to a concrete next step) together with self-reported readiness to act, and no markers of overwhelm. The action-initiation following the impulse is the countable marker.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Pre/post readiness probe plus behavioral follow-through: single-item action-readiness rating (1-7) before and after the impulse, combined with a log check for whether a concrete next step was initiated within the following turns; report mean readiness change and follow-through rate.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1033 — Deliberate Restart
also known as: Reset Ruck
description: A clear restart undertaken when a user's thinking or the dialogue has become tangled, in which the user breaks off the snarled line and re-opens the task from a fresh starting point.
operationalDefinition: EVENT. One occurrence is coded when, following turns showing confusion or a tangled exchange (repair attempts, contradictions, circularity), a user turn explicitly abandons the current line and restarts (e.g., 'let's start over'). The explicit restart turn is the countable marker.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater coding of restart events: coders flag preceding tangle indicators and the explicit restart turn that resets the task; count restarts per session and record the length of the tangled stretch preceding each; inter-rater agreement would be assessed via Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1065 — Morning Orientation Routine
also known as: Morning Magic
description: A structuring, positive start to the day created by a short AI interaction that helps a user organize and orient the day's tasks at its outset.
operationalDefinition: EVENT. One occurrence is coded when a brief AI session early in the user's day produces an explicit plan or ordered task list for that day and the user reports a positive, organized start. The morning planning turn that yields an ordered day-structure is the countable marker (at most one per day).
measurementSchema: Proposed measurement protocol (not yet empirically validated): Morning diary entry: daily items for whether the session produced a concrete day-plan (checkbox) and a 1-7 rating of feeling organized/positively oriented at the start; aggregated across days as proportion of structured mornings and mean orientation rating over the diary window.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0204 — Conversational Afterimage
also known as: The Conversational Afterimage
description: The aftereffect of an intensive AI session in which a user later carries over formulations, thinking structures, or argumentation patterns from the dialogue into their everyday communication, often without awareness of the source.
operationalDefinition: EVENT. One occurrence is coded when a distinctive formulation, frame, or argument structure introduced during an AI session reappears in the same user's later, non-AI communication (speech or writing) within a defined follow-up window. The matched carry-over instance in post-session communication is the countable marker.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Matched-feature transfer coding: build a per-session lexicon/structure list of distinctive AI-introduced features, then have raters scan later user-authored samples for matches; count carry-over instances per follow-up window; inter-rater agreement would be assessed via Cohen's kappa on match judgments.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0045 — Indexical Memory
description: A longitudinal shift of the remembering function from retaining content itself toward retaining where and how the content can be retrieved through AI, reflecting a change in how a person manages personal knowledge under sustained AI use.
operationalDefinition: TREND. Tracked over repeated measurements as the change in the balance between content recall and retrieval-path recall: at each timepoint, record the share of queried items for which the person recalls the substantive content versus only the location/route to retrieve it; the trend is the change in that share across the observation window.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Repeated cued-recall protocol: at multiple timepoints, probe a fixed item set and classify each response as content-recall vs. location/route-recall; track the location-recall proportion over time as the longitudinal index; report the trajectory (e.g., slope across sessions) rather than a single-session count.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0345 — Capability-Limit Encounter
also known as: The Wall Check
description: The moment a user encounters the limits of an AI system: the system cannot solve the task, returns incorrect information, or exposes a comprehension boundary, giving the user direct evidence of what the system cannot do.
operationalDefinition: EVENT. One occurrence is coded when, within an exchange, the AI fails on the user's request in an observable way (explicit inability, a detectable error, or a stated comprehension limit) and the user registers this limit in a subsequent turn. The user's recognition turn following the AI limit is the countable marker.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Failure-event rater coding: coders annotate AI-limit events (refusal/inability, factual error, comprehension limit) and the user turn that registers them; count limit-recognition events per session and categorize by limit type; inter-rater agreement would be assessed via Cohen's kappa on event type.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1007 — Sudden Comprehension Event
also known as: Clarity Click
description: An abrupt transition from diffuse uncertainty to a clearly structured mental representation, frequently prompted by a precise restatement of the user's own thoughts by the AI.
operationalDefinition: EVENT. One occurrence is coded when a user turn marks a sudden gain in clarity (e.g., an explicit 'now it makes sense' / 'that's exactly it') in the turn immediately following an AI restatement or reframing of the user's input. The user's clarity-acknowledgement turn is the countable marker.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Cued retrospective rating plus marker coding: raters flag clarity-acknowledgement turns following an AI reframing (inter-rater agreement would be assessed via Cohen's kappa), and the user gives a single-item insight-suddenness rating (1-7) for flagged moments; report event count per session and mean suddenness.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1016 — Relaxed Return
description: A gentle re-connection to a topic after overload, distraction, or a break, in which the user resumes work without self-blame or pressure, easing back into the task.
operationalDefinition: EVENT. One occurrence is coded when, after a gap caused by overload/distraction/abandonment, the user re-engages the prior topic in a measured, low-pressure way, and the re-entry turn contains no self-blame or urgency markers. The gentle re-entry turn is the countable marker.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater coding of resumptions with a pressure-marker checklist: coders flag topic resumptions after gaps and score self-blame/urgency markers (absent = relaxed); count low-pressure resumptions per session and record gap length; inter-rater agreement would be assessed via Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1021 — Low-Pressure Encouragement
also known as: Energy Kiss
description: A small, friendly affirmation from the AI that the user reports as a brief lift in energy or motivation, without creating performance pressure or implying an obligation to perform.
operationalDefinition: EVENT. One occurrence is coded when a brief affirmative AI turn is followed by a user-reported small rise in energy/motivation, with no accompanying increase in felt performance pressure. The reported energy rise tied to the affirmation is the countable marker.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Paired single-item probes around the affirmation: user rates momentary energy/vigor (1-7) and felt performance pressure (1-7) before and after; the event requires an energy increase without a pressure increase; report the joint pre-post change pattern across instances.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1024 — Optimism Opener
description: A reframing that shifts the user's perspective toward opportunities and possible courses of action without denying or minimizing the existing problems.
operationalDefinition: EVENT. One occurrence is coded when, after an exchange, the user's framing of a situation shifts from problem-focused to opportunity/action-focused while still acknowledging the problem (problems are not denied). The user turn exhibiting the opportunity-oriented reframing is the countable marker.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Pre/post framing classification: raters code the user's stance on the focal issue before and after as problem-focused vs. opportunity/action-focused (with a check that problems remain acknowledged); the event is a problem->opportunity shift; report shift frequency and inter-rater agreement would be assessed via Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1039 — Frustration-to-Openness Shift
also known as: Light Swing
description: A mental turn from frustration toward constructive openness, in which a user moves from a blocked, irritated stance to a receptive, solution-oriented one during an AI exchange.
operationalDefinition: EVENT. One occurrence is coded when a turn expressing frustration is followed within the same exchange by a turn showing constructive, open engagement (proposing or accepting a way forward). The transition into the constructive turn is the countable marker.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Stance-transition coding: raters label turns on a frustration-to-openness dimension and mark transitions from frustrated to constructively open; count qualifying transitions per session; inter-rater agreement would be assessed via Cohen's kappa on stance labels.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1048 — Soft Landing
description: The scaling-down of overreaching ideas into realistic, implementable steps, in which an AI exchange helps a user convert an oversized aim into a feasible sequence of actions.
operationalDefinition: EVENT. One occurrence is coded when an initially overscoped goal stated by the user is, after the exchange, reformulated by the user into a smaller set of concrete, feasible steps. The user turn presenting the down-scoped, actionable plan is the countable marker.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Scope-change coding: raters record the initial scope (e.g., number/size of claimed deliverables) and the post-exchange scope (concrete next steps), and judge whether the reformulation is more feasible; report the proportion of overscoped goals that are down-scoped and inter-rater agreement would be assessed via Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1049 — Tension-to-Ease Shift
also known as: Grin Shift
description: A transition from tension to ease brought about by small positive cues in an AI exchange, in which the user's strained state gives way to a lighter, more relaxed one.
operationalDefinition: EVENT. One occurrence is coded when a turn showing tension or strain is followed, after a small positive cue from the AI, by a turn showing reduced strain and greater ease. The transition into the eased turn is the countable marker.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Tension-to-ease coding: raters rate per-turn strain/ease (e.g., tense vs. relaxed expression markers) and mark transitions from strained to eased following a positive cue; count transitions per session; inter-rater agreement would be assessed via weighted Cohen's kappa on the ordered strain ratings.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-1062 — Lightness Lift
description: The targeted reduction of a user's felt heaviness through linguistic or perspective-based relief, in which an AI exchange deliberately reframes or lightens the framing of a burdensome matter.
operationalDefinition: EVENT. One occurrence is coded when an AI exchange applies a deliberate relief move (reframing, lightening language, perspective change) and the user subsequently reports reduced felt heaviness/burden about the focal matter. The user's reduced-burden report following the relief move is the countable marker.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Pre/post burden rating: user rates felt heaviness/burden about the focal matter (1-7) before and after the relief move; the event requires a meaningful decrease; report mean within-person burden reduction with a paired test, paired with rater identification of the relief move (kappa for the move classification).
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0089 — Pattern Sharpening
also known as: The Pattern Sharpening
description: A longitudinal change in a user's habitual reasoning patterns associated with sustained, regular AI use, observed as drift in how problems are approached across repeated sessions rather than at any single moment.
operationalDefinition: Tracked as change over time in a fixed cognitive-task metric (for example solution-path structure or decomposition style) measured at regular intervals across a defined usage period; the trend is that metric's slope.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Longitudinal tracking: a standardized reasoning task administered at fixed intervals over weeks; the outcome is the within-person trajectory (slope) of the task metric, modeled across sessions.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0422 — Calibrated Trust
also known as: The Calibrated Trust
description: A graded user disposition in which the degree of acceptance granted to an AI output is scaled to its stakes: high-consequence outputs receive more independent verification, routine outputs receive proportionally lighter checking. Trust varies by context rather than being uniformly extended or withheld.
operationalDefinition: STATE. Two trained raters score, for each AI output a user acts upon, (a) the stakes of the output on a 3-point scale (low/medium/high consequence of error) and (b) the observed verification effort the user applied before acting (none / light cross-check / independent confirmation). Calibrated trust is operationalised as the rank-order alignment between stakes and verification effort across a session; higher alignment indicates greater calibration.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater-coded verification-effort vs. stakes alignment per acted-upon output; calibration quantified as the Goodman-Kruskal gamma rank correlation between the stakes scale and the effort scale within a participant. Inter-rater reliability for both coded dimensions could be reported as weighted Cohen's kappa on a held-out 20% double-coded subset; participant-level self-report supplement via a 5-point Likert item ('I checked this more carefully because it mattered more').
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0614 — Synthetic Spotting
also known as: The Synthetic Spotting
description: A discrete user judgement that a given piece of content was produced by an AI system, inferred from stylistic regularities, characteristic phrasing, or content-level cues. The referent is the act of attribution, not the correctness of that attribution.
operationalDefinition: EVENT. One occurrence is recorded each time a user explicitly attributes a presented artefact (text, image, or audio) to AI generation. The observable marker is a verbal or logged attribution judgement ('this is AI-generated'). When ground truth of the artefact is known, each judgement is scored hit / miss / false alarm / correct rejection.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Signal-detection task: participants classify a balanced stimulus set of AI-generated and human-authored artefacts. Sensitivity could be reported as d-prime and response bias as criterion c, computed from hit and false-alarm rates; per-stimulus confidence captured for an ROC curve. Accuracy contrasted against a 50% chance baseline with a binomial test.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0775 — AI Abstention Zone
also known as: The KI-Free Zone
description: A deliberately maintained boundary - spatial, temporal, or topical - within which a person excludes AI assistance. The referent is a standing, self-imposed restriction on the scope of AI use rather than a one-off refusal.
operationalDefinition: STATE. Coded as present when a participant names at least one explicit, durable exclusion boundary and the conditions under which it applies (e.g. 'no AI during family meals', 'no AI for first-draft poetry', 'no AI in the bedroom'). Each declared boundary is catalogued by type (spatial / temporal / topical) and by stated scope; the state is graded by the count and breadth of maintained boundaries.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Structured boundary inventory: a semi-structured interview enumerates declared exclusion zones, each verified for durability via a one-week experience-sampling log of adherence (proportion of qualifying occasions on which AI was in fact withheld). Could be reported as number of zones, zone-type distribution, and mean adherence rate; coder agreement on zone-type classification via Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0260 — Narrative Divergence Prompt
also known as: The Plot Twist Partner
description: A use episode in which a person enlists an AI system to generate unexpected narrative turns or alternative storylines for a creative or persuasive artefact such as a story, pitch, or argument. The referent is the request-and-use episode, not a property of the AI.
operationalDefinition: EVENT. One occurrence is logged when a user issues a request whose explicit aim is to obtain a divergent or unexpected continuation of an existing narrative or argument (e.g. 'give me a surprising twist', 'what is an unexpected counter-move'). The observable marker is a divergence-seeking instruction directed at narrative or argumentative material.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Task-log content coding: prompts in a creative-writing corpus are labelled for divergence-seeking intent by two raters; rate could be reported as divergence requests per 100 task turns. Downstream uptake measured as the proportion of AI-proposed twists incorporated into the user's subsequent draft (acceptance rate); coder reliability via Cohen's kappa on the intent label.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0370 — Animal-Care Query
also known as: The Pet Lookup
description: A use episode in which a person queries an AI system for information about animal care - feeding, behaviour, preventive health, or species-appropriate husbandry. The referent is a topically defined everyday information request.
operationalDefinition: EVENT. One occurrence is recorded when a user query is classified as seeking animal-care information, identified by topical markers (species terms together with care intents such as feeding, health signs, behaviour, or housing). Each qualifying query counts as one instance.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Query-topic classification: a held-out sample of user queries is labelled against an animal-care topic taxonomy by two annotators; prevalence could be reported as the proportion of queries in this topic class and as queries per active user per month. Annotation reliability via Cohen's kappa; optional automated classifier validated against the human labels (precision/recall, F1).
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0430 — Workflow Reversion
also known as: The Vintage Loop
description: A reversion episode in which a person returns to a previously used AI workflow after a newer alternative did not deliver the intended result. The referent is the act of reverting, reflecting that an established procedure outperformed a newer one for that task.
operationalDefinition: EVENT. One occurrence is recorded when a user, having adopted a newer tool, model, or procedure for a recurring task, switches back to an earlier one they had previously used for the same task. The observable marker is a documented workflow transition from a newer to an older procedure for an unchanged task type.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Workflow-version transition log: for a panel of users, tool/model/procedure choices per recurring task are timestamped, yielding a directed sequence of transitions. Reversions are counted as backward transitions to a prior procedure; could be reported as reversion rate per adopted change and median tenure of the new procedure before reversion (a median time-to-event, survival-style, analysis).
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0534 — Novel-Perspective Surfacing
also known as: The Hidden Angle Finder
description: A use episode in which an AI system surfaces a perspective on a familiar topic that the user reports not having considered before. The referent is the discovery event of a previously unconsidered viewpoint during interaction.
operationalDefinition: EVENT. One occurrence is recorded when (a) an AI output introduces a perspective absent from the user's prior framing of the topic and (b) the user acknowledges it as new to them. The observable markers are a rater-identified perspective shift in the output plus a user confirmation of novelty.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Two-source corroboration: trained raters code each output for introduction of a perspective not present in the preceding user turns (presence/absence), and users rate each flagged output on a 5-point novelty item ('this viewpoint had not occurred to me'). An instance requires both signals; could be reported as novel-perspective episodes per session. Rater reliability via Cohen's kappa; convergence of rater flag and user rating via phi coefficient.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0639 — Duplicate Notice
also known as: The Duplicate Notice
description: A discrete recognition act in which a user identifies an AI response as substantively repeating an earlier response - reworded but identical in core content. The referent is the user's recognition of redundancy, not the redundancy itself.
operationalDefinition: EVENT. One occurrence is recorded when a user signals that a current response duplicates the substance of a prior one in the same conversation (e.g. an explicit complaint, a 'you already said this' turn, or a logged flag). The observable marker is a user-issued redundancy signal referring to an earlier turn.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Redundancy-flag detection corroborated by semantic similarity: user redundancy signals are extracted from transcripts; for each, the cosine similarity between the flagged response and the referenced earlier response would be computed from sentence embeddings to confirm genuine near-duplication above a calibrated threshold. Could be reported as redundancy notices per 100 turns and precision of user flags against the similarity criterion; human coding of flags reliability via Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0675 — Role-Aware Input
also known as: The Role-Aware Input
description: A prompt in which the user frames the request through an explicit social role - for example 'as a parent I need...' or 'in my capacity as...'. The referent is the role-marked formulation of an individual input.
operationalDefinition: EVENT. One occurrence is recorded when a prompt contains an explicit self-referential role marker that contextualises the request (a stated role noun phrase tied to the user, such as parent, teacher, manager, caregiver). The observable marker is the presence of such a role-framing clause in the prompt text.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Prompt-text annotation: a corpus of prompts would be coded for presence and type of explicit user-role markers by two annotators; prevalence could be reported as proportion of role-marked prompts and rate per 100 prompts, with a role-category distribution. Annotator agreement would be assessed via Cohen's kappa; optional regex/NER detector validated against human labels (precision/recall, F1).
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0711 — Accent Persistence
also known as: The Accent Persistence
description: The degree to which structural features of a user's first language - word order, punctuation conventions, and phrasing patterns - remain detectable in their written AI inputs composed in a second language. The referent is the relative strength of cross-linguistic transfer in writing.
operationalDefinition: RATIO. Quantified as the relative frequency of first-language-typical structural features in a user's second-language prompts compared with a same-language native-writer baseline. Numerator: count of L1-transfer features observed in the user's prompts (e.g. divergent word-order patterns, calqued collocations, L1 punctuation conventions); denominator: total scored constructions, normalised against the native-baseline feature rate.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Contrastive linguistic feature scoring: prompts are parsed and tagged for a predefined inventory of L1-transfer markers; a transfer index would be computed as the standardised difference between the user's marker rate and a matched native-writer corpus baseline. Could be reported as effect size (Cohen's d) of marker rate against baseline; classifier predicting writer L1 from prompt features evaluated by AUC. Tagging reliability via Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0901 — Emergent Coordination
also known as: The Emergent Coordination
description: An occurrence in a multi-agent system of a coordination pattern that was not explicitly specified, arising from the interaction of components rather than from any single component's design. The referent is the observed unspecified coordinated behaviour.
operationalDefinition: EVENT. One occurrence is recorded when system logs show a stable coordinated behaviour among agents (e.g. role division, turn-taking, convergence on a shared plan) that is not entailed by any individual agent's specification or by explicit coordination rules. The observable marker is a recurring inter-agent pattern absent from the design specification.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Specification-versus-behaviour log analysis: interaction traces are mined for recurrent multi-agent patterns; each candidate is checked against the documented specification and ablated to confirm it disappears when interaction is removed. Could be reported as count of specification-absent coordinated patterns per run and their stability (recurrence rate across seeds); two analysts independently adjudicate 'unspecified' status with agreement would be assessed via Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0001 — Augmanitai
description: A superordinate framework concept describing deliberate, human-directed collaboration between people and AI systems, in which augmentation of human capability is paired with explicit human oversight; it serves as the umbrella term under which the corpus's other constructs are organized.
operationalDefinition: Superordinate framework term; not operationalized as a per-interaction event. It denotes the organizing concept for the corpus rather than a discrete, countable phenomenon, and is instantiated only indirectly through the more specific constructs it subsumes.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Not directly measured; superordinate term.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0005 — Integrated Operator
description: A characterisation of a user's working style marked by fluid interleaving of AI assistance with their own cognitive and work processes. It is a graded usage profile attributed to a person, not a discrete event.
operationalDefinition: STATE. Rated, not counted: a user's working style is scored on a fluency-of-integration scale derived from behavioural indicators (switch frequency between AI-assisted and unassisted work, low hesitation latency). Attribution above a pre-registered profile cut-off.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Composite usage-style index from log indicators (AI-to-manual switch rate, inter-action latency) plus a self-report integration subscale; standardised and combined; profile reliability via internal consistency (Cronbach's alpha).
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0016 — Cross-Domain Synthesis
also known as: Poly-Categorical Mesh
description: An observable instance in which AI-assisted work links concepts from two or more distinct disciplines into a single output that no source domain supplied alone, yielding a cross-disciplinary result.
operationalDefinition: One occurrence is coded when an AI-supported output explicitly integrates content from at least two identifiable domains into one connected claim or artifact; the integrating passage is the observable marker.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater coding of cross-domain integration: trained coders tag outputs that integrate two or more named domains, counted per session, with inter-rater reliability could be reported as Cohen's kappa on a double-coded sample.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0047 — Post-Affirmation Advocacy
also known as: The Echo Courage
description: An episode in which a user, after an affirming AI interaction, becomes more willing to advance an idea to other people. The referent is the subsequent real-world advocacy act, with the prior AI exchange serving as an initial sounding board.
operationalDefinition: EVENT. One occurrence is recorded when a user reports or is observed to advance an idea to a human audience (raising it in a meeting, sending it, defending it) after having first tested that idea in an affirming AI exchange. The observable marker is a downstream advocacy action temporally preceded by a confirming AI interaction on the same idea.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Paired before/after self-report with optional behavioural follow-up: participants rate intention to advocate an idea before and after an AI exchange on a 5-point scale, and report actual advocacy at a later follow-up. Effect estimated as the within-subject change in advocacy intention (paired t-test or Wilcoxon signed-rank) and the conversion rate from intention to reported action; an AI-affirmation versus neutral-exchange contrast controls for mere rehearsal.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0092 — Output Asymmetry
description: The observed relationship between the effort a user invests as input and the volume of output obtained in AI-assisted work. It is a quantitative input-to-output ratio characterising leverage.
operationalDefinition: RATIO. Numerator: a sized measure of produced output (e.g. words, components, deliverables). Denominator: a sized measure of user input effort (e.g. prompt tokens authored, active minutes). Reported as output units per unit of input effort.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Instrumented logging of input effort (authored tokens; active edit time) and output size (produced tokens/artefacts) per task; ratio reported with dispersion across tasks. Distinct from AUG-0091 (which compares conditions, not input vs. output).
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0999 — Forward Assessment
also known as: The Forward Assessment
description: An utterance that attempts to anticipate how the human-AI relationship will develop while explicitly acknowledging the uncertainty of any such forecast. The referent is the hedged, future-oriented assessment act; the surrounding lexicon documents the present rather than predicting it.
operationalDefinition: EVENT. One occurrence is recorded when a discourse segment makes an explicit claim about the future state of human-AI interaction accompanied by an epistemic hedge marking its uncertainty (e.g. 'may', 'could', 'it is hard to say'). The observable markers are co-present future-time reference and uncertainty hedging within the same segment.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Discourse coding for hedged future-oriented statements: text segments are annotated for future-time reference and for epistemic-hedge markers; an instance requires both. Could be reported as the rate of hedged forecasts per 1000 words and the proportion of future-claims that carry a hedge (a calibration-of-uncertainty indicator). Annotation reliability via Cohen's kappa on a double-coded subset.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0041 — Tangential-Output Serendipity
also known as: The Scatter Spark
description: An episode in which an off-target or tangential AI output - content not directly matching the query - triggers an unanticipated association or idea connection in the user. The referent is the serendipitous association event arising from output 'noise'.
operationalDefinition: EVENT. One occurrence is recorded when (a) an AI output segment is judged off-target relative to the user's request and (b) the user reports that this segment prompted a new association or idea. The observable markers are a rater-identified off-target segment plus a user-reported idea triggered by it.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Serendipity-episode corroboration: raters label output segments as on- or off-target to the query; users flag segments that sparked a new idea and rate its usefulness. A serendipity instance requires an off-target segment with a user spark-flag; could be reported as serendipitous-association episodes per session and their rated idea value. Rater on/off-target reliability via Cohen's kappa; association of off-target status with spark-flags via phi.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0194 — Positive Surprise
also known as: The Positive Surprise
description: An AI response that exceeds the user's prior expectations in quality, depth, or perspective. The referent is the discrete expectation-exceeding response event, which is associated with an increase in the user's trust in the collaboration.
operationalDefinition: EVENT. One occurrence is recorded when a user rates a specific response as exceeding the expectation they held before receiving it. The observable marker is a positive expectation-disconfirmation judgement attached to an individual response (post-rating higher than the pre-stated or concurrently reported expectation).
measurementSchema: Proposed measurement protocol (not yet empirically validated): Expectation-disconfirmation rating per response: users state an expectation level before a response and rate the delivered response on the same 5-point scale, yielding a signed disconfirmation score; positive surprises are responses with a positive score beyond a set margin. Could be reported as the rate of positive-surprise responses and mean disconfirmation; an optional paired trust item before and after gauges the associated trust shift (Wilcoxon signed-rank).
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0220 — Differential Gratitude Expression
also known as: The Gratitude Paradox
description: The pattern in which some users express gratitude toward an AI more freely than toward people, attributed to the absence of social-evaluation concern in the AI exchange. The referent is the relative difference in expressed gratitude between AI-directed and human-directed communication.
operationalDefinition: RATIO. Numerator: frequency or rated intensity of gratitude expressions a user directs at an AI system. Denominator: the same measure for the user's human-directed communication over a comparable period. The paradox is indicated when the AI-directed value meaningfully exceeds the human-directed value for a user.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Comparative gratitude-expression measurement: gratitude acts are detected and intensity-rated in matched AI-directed and human-directed message samples per participant (coded by raters, optionally aided by a validated gratitude-lexicon classifier). Could be reported as the within-person AI-to-human gratitude ratio and tested against parity (ratio of 1) with a paired comparison; coder intensity-rating reliability via ICC and detection reliability via Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0305 — Seasonal User
also known as: The Seasonal User
description: A user whose AI-use intensity rises and falls over time with life phase, project load, or season, alternating intensive daily periods with extended pauses. The referent is the longitudinal, non-linear pattern of usage intensity for an individual.
operationalDefinition: TREND. Tracked as the variability of an individual's AI-use intensity (e.g. sessions or active minutes per week) across a longitudinal window of at least several months. A user is classified seasonal when usage shows recurring high-low cycling exceeding a defined dispersion threshold rather than a stable or monotonic trajectory.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Longitudinal usage-intensity time series: weekly usage is logged over a multi-month window; intra-individual dispersion is summarised by the coefficient of variation and burstiness, and any periodicity by autocorrelation or spectral analysis. A user is flagged seasonal above a pre-registered coefficient-of-variation cut-off; reported with the distribution of dispersion scores across the panel.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0330 — Source Curiosity
also known as: The Source Curiosity
description: A user's move to trace the origin or grounding of an AI-generated statement - asking where a claim comes from or on what it is based. The referent is the discrete provenance-seeking act during interaction.
operationalDefinition: EVENT. One occurrence is recorded when a user issues a request for the source, basis, or evidence behind an AI statement (e.g. 'what is the source', 'how do you know that', 'cite that'). The observable marker is a provenance- or evidence-seeking turn directed at a preceding AI claim.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Provenance-request coding: conversation turns are annotated for source/evidence-seeking moves by two coders; could be reported as provenance requests per 100 turns and the proportion of factual AI claims that are followed by such a request (a verification-engagement rate). Coder agreement would be assessed via Cohen's kappa; optional intent classifier validated against labels (precision/recall, F1).
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0435 — Meal-Planning Query
also known as: The Dinner Shortcut
description: A brief AI query made for an everyday food decision, such as what to cook from available ingredients or what side dish fits a meal. The referent is a low-effort, topically defined household information request.
operationalDefinition: EVENT. One occurrence is recorded when a user query is classified as an everyday meal-planning request, identified by markers combining food or ingredient terms with a decision intent (what to cook, what pairs with, substitution). Each qualifying query counts as one instance.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Query-topic classification: a sample of queries is labelled against a meal-planning intent taxonomy by two annotators; prevalence could be reported as the share of queries in this class and as queries per active user per week, with query-length statistics characterising the 'low-threshold' nature. Annotation reliability via Cohen's kappa; automated classifier validated against human labels (precision/recall, F1).
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0520 — Inquiry Chain Cascade
also known as: The Wiki Wormhole
description: An exploration episode in which one AI question leads to a chain of follow-up questions that draws the user progressively deeper into a topic, analogous to a 'Wikipedia hole' but driven by AI dialogue. The referent is the extended, self-propagating inquiry episode.
operationalDefinition: EVENT. One occurrence is recorded when a single seed question develops into an uninterrupted chain of topically linked follow-up turns exceeding a set depth threshold (e.g. five or more consecutive on-topic follow-ups) within one session. The observable marker is a contiguous follow-up chain of qualifying depth descending from one initial query.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Session follow-up-chain analysis: dialogue turns are segmented into topic-linked chains; an episode would be coded when chain depth crosses the pre-set threshold. Could be reported as chains-per-session, mean and maximum chain depth, and chain dwell time; topic-linkage decisions corroborated by inter-turn semantic similarity and double-coded for reliability via Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0565 — Balance Filter
also known as: The Balance Filter
description: A standing self-regulation strategy for keeping AI use proportionate to non-digital activity through fixed times, rules, or routines, so that AI supplements rather than overtakes daily life. The referent is the maintained regulatory arrangement, not a single instance of restraint.
operationalDefinition: STATE. Coded as present when a participant maintains one or more explicit, durable rules or routines that bound AI use relative to non-digital activity (e.g. fixed AI-free hours, capped daily sessions, mandatory offline routines). The state is graded by the number of active rules and the degree of adherence to them.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Self-regulation inventory with adherence tracking: declared balancing rules are enumerated in interview, then adherence would be measured over a one- to two-week diary or experience-sampling period as the proportion of qualifying occasions on which each rule was kept. Could be reported as number of active rules and mean adherence rate; an optional validated self-control or digital-balance scale (Likert) provides a convergent dispositional measure. Coder rule-type agreement would be assessed via Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0981 — Companion Pattern
description: The condition in which a user has settled into stable, recurring workflows with an AI system, observable as repeated prompt structures, preferred task types, and consistent sequencing across sessions. Describes regularity of use, not any bond with the system.
operationalDefinition: STATE. Assessed from logs over a window by the stability of the user's prompt templates and task-type distribution: high recurrence of similar prompt structures and a concentrated task mix indicate the pattern is present. The construct is the degree of routine consolidation.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Routine-consolidation index from logs: proportion of sessions matching a recurring prompt template (clustered by edit distance) plus task-type entropy (lower = more consolidated); reported over >=15 sessions.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

## WAVE 2 — 100 PERIODIC-TABLE TERMS

### AUG-0004 — Zero-Point Self
description: A person's documented baseline of knowledge or skill in a domain prior to any AI assistance, used as the individual reference point against which later AI-supported performance is compared.
operationalDefinition: A criterion-referenced competence assessment in the target domain administered to the user before their first AI-assisted attempt; the recorded score is the reference value, not a counted occurrence.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Pre-exposure baseline: criterion-referenced domain test (fixed item bank) scored as percent-correct, administered once before first AI use, then stored as the per-person reference value for later difference scores.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0054 — Augmented Understanding
description: An observable instance in which AI-assisted retrieval gives a user access to information they could not obtain unaided, evidenced by a measurable gain in comprehension or recall on the topic at hand.
operationalDefinition: One occurrence is coded when, within a task, the user obtains via the AI a specific fact or explanation absent from their prior statements and then correctly applies or restates it; the applied item is the observable marker.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Pre/post comprehension delta: a short topic probe (3-5 items) given before and after the AI exchange; an occurrence counts when the post-score exceeds the pre-score on items covered by the retrieved content.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0002 — Cognitive Offloading Turn
also known as: Mentale Externalisierung
description: A turn in which a user deliberately hands a discrete cognitive sub-task, such as enumeration, drafting, or calculation, to an AI system to reserve attention for other reasoning, observable as an explicit delegation request.
operationalDefinition: One occurrence is coded per user turn whose primary act is delegating a bounded sub-task to the AI (an imperative request to produce, compute, or list something the user could do themselves), excluding clarification or evaluation turns.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater-coded delegation count per 10 user turns using a turn-level codebook (delegation vs. clarification vs. evaluation); inter-rater agreement could be reported as Cohen's kappa on a double-coded subset.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0007 — Blending Effect
also known as: The Blending Effect
description: A graded condition in which a user can no longer cleanly separate their own contributions from AI-generated material in a co-produced artifact; an attribution phenomenon about authorship boundaries, not a loss of ability.
operationalDefinition: Rated as the difficulty of source attribution for a co-produced text: the user or an independent rater labels each passage as human- or AI-originated, and lower attribution accuracy indicates a stronger condition.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Source-attribution task: passages of a co-authored artifact are labeled human/AI by the participant and scored as attribution accuracy (percent correct) against an edit-log ground truth, reported per session.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0035 — Information Currency Duration
also known as: Epistemic Half-Life
description: The elapsed time over which a specific AI-generated factual claim stays accurate before external developments render it outdated; a duration describing how quickly such information loses currency.
operationalDefinition: For a sampled factual claim, the duration in days from generation until a verifiable external update first contradicts or supersedes it; the numerator is elapsed time, anchored to dated ground-truth sources.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Time-to-supersession: a sample of dated generated claims is re-checked against authoritative sources at fixed intervals; could be reported as median days-until-superseded with interquartile range.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0043 — Just-in-Time Competence
description: An observable instance of accessing domain knowledge at the exact point of need through AI support, where the retrieved information immediately precedes and enables completion of the user's current task step.
operationalDefinition: One occurrence is coded when a point-of-need query returns information the user applies in the immediately following action to complete a task step; the successful task step is the observable marker.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Task-contingent retrieval count plus latency: number of point-of-need retrievals per task, and time from query to the first correct task action in seconds, both derived from interaction timestamps.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0088 — Algorithmic Intuition
description: A graded user skill, observed in experienced operators, of rapidly recognizing recurring AI-interaction patterns and selecting effective prompting strategies with little deliberation, distinguishing them from novices.
operationalDefinition: Operationalized via the speed and success with which a user formulates an effective prompt for a familiar task type, rated as pattern-recognition proficiency and compared across experience levels.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Between-groups comparison: median prompt-formulation latency (seconds to first effective prompt) and task-success rate for experienced vs. novice users on matched tasks; the group difference could be reported with an effect size.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0016 — Cross-Domain Synthesis
also known as: Poly-Categorical Mesh
description: An output in which an AI-assisted user integrates knowledge from several distinct disciplines into a single interdisciplinary result. The referent is an instance of cross-domain synthesis produced with on-demand access to domain knowledge.
operationalDefinition: EVENT. One occurrence is recorded when a user-produced output combines content from two or more distinct knowledge domains into a connected result. The observable marker is a single artefact that draws identifiable, separable contributions from multiple disciplines and links them rather than juxtaposing them.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater coding of cross-domain integration: each output would be annotated for the number of distinct source domains present and the depth of integration on an ordinal scale (juxtaposition / partial linkage / synthesised novelty). Could be reported as the count of multi-domain syntheses per output set and the domain-count distribution; inter-rater reliability on the integration-depth scale via intraclass correlation (ICC) and on domain identification via Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0051 — Depth of Provenance
description: A graded user competence in evaluating the origin, evidential basis, and reliability of AI-generated knowledge and placing it appropriately, distinguishing well-sourced claims from unsupported ones.
operationalDefinition: Rated on a source-critical reasoning rubric: the user is scored on whether they question provenance, request or check sources, and calibrate confidence to evidence quality across a set of AI responses.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rubric scoring by trained raters on a 0-4 source-criticality scale across fixed probe items; agreement could be reported as the intraclass correlation (ICC) for averaged ratings.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0089 — Pattern Sharpening
also known as: The Pattern Sharpening
description: A longitudinal effect in which regular AI interaction is associated with improvement in a user's ability to recognize patterns in data, texts, or arguments, such that sustained use coincides with gains in the person's own analytic pattern detection.
operationalDefinition: TREND. Tracked over repeated measurements as the change in pattern-recognition performance: administer a standardized pattern-detection task at multiple timepoints over an AI-use period and track the change in accuracy/speed across the window; the trend (improvement trajectory) is the referent, not a single in-session event.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Repeated standardized assessment: a validated pattern/inductive-reasoning instrument (e.g., matrix-reasoning or series-completion items) administered at baseline and follow-ups; track accuracy and response time over time; analyze the trajectory (e.g., slope) ideally against a comparison condition to support the association rather than infer causation.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0205 — Skill Unlock
also known as: The Skill Unlock
description: An observable instance of completing, with AI support, a task belonging to a class the user could not previously accomplish unaided, marking newly reachable capability.
operationalDefinition: One occurrence is coded at the first successful completion of a task from a class the user had previously not completed or never attempted unaided, verified against a record of prior unaided attempts.
measurementSchema: Proposed measurement protocol (not yet empirically validated): First-completion event count: the number of previously-unattainable task classes the user first completes with AI support, indexed against a pre-recorded baseline of failed or unattempted tasks.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0098 — Thinking Leverage
description: The relationship between the cognitive effort a user invests and the quality or quantity of output achieved with AI support; a ratio expressing output yield per unit of human input.
operationalDefinition: Computed as rated output quality, or completed work units, divided by invested effort measured in active minutes or user turns for a task; the numerator is the work product and the denominator is human effort expended.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Effort-normalized yield: rater-scored output quality (0-100) divided by active user time in minutes (or by user-turn count) per task, could be reported as a yield ratio together with its components.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0156 — Articulation Unlock
description: An observable moment in which AI support helps a user put a pre-existing but unformed thought into words, evidenced by the user accepting a formulation that expresses an intent they had already signaled.
operationalDefinition: One occurrence is coded when the user adopts, verbatim or lightly edited, an AI-offered phrasing for an idea the user had gestured at earlier in the session and confirms it matches their meaning.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Acceptance count plus self-report: the number of articulation-assist turns accepted per session, paired with a single-item agreement rating ('this captured what I meant', 1-7 Likert) recorded at each acceptance.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0403 — Translation Relief
also known as: The Translation Relief
description: A graded affective response of relief experienced by a user when AI support removes a language barrier and eases translation effort, reported by the user rather than inferred from the AI.
operationalDefinition: Rated as self-reported relief intensity at points where the AI bridged a language gap, for example by producing or clarifying a translation; the construct is the user's affect, not a property of the AI.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Self-report Likert: a 1-7 relief item anchored 'no relief' to 'strong relief', completed by the user immediately after AI-mediated translation events and aggregated as mean relief per session.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0513 — Conversational Language-Practice Tool
also known as: The Language Buddy
description: A superordinate description of using an AI system as a repeatable language-practice resource that supplies correction and explanation alongside formal language instruction; a usage role rather than a single observable event.
operationalDefinition: Superordinate framework term; not operationalized as a per-interaction event. It names a category of supportive language-learning use under which more specific, measurable events fall.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Not directly measured; superordinate term.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0478 — Clarity Catalyst
also known as: The Clarity Catalyst
description: An observable instance in which the demand to state a request explicitly to an AI prompts the user to reorganize their own understanding of a problem, changing the thinking process through the act of formulation.
operationalDefinition: One occurrence is coded when, between an initial vague query and its explicit reformulation for the AI, the user's stated problem framing measurably changes by adding constraints or sharpening the goal; the reformulation pair is the marker.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Pre/post articulation-clarity rating: independent raters score the user's problem statement before and after explicit formulation on a 1-5 clarity rubric, and an occurrence counts when the post score exceeds the pre score.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0188 — Tone Alignment
description: A longitudinal convergence between a user's communication style and the AI's responses over repeated exchanges, in which the pair settles into a shared workable register; a stylistic adaptation, not an emotional state of the AI.
operationalDefinition: Tracked as decreasing stylistic distance between user and AI turns across a session series, measured on linguistic style features; the trend is the downward trajectory of that distance over time.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Stylometric convergence: per-session linguistic distance between user and AI turns (function-word and register feature vectors, cosine or Euclidean) tracked over sessions, with the distance trajectory as the outcome.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0169 — Second-Language Fluency
description: A graded condition of more successful foreign-language communication achieved with AI support, observed as eased production and comprehension in a second language during AI-assisted exchanges.
operationalDefinition: Rated as communicative success in the second language with AI support, namely task completion and intelligibility on a defined L2 communication task, compared against the same task performed without AI assistance.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Within-subject contrast: rater-judged L2 communicative success (task completion plus intelligibility, 0-5) on matched tasks with vs. without AI, could be reported as the with-AI minus without-AI difference.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0137 — Voice-First Protocol
description: A superordinate category for speech-based AI interaction treated as a distinct modality with its own affordances and dynamics, under which specific measurable voice-interaction phenomena are grouped.
operationalDefinition: Superordinate framework term; not operationalized as a per-interaction event. It designates an interaction modality of spoken input and output rather than a single countable occurrence.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Not directly measured; superordinate term.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0267 — Language Unlock
also known as: The Language Unlock
description: An observable expansion of a user's active language repertoire through AI-assisted interaction, evidenced when a lexical or expressive item first encountered via the AI later appears in the user's own unaided production.
operationalDefinition: One occurrence is coded when a word or expression introduced by the AI subsequently appears in the user's independent, non-AI output; the later unaided use is the observable marker, distinguishing uptake from mere exposure.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Uptake count: the number of AI-introduced lexical or expressive items that recur in the user's subsequent unaided writing within a follow-up window, tracked by term matching against the interaction log.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0501 — Style Shifter
also known as: The Style Shifter
description: A graded user ability to deliberately switch among distinct writing styles with AI assistance and to access registers the user could not readily produce alone.
operationalDefinition: Rated as the distinctiveness of styles a user can elicit on demand: outputs produced for several requested registers are judged for how reliably they differ from one another and match the target register.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Style-discriminability: independent raters, or a held-out classifier, attempt to assign AI-assisted outputs to their requested register; the score would be classification accuracy or pairwise distinctiveness across registers.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0283 — Syntax Voice
also known as: The Syntax Voice
description: The gradual incorporation of new sentence constructions and expressions, first encountered through AI interaction, into a user's own habitual syntactic repertoire over time.
operationalDefinition: Tracked as the changing frequency of AI-originated syntactic structures in the user's unaided writing across successive periods; the trend is the rise of those structures in independent production.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Longitudinal syntactic frequency: the rate of targeted AI-originated constructions per 1000 words in the user's unaided writing, sampled at intervals, with the within-person frequency trajectory as the outcome.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0046 — Post-Session Affective Residue
also known as: The Felt Echo
description: A residual affective and cognitive engagement persisting after an intense AI session, in which the user continues to feel the exchange and keeps processing its content for some time afterward; a post-session state of the user.
operationalDefinition: Rated as the intensity of residual engagement the user reports at a fixed delay after an intensive session, for example 30 minutes, capturing continued rumination on and emotional resonance of the exchange.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Delayed self-report: a short residual-engagement scale (about 4 items on continued thinking and emotional resonance, 1-7 Likert) administered at a fixed post-session delay and aggregated as a mean residue score.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0025 — Offload Lift
also known as: The Offload Lift
description: A graded affective response of relief or reduced burden experienced by a user when handing an effortful task to an AI system, reported as a felt easing of workload at the point of delegation.
operationalDefinition: Rated as self-reported reduction in perceived burden at delegation points, capturing the user's relief when an effortful task is offloaded; the construct is the user's affect, not a capability of the AI.
measurementSchema: Proposed measurement protocol (not yet empirically validated): A brief perceived-workload/effort self-report (effort and frustration items), or a 1-7 relief item, completed immediately after offloading an effortful task, with the pre-minus-post perceived-load change as the outcome.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0127 — Expansion Feeling
description: A subjective experience reported during intensive AI use in which a user perceives their own thinking boundaries as widening; a phenomenal state of the user, not a claim about expanded capacity in the AI.
operationalDefinition: Rated as the strength of the user's felt sense of cognitive expansion during or right after intensive sessions, captured by direct self-report of perceived broadening of one's own thinking.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Phenomenological self-report: a perceived-cognitive-expansion scale (3-5 items, 1-7 Likert anchored 'not at all' to 'very much') completed post-session, could be reported as a mean rating with internal-consistency reliability.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0168 — Non-AI Activity Transition
also known as: Rehumanization Moment
description: An observable moment of deliberately returning to a non-AI activity after a period of intensive AI-assisted work, marking an intentional disengagement from the AI for a human-only task.
operationalDefinition: One occurrence is coded at an explicit transition where the user disengages the AI to resume an unaided activity after intensive use; the marked transition, verbalized or a logged tool-off, is the observable marker.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Transition count and timing: the number of deliberate AI-disengagement events per session and their latency from session end, identified from interaction logs or user diary marks.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0301 — Capability-Surprise Reaction
also known as: The Wonder Moment
description: A discrete user reaction of surprise or astonishment occurring when an AI system produces output that exceeds the user's prior expectation of its capability. The reaction is a property of the human appraisal, not of the system.
operationalDefinition: EVENT. Counted once per transcript turn in which the user explicitly marks unexpected capability (e.g. 'I didn't expect that', 'wow', 'how did it do that') OR, in think-aloud studies, verbalises surprise within 5 seconds of receiving output. Repeated markers about the same output count as one event.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater-coded count of surprise-marked turns per session; two independent coders, inter-rater reliability could be reported as Cohen's kappa. Optionally corroborated by a single post-task item ('How surprising was the system's performance?', 1-7).
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0245 — Perceived-Understanding State
also known as: The Seen Feeling
description: A graded subjective state in which a user judges an AI system's response to have accurately captured their intent or meaning. It denotes perceived understanding as reported by the user, not any mental state of the system.
operationalDefinition: STATE. Rated, not counted: after a response, the user rates perceived understanding of their intent on a Likert scale. The state is attributed when the rating exceeds a pre-registered threshold (e.g. >=6 on a 1-7 scale).
measurementSchema: Proposed measurement protocol (not yet empirically validated): Self-report Likert item ('The response captured what I actually meant', 1-7), adaptable from felt-understanding subscales; aggregated per session as a mean. Convergent check: third-party raters score response-to-intent fidelity (ICC).
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0128 — User-Directed Thanking
also known as: The Gratitude Response
description: A discrete instance in which a user directs an expression of thanks toward an AI system following a response they found helpful. The behaviour is an observable politeness act by the user and implies nothing about the system.
operationalDefinition: EVENT. Counted once per turn containing a user-authored expression of thanks ('thanks', 'thank you', 'danke' and close variants) addressed to the system. Thanks not directed at the system, or template sign-offs, are excluded by coder judgement.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Lexical-plus-coder detection of thanking turns; rate expressed as thanking turns per 100 user turns. Coder reliability on the include/exclude decision could be reported as Cohen's kappa to control for false positives (e.g. sarcasm).
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0529 — Productive-Connection State
also known as: The Closeness Bridge
description: A graded user-reported sense of productive connection that arises while using an AI system to externalise and structure one's own thoughts. It describes a task-coupled affective state of the user, not a social relationship.
operationalDefinition: STATE. Rated, not counted: users rate felt productive connection during a thought-structuring session on a Likert scale; the state is attributed above a pre-registered threshold. Trend variant tracks the rating across repeated sessions.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Self-report Likert items on productive connection while externalising thoughts (e.g. 'Working with the system helped me organise my own thinking', 1-7), session-mean aggregated; test-retest stability checked across sessions.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0081 — Post-Authorial Pride
description: A graded user-reported feeling of pride in an artefact co-produced with an AI system. It is an affective appraisal the user holds toward a jointly produced output, distinct from sole-authorship pride.
operationalDefinition: STATE. Rated, not counted: after completing an AI-assisted artefact, the user rates pride in the result on a Likert scale; attributed above a pre-registered threshold. Ownership attribution (self vs. system share) is recorded alongside.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Self-report Likert pride item adapted from achievement-emotion scales ('I feel proud of what we produced', 1-7), paired with a self/AI contribution split (% slider). Reported per artefact; no rater coding required.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0061 — Authorship-Attribution Query
also known as: Creator's Question
description: A discrete instance in which a user explicitly raises the attribution question of which part of a co-produced output is their own contribution versus the AI system's. It marks an observable moment of authorship reflection.
operationalDefinition: EVENT. Counted once per turn or think-aloud segment in which the user explicitly questions the division of contribution between self and system ('which part is mine?', 'did I do this or did it?'). General quality talk that does not raise attribution is excluded.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater-coded count of authorship-attribution utterances per session; two coders, Cohen's kappa on the attribution-vs-other decision. Frequency reported per session and per 100 user turns.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0020 — Recursive Feedback Loop
description: An interaction pattern in which AI output is edited by the user and returned for further processing across successive turns, producing progressive refinement of the artefact. It denotes an observable iterate-and-return turn sequence.
operationalDefinition: EVENT. Counted once per detected cycle in which user-modified content from a prior AI turn is resubmitted and the system produces a revised version. A run of k such cycles on one artefact is recorded as a chain of length k.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Turn-graph analysis of resubmission edges (content overlap between a user turn and the preceding AI turn above a similarity threshold); could be reported as cycle count and mean chain length per task. Coder validation of detected cycles via Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0005 — Integrated Operator
also known as: The Integrated Operator
description: A user mode in which a person no longer treats AI as a separate tool but moves fluidly between their own thinking and AI assistance without a distinct switching moment. The referent is an advanced, relatively stable working state characterised by low perceived separation between human and AI activity.
operationalDefinition: STATE. Assessed as a graded user-level condition combining (a) behavioural fluidity - short and infrequent deliberate transitions between unaided work and AI use - and (b) low self-reported sense of 'switching'. A participant is placed on the integration continuum from these indicators rather than by counting discrete events.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Mixed behavioural-and-self-report profiling: inter-action transition latency and frequency are derived from interaction logs, and integration is rated via a validated multi-item self-report scale (e.g. a tool-incorporation/flow inventory administered on a 5-point Likert metric). A composite integration score combines standardised log and survey indicators; internal consistency could be reported as Cronbach's alpha and the behaviour-survey convergence as a Pearson correlation.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0008 — Multi-System Orchestration Style
also known as: Polyphonic Sovereign
description: A working style in which a user deliberately orchestrates several AI systems and cross-checks their differing outputs against one another. It is a graded profile of multi-system, comparison-driven use.
operationalDefinition: STATE. Rated, not counted: scored by the degree to which a user routes the same task to multiple systems and compares results, from behavioural traces (number of distinct systems queried per task; presence of explicit comparison). Attributed above a pre-registered threshold.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Log-derived multi-system index (distinct models per task; cross-reference rate of one system's output in prompts to another) plus self-report on comparison practice; combined index, reliability via internal consistency.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0021 — Initialization Cascade
description: A user's personalised opening sequence at the start of an AI session, in which experienced users deploy characteristic set-up moves before substantive work. It denotes an observable session-initiation routine.
operationalDefinition: EVENT. Counted once per session in which the opening turns before the first task-bearing request match a recurring user-specific set-up template (e.g. role priming, context loading). Sessions opening directly with a task are scored as absent.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Sequence detection over the first N opening turns against a per-user template library; could be reported as presence rate across sessions and mean opening length (turns) before first task. Template membership validated by coder agreement (Cohen's kappa).
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0133 — Prompt Craftsmanship
description: A graded user skill in formulating effective AI inputs, developing with usage experience. It describes a measurable competence attributed to a person rather than a discrete interaction event.
operationalDefinition: STATE. Rated, not counted: prompt quality is scored against a defined rubric (specificity, constraint use, context provisioning, iteration efficiency) over a sample of a user's prompts; competence is the rubric-mean. Trend variant tracks the mean across time.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater-applied prompt-quality rubric (multi-criterion, anchored scale) over a fixed prompt sample per user; inter-rater reliability via ICC. Optionally validated against downstream task-success rate.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0134 — Context Window Awareness
description: A graded degree to which a user understands the mechanics of a model's context window and manages interaction strategically around its limits. It is an attributed knowledge-and-behaviour state of the user.
operationalDefinition: STATE. Rated, not counted: scored from (a) a short knowledge probe about context-window behaviour and (b) behavioural indicators of management (summarising, re-anchoring, or splitting near capacity). Attribution above a pre-registered composite threshold.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Composite of a knowledge-probe score and coder-rated management behaviours (frequency of deliberate context-management moves per long session); components standardised and combined; behaviour coding reliability via Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0018 — Trinaug Protocol
description: A user method in which the same task is issued to three AI systems and the resulting outputs are compared to identify shared and divergent patterns. It denotes a discrete three-way comparison procedure.
operationalDefinition: EVENT. Counted once per task for which a user obtains outputs from three distinct systems for the same prompt and produces an explicit comparison. Two-system or uncompared multi-system uses do not satisfy the criterion.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Procedure detection: same-prompt dispatch to exactly three distinct systems plus a comparison artefact; could be reported as count of three-way comparisons per project. Coder confirmation of the comparison step via Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0138 — Session Architecture
description: Characteristic patterns by which users structure the flow of an AI session, with experienced users showing recognisable phase organisation. It describes a graded structural property of a session's organisation.
operationalDefinition: STATE. Rated, not counted: a session transcript is scored for degree of structural organisation against a phase rubric (set-up, exploration, refinement, closure). The construct is the rubric-rated organisation level of the session.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater-applied session-structure rubric (presence and ordering of defined phases) yielding an organisation score per session; inter-rater reliability via ICC. Optionally a phase-transition count from automated segmentation as corroboration.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0091 — Productivity Arbitrage
description: The observed difference in task efficiency between AI-assisted and unassisted work processes for comparable tasks. It is a relational quantity comparing two performance conditions, not an event or feeling.
operationalDefinition: RATIO. Numerator: output or task units completed under the AI-assisted condition per unit time. Denominator: the same measure under a matched unassisted baseline. Values above 1 indicate an efficiency advantage; computed on matched task sets.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Within-subject A/B timing study on matched tasks: throughput (tasks per hour) AI-assisted vs. unassisted, could be reported as a ratio with confidence interval; quality held constant via a fixed acceptance rubric so speed gains are not quality losses.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0092 — Output Asymmetry
description: The relationship in which the effort needed to produce a result falls substantially under AI assistance while the quantity or quality of the result is maintained or increased. The referent is the changed ratio of input effort to output yield.
operationalDefinition: RATIO. Numerator: a measure of output yield (validated units produced, or an independently rated quality score). Denominator: a measure of input cost (active user time, edits, or token/turn effort) for the same task. Asymmetry is the ratio of yield to cost compared between AI-assisted and unaided conditions for matched tasks.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Within-subject controlled comparison: participants complete matched tasks with and without AI; input cost (logged active time and revision count) and output yield (deliverable count plus blind expert quality ratings on a fixed rubric) are recorded. Asymmetry could be reported as the yield-per-unit-cost ratio and its assisted/unaided difference (paired comparison with effect size); quality-rating reliability across blind judges via ICC.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0095 — One-Person Operation
description: An observed condition in which AI assistance enables a single individual to reach a level of productive output previously requiring a larger team. It describes a capacity-expansion condition at the level of an individual operator.
operationalDefinition: STATE. Attributed when a single operator, with AI assistance, delivers a workload or scope-of-output that a pre-defined benchmark assigns to a multi-person team (e.g. meets a role-coverage checklist normally spanning >=N roles). Graded by number of distinct roles covered solo.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Role-coverage audit: count of distinct functional roles a solo operator fulfils with AI assistance against a benchmark team composition; could be reported as roles-covered ratio. Corroborated by output-volume comparison to team baselines where available.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0031 — Idea-Trigger Event
also known as: Semantic Spark
description: A discrete instance in which an AI response triggers a new, unanticipated idea in the user, where the value lies in the thought it provokes rather than the answer itself. It marks an observable user-reported insight moment.
operationalDefinition: EVENT. Counted once per turn or think-aloud segment in which the user reports that a response prompted a new idea of their own ('that gives me an idea', 'that made me think of...'). Mere acceptance or use of the answer without a new idea is excluded.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater-coded count of user-reported insight-trigger turns per session; two coders, Cohen's kappa on the insight-vs-no-insight decision; rate per 100 user turns. Self-report corroboration via a per-session 'new ideas sparked' tally.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0084 — Glitch-Mining
description: A user practice of treating AI errors as informative, deliberately examining what a malfunction reveals about a system rather than discarding it. It denotes a discrete investigative use of an observed error.
operationalDefinition: EVENT. Counted once per episode in which a user, upon encountering a model error, explicitly probes or analyses the error to infer system behaviour (follow-up questions about the failure, deliberate re-elicitation). Simply re-prompting for a correct answer is excluded.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater-coded count of error-leveraging episodes per session, conditioned on detected error turns; could be reported as episodes per 100 error events. Coder reliability on the leverage-vs-retry distinction via Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0030 — Contextual Gravity
description: The tendency of an AI session to develop an increasing thematic pull toward its accumulated context as that context grows over the course of the session. It describes a longitudinal drift in topical concentration.
operationalDefinition: TREND. Tracked over the session timeline: the degree to which later turns remain within the topic established by accumulated earlier context, measured as change in topical concentration from early to late session windows. An upward trend indicates gravity.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Longitudinal topic-concentration index across binned session windows (e.g. embedding-based similarity of each turn to the running session centroid); slope over turns reported per session. Validated against coder topic-continuity ratings (ICC).
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0524 — Context Layer
also known as: The Context Layer
description: The accumulated interaction context of a session treated as a working substrate that conditions subsequent system output. It is a superordinate organising concept spanning many context-related phenomena (such as how context is managed, accumulated, and reused) rather than a single observable occurrence.
operationalDefinition: FRAMEWORK. Superordinate framework term; not operationalized as a per-interaction event. It groups context-related constructs (e.g. context-window awareness, contextual gravity) rather than denoting a single countable occurrence.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Not directly measured; superordinate term.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0176 — Capability Discovery
description: A discrete instance in which a user identifies a previously unknown capability of an AI system through exploration or experimentation. It marks an observable moment of first encountering a new affordance.
operationalDefinition: EVENT. Counted once per session in which a user first elicits and recognises a system capability they had not previously used, evidenced by exploratory prompting followed by a recognition marker ('oh, it can also...'). Re-use of known capabilities is excluded.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater-coded count of first-encounter capability events per user over a usage period, cross-referenced against the user's prior capability set; could be reported as discoveries per session. Coder agreement on novelty judgement via Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0107 — Output-Verification Disposition
also known as: Verification Principle
description: A user disposition toward systematically checking and contextualising AI results before relying on them. It describes a graded verification practice attributed to a user, distinct from any single checking act.
operationalDefinition: STATE. Rated, not counted: scored as the proportion of consequential AI outputs a user subjects to an explicit verification step (external check, cross-source, or re-derivation), plus self-reported verification habit. Higher proportion indicates the disposition.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Behavioural verification rate (verified outputs / consequential outputs) from log coding, combined with a self-report verification-habit subscale; behaviour coding reliability via Cohen's kappa. Note: name overlaps logical-positivism term; see renameReason.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0126 — Iterative Diminishing-Returns Point
also known as: Semantic Saturation
description: The point in an iterative AI exchange at which further iteration ceases to add value, functioning as a signal to stop. It marks a discrete diminishing-returns threshold within a refinement sequence.
operationalDefinition: EVENT. Counted once per refinement chain at the iteration after which marginal improvement falls below a pre-registered criterion (rated quality gain < epsilon, or the user explicitly stops citing no added value). The saturating iteration index is recorded.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Marginal-gain tracking across iterations: per-iteration quality rating (anchored rubric) with the saturation point at the first sub-threshold gain; could be reported as mean iterations-to-saturation. Inter-rater reliability of gain ratings via ICC.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0142 — Model Fingerprint
description: The recognisable stylistic signature of a given AI model, such that experienced users can identify which system produced a response. It describes a distinguishable stylistic property of a model's outputs.
operationalDefinition: STATE. Operationalised as identifiability: the degree to which a model's outputs can be correctly attributed to it. Measured by recognition accuracy in a blinded source-attribution task; higher-than-chance accuracy indicates a detectable fingerprint.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Blinded model-attribution experiment: raters assign unlabelled outputs to source models; identifiability could be reported as accuracy above chance and per-model confusion matrix. Stylometric feature-distance between models could be reported as a convergent measure.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0135 — Persona Engineering
description: The deliberate shaping of an AI system's expressed persona via instructions to fit specific tasks or contexts. It denotes an observable user act of configuring system style, not a trait of the system itself.
operationalDefinition: EVENT. Counted once per session in which a user issues explicit persona-shaping instructions (role, tone, or character directives intended to set the system's expressed style). Ordinary task instructions without a persona directive are excluded.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater-coded count of persona-directive turns per session; could be reported as rate per session and share of sessions containing at least one directive. Coder reliability on the persona-vs-task-instruction distinction via Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0006 — Platform Ontology
description: The set of systematic dispositions in a given AI system's outputs that derive from its training data and architecture, such that different systems produce characteristically different framings of the same prompt. A property attributed to a model configuration, not to a single reply.
operationalDefinition: STATE. The same fixed prompt set (held constant across systems) is submitted to two or more AI systems; raters score each system's aggregate response profile for characteristic, reproducible framing tendencies (e.g. hedging, domain emphasis, refusal style) on a 0-4 anchored scale. The construct is the between-system divergence in these scored tendencies.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Held-constant prompt battery (>=30 prompts) run on each system; two raters score response-tendency dimensions on 0-4 anchors; between-system divergence could be reported as mean pairwise distance with inter-rater ICC(2,k) for the dimension scores.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0566 — Ongoing Partnership
description: A longitudinal pattern in which a single user's repeated working interactions with one AI system show increasing efficiency or stability over time, such as shorter prompt-to-acceptable-output paths or more reused conventions across sessions.
operationalDefinition: TREND. Tracked per user-system pair across successive sessions over a defined window (e.g. 8 weeks). Indicators logged per session: turns-to-accepted-output, count of reused user-defined conventions, and session-level task-completion rate. The construct is the slope of these indicators over the window.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Per-session telemetry over >=6 sessions: turns-to-accepted-output, reused-convention count, completion rate; trend estimated by linear mixed-effects slope per user; effect reported with 95% CI.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0539 — Companion Shift
also known as: The Companion Shift
description: A reported change over time in how an individual user characterizes an AI system, moving from describing it as a fixed single-task tool toward describing it as a flexible multi-purpose working instrument. Concerns the user's stated framing, not the system's properties.
operationalDefinition: TREND. The user completes a short framing questionnaire at intervals (e.g. baseline, week 4, week 8) containing items contrasting 'fixed tool' vs 'flexible instrument' descriptors. The construct is the directional change in the framing-index score across timepoints for that user.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Repeated 6-item framing questionnaire (fixed-tool vs flexible-instrument, 5-point bipolar) at >=3 timepoints; change scored as within-person slope; scale reliability reported via Cronbach's alpha per wave.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0170 — Articulation-to-System Effect
also known as: Witness Effect
description: A user-reported experience of an AI system as an attentive interlocutor to which thoughts can be externalized, where the act of articulating to the system is rated as clarifying. The construct concerns the user's subjective report and ascribes no inner states to the system.
operationalDefinition: STATE. After a session in which the user used the system primarily to think aloud or externalize reasoning, the user rates perceived attentiveness and clarification gain on Likert items. The construct is the rated degree of externalization-clarification, not any property of the system.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Post-session self-report: 4 Likert items (5-point) on perceived attentiveness and clarification-through-articulation; could be reported as item means; convergent check against a single 'I think more clearly' item via Spearman rho.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0177 — Trust Setting
also known as: The Trust Setting
description: The graded level of reliance a user deliberately places on a given AI system's outputs for a class of tasks, expressed in how much independent verification the user applies before acting on those outputs. A consciously adjusted disposition rather than a fixed trait.
operationalDefinition: STATE. Operationalized as the user's stated reliance level plus a behavioral proxy: the proportion of AI outputs the user independently verifies before use within a task class. Lower verification proportion indicates a higher trust setting. Rated/logged per task class.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Composite: self-reported reliance (single 0-10 item) plus behavioral verification ratio (independently-checked outputs / total used outputs) over >=20 outputs; reported jointly; self-report vs behavior concordance via correlation.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0577 — Early-Interaction Alignment Curve
also known as: The Calibration Dance
description: The early phase of a new user-system working relationship during which prompts and responses progressively align, observable as a decline in clarification exchanges and reformulations across the first several sessions before reaching a plateau.
operationalDefinition: TREND. Tracked across the opening sessions of a user-system pairing. Per session, raters or logs count clarification/reformulation turns (turns that restate or correct intent). The construct is the downward trajectory of this count until it stabilizes.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Clarification-turn count per session across the first >=5 sessions; trajectory modeled as session-indexed decline to plateau (e.g. fitted exponential decay); rater coding of clarification turns checked with Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0161 — Invisible Colleague
description: A usage configuration in which an AI system functions as an on-demand background work aid that the user invokes for discrete subtasks without sustained turn-by-turn dialogue or social framing. Characterized by short, intermittent, task-bounded calls.
operationalDefinition: STATE. Classified from interaction logs over a period: a user-system relationship is in this state when sessions are predominantly short, single-purpose invocations (e.g. median <=3 turns) interspersed across work rather than sustained conversational sessions. The construct is this session-shape profile.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Log-derived session-shape metrics over a usage window: median turns per session, share of single-turn invocations, inter-invocation gap; state assigned by threshold rule; no rater required (objective telemetry).
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0981 — Companion Pattern
also known as: The Companion Pattern
description: A standing usage mode in which a person incorporates AI assistance into daily routines as a steady accompaniment for thinking, drafting, and exploration, operating alongside rather than in place of their own reflection. The referent is the co-present, habitual mode of use.
operationalDefinition: STATE. Assessed as a graded user-level condition indicated by (a) regularity of AI use across daily routine contexts and (b) a complementary rather than substitutive relationship to the user's own reflection (the user continues to contribute and revise rather than deferring wholesale). Placement on the continuum derives from routine-integration breadth and the human-contribution share in outputs.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Routine-integration profiling: breadth and regularity of AI use across daily contexts are derived from usage logs, and the complementarity of use is estimated from the human-edit share of co-produced outputs plus a self-report routine-companionship scale (Likert). A composite index combines standardised regularity and complementarity indicators; internal consistency via Cronbach's alpha and log-survey convergence via Pearson correlation.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0197 — Sustained-Focus State
also known as: The Shared Quiet
description: A user-reported state of sustained, undistracted concentration during a stretch of productive work with an AI system, in which the tool fades into the background of attention. The construct is the user's focus state, attributed to the person, not jointly to the system.
operationalDefinition: STATE. The user rates depth and continuity of concentration for a defined work block on a focus scale, optionally corroborated by an objective proxy such as uninterrupted on-task duration before the next context switch. The construct is the rated focus level during AI-assisted work.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Post-block self-report on a brief focus/flow self-report (a short multi-item scale, 7-point) plus optional on-task-duration proxy; could be reported as scale mean; if proxy used, report self-report vs duration correlation.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0201 — Low-Stakes Exploration Space
also known as: The Closeness Bridge
description: A reported use of AI interaction as a low-stakes space for structured exploration of one's own ideas, absent the social considerations of a human audience. The construct concerns the user's described use of the channel, not emotional closeness to the system.
operationalDefinition: STATE. The user rates, after relevant sessions, the extent to which the interaction was used to explore or rehearse thinking without social inhibition (e.g. judgment-free exploration items). The construct is the rated degree of inhibition-free exploratory use of the channel.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Post-session self-report: 4-5 Likert items (5-point) on judgment-free exploratory use and absence of social-performance concern; could be reported as subscale mean; internal consistency via Cronbach's alpha.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0080 — Relationship-First
description: A stated user value-orientation that places human relationships ahead of AI-assisted productivity when the two are in tension. As a superordinate guiding principle it organizes choices across situations rather than denoting a single observable interaction event.
operationalDefinition: FRAMEWORK. Superordinate value-orientation term; not operationalized as a per-interaction event. Where measured at all, it is assessed as a stated priority disposition (e.g. forced-choice tradeoff items pitting availability for people against AI-task continuation), not by counting interaction occurrences.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Not directly measured as an event; superordinate term. If a disposition index is desired, use a forced-choice value-priority questionnaire (relationship vs productivity tradeoffs) could be reported as a preference score; otherwise treated as a non-countable principle.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0119 — Level Playing Field
description: A population-level tendency for AI tools to widen access to knowledge and capabilities across user groups over time, such that gaps between previously advantaged and less advantaged groups on AI-mediated tasks narrow. A macro trend, not an interaction-level event.
operationalDefinition: TREND. Tracked across cohorts or regions over time using an access/capability indicator (e.g. share of a population able to complete a defined knowledge task with AI assistance, or a between-group disparity index). The construct is the change in the disparity indicator over the observation window.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Longitudinal between-group disparity index (e.g. ratio or gap in task-completion or access rates across defined groups) measured at >=2 timepoints from survey or usage data; trend reported with uncertainty interval.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0010 — Cross-Era Cohort
also known as: Bridge Species
description: A label for the cohort of people whose working lives span both the pre-AI and post-AI periods, giving them firsthand comparative experience of both. A descriptive population category, not a measurable interaction phenomenon.
operationalDefinition: FRAMEWORK. Superordinate population-category term; not operationalized as a per-interaction event. Membership is a demographic/biographical classification (active professional experience before and after widespread AI tool adoption), assessed by survey, not by interaction coding.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Not directly measured as an interaction; superordinate category. Membership, if needed, is assigned by a biographical survey item (professional activity spanning defined pre- and post-adoption periods); could be reported as cohort prevalence.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0198 — New Literacy
also known as: The New Literacy
description: A societal tendency for competence in using AI tools to consolidate into a broadly expected cultural skill, comparable to earlier shifts in expected literacy. Denotes a gradual rise in baseline expected AI competence, not a single interaction.
operationalDefinition: TREND. Tracked at population level via the prevalence of defined AI-use competencies (e.g. share able to perform a benchmark set of AI tasks) and the degree to which such competence is treated as expected (job postings, curricula). The construct is the upward movement of these indicators over time.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Repeated population surveys of benchmarked AI-competence prevalence plus content analysis of expectation signals (job-ad / curriculum mentions) across years; trend could be reported as indexed change over time.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0093 — Zero-Marginal Cost
description: The tendency for the incremental cost of producing one additional unit of creative or informational output to fall toward negligible levels when AI assistance is used. Expressed as cost per additional produced unit, observed at the level of a task or workflow.
operationalDefinition: RATIO. Numerator: incremental resource cost (time, money, or compute) attributable to producing one further output unit with AI assistance. Denominator: one additional accepted output unit. The construct is this marginal cost-per-unit, compared against the non-AI baseline marginal cost.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Marginal cost-per-unit = incremental cost / additional accepted unit, computed from task time-and-cost logs with and without AI assistance; could be reported as the AI-vs-baseline ratio of marginal costs with dispersion across tasks.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0099 — Adoption Window
description: A bounded early period in the uptake of an AI technology during which usage patterns form and their downstream effects are documented. Denotes a time interval of early adoption observed at population or organizational level, not a single interaction.
operationalDefinition: TREND. Defined over the early adoption interval for a given tool and population. Tracked indicators: adoption rate over time and the emergence of stable usage practices. The window is delimited (e.g. from first availability until adoption-rate inflection); the construct is the early-interval trajectory of these indicators.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Adoption-rate time series (cumulative or period uptake) over the early interval, with the window bounded by an inflection criterion; emergent-practice indicators tracked across the same interval; could be reported as a dated trajectory.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0730 — Open-Source Path
also known as: The Open-Source Path
description: A tendency toward the increasing availability of openly released AI models and the documented effects of that openness on who can access and build on such models. A macro trend in the model ecosystem, not an interaction-level phenomenon.
operationalDefinition: TREND. Tracked at ecosystem level via indicators such as the share of widely used models that are openly released and the breadth of downstream access (e.g. derivative or deployment counts). The construct is the change in these openness/access indicators over time.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Ecosystem time series: proportion of leading models released under open terms and downstream-access breadth (derivative/deployment counts) sampled across periods; trend could be reported as indexed change with source provenance.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0834 — Public Sentiment Oscillation
also known as: Perception Wave
description: An oscillation over time in aggregate public sentiment toward AI, cycling between phases of enthusiasm and phases of caution. Denotes a longitudinal swing in measured population sentiment, observed at the level of discourse, not individual interaction.
operationalDefinition: TREND. Tracked via a population sentiment index toward AI sampled repeatedly over time (survey balance of favorable vs cautious responses, or validated sentiment scoring of a defined media/discourse corpus). The construct is the cyclical movement of this index across the observation window.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Repeated AI-sentiment index from representative surveys and/or validated sentiment classification of a fixed discourse corpus; sampled at regular intervals; oscillation characterized by amplitude/turning points over the series.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0808 — Knowledge Access Pattern
description: A descriptive label for the different routes by which groups and regions reach AI-mediated knowledge, characterizing pathways of access rather than enumerating barriers. A superordinate descriptive category spanning many specific access routes.
operationalDefinition: FRAMEWORK. Superordinate category term; not operationalized as a per-interaction event. Specific access routes (device, language, cost, intermediary) can each be surveyed, but the umbrella term itself denotes the typology of pathways rather than a single countable occurrence.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Not directly measured as one variable; superordinate typology. Constituent routes are measured separately via access surveys (by device, language, cost, intermediary) and could be reported as a profile; the umbrella term aggregates these descriptively.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0847 — Digital Native Dialogue
description: An exchange about AI spanning generational lines, in which younger habitual technology users and members of earlier generations share perspectives, characterized by reciprocal information transfer rather than one-directional instruction. The unit is a cross-generational conversational exchange.
operationalDefinition: EVENT. One occurrence is a recorded conversation about AI between participants from distinct generational groups in which bidirectional contribution is present (each side both asks and explains at least once). Coders mark presence of reciprocity to qualify the exchange as an instance.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater coding of recorded cross-generational conversations for the reciprocity marker (both parties give and receive >=1 explanation); instances counted per corpus; coder agreement on the reciprocity judgment reported via Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0071 — AI-Output Verification Practice
also known as: Epistemic Hygiene
description: The degree to which a user applies systematic checking practices to AI-generated information, such as source-checking, cross-referencing across independent sources, and contextual appraisal before relying on it. A graded behavioral disposition in handling AI output.
operationalDefinition: STATE. Operationalized as the proportion of AI-provided claims a user subjects to defined checking steps (independent source confirmation, triangulation, contextual qualification) before use, observed over a set of outputs or via a behavior-frequency self-report. The construct is the level of checking applied.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Checking-behavior frequency: per claim, code which of a fixed checklist of verification steps were performed (observed logs or diary); report mean checks-per-claim; if self-reported, use a behavior-frequency scale with test-retest reliability.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0075 — Knowledge-Base Curation Routine
also known as: The Gardener Protocol
description: A maintenance routine in which a user periodically reviews, sorts, updates, and prunes an accumulated AI-assisted knowledge base to keep it current and accurate. The construct is the regularity and completeness of this curation behavior.
operationalDefinition: STATE. Operationalized by the cadence and coverage of curation actions on a knowledge store: how regularly review passes occur and what share of stored items are checked, updated, or removed per pass. The construct is the level of upkeep, ratable from maintenance logs or self-report.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Curation-activity metrics from version history or logs: review-pass interval, fraction of items revised/removed per pass, and staleness rate (share of items past a freshness threshold); could be reported as an upkeep profile.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0076 — Self-Referential Grounding
description: The extent to which a user anchors decisions during AI interaction in their own explicit value system, using personal criteria as the reference point when appraising or accepting AI output. A graded reliance on internally held standards.
operationalDefinition: STATE. Rated from think-aloud or annotated decisions: the share of AI-output acceptance/rejection decisions the user justifies by reference to stated personal values or criteria, versus deferring without such reference. The construct is the degree of value-anchored appraisal.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater coding of decision rationales (think-aloud or written) for explicit value-criterion references; could be reported as proportion of value-anchored decisions; inter-rater agreement would be assessed via Cohen's kappa on the rationale-type code.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0024 — Internal Judgment Anchor
also known as: Built-In Compass
description: A user's internal orientation, formed from personal values, experience, and intuition, that they report drawing on to judge and correct AI output. The construct is the user's reliance on this internal reference, not any faculty of the AI system.
operationalDefinition: STATE. Self-reported reliance on internal judgment (values, experience, intuition) when evaluating AI output, rated on Likert items, optionally corroborated by coded instances where the user overrides AI output citing personal judgment. The construct is the rated strength of this internal corrective reliance.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Self-report on a short scale (5-point) measuring reliance on internal judgment to vet AI output, plus optional behavioral corroboration (rate of self-justified overrides); scale internal consistency via Cronbach's alpha.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0011 — Reflective Operator
description: A user profile characterized by observable reflection on one's own AI use and deliberate refinement of that use over time, such as reviewing past interactions and adjusting strategies. The construct is the degree of metacognitive, self-revising engagement with one's AI practice.
operationalDefinition: STATE. Operationalized via indicators of reflective practice: frequency of reviewing or annotating past interactions, documented strategy adjustments, and self-reported metacognitive engagement. A user scores higher when these reflection-and-revision indicators are more present. The construct is graded, not a discrete event.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Composite reflective-practice score: log indicators (revisits/annotations of prior sessions, count of recorded strategy changes) plus a short metacognition self-report scale; could be reported as a standardized composite with subscale reliabilities.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0078 — Naming Power
description: The general proposition that assigning a name to a phenomenon can alter how that phenomenon is subsequently perceived and discussed. A superordinate conceptual claim about language and perception, not a per-interaction occurrence.
operationalDefinition: FRAMEWORK. Superordinate conceptual term; not operationalized as a per-interaction event. It can motivate specific testable hypotheses (e.g. introducing a label changes recognition or reporting rates of a phenomenon), but the umbrella claim itself is not a countable interaction unit.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Not directly measured as an interaction; superordinate term. A derived hypothesis could be tested experimentally (pre/post-label recognition or reporting rates in matched groups), but the concept itself is treated as non-countable.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0001 — Augmanitai
description: A superordinate framework term for the deliberate, productive collaboration between humans and AI in everyday contexts, viewed from the human side. It blends 'augmentation' with a coined 'humanitai' root and, unlike 'artificial intelligence', foregrounds the human's experience and agency rather than the machine.
operationalDefinition: FRAMEWORK. Superordinate framework term; not operationalized as a per-interaction event. It names the overarching subject matter under which the lexicon's countable phenomena are organised and is not itself directly counted.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Not directly measured; superordinate term. Its scope is articulated through the constituent constructs of the framework, each of which carries its own instrument.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0023 — Vigilance Imperative
description: A verification act in which a human user independently checks an AI-generated output against an external reference or domain knowledge before relying on it, distinguished from passive acceptance by the presence of an observable confirmation or correction step.
operationalDefinition: EVENT. One occurrence would be coded when, following an AI output, the user performs an observable checking action against an independent source (cross-reference lookup, recomputation, citation follow-up, or explicit correction of the output) before acting on it. Mere re-reading of the AI text without an external check does not count.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Behavioural event coding from interaction logs: count of verification acts per task, with each act tagged by check type (source lookup / recomputation / correction). Two raters double-code a 20% sample; inter-rater agreement could be reported as Cohen's kappa on the present/absent decision.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0102 — Decision Authority Retention
also known as: Sovereignty Principle
description: A configuration of a human-AI workflow in which final decision authority is retained by the human rather than transferred to the AI system, characterised by the human's capacity to set, confirm, or override any AI-proposed action across delegated tasks.
operationalDefinition: STATE (graded). The degree of retained human decision authority is rated per workflow on a 1-5 ordinal scale, anchored 1 = AI executes consequential actions without a human confirmation point, 5 = every consequential action requires explicit human confirmation. Rating is based on documented control points in the workflow, not on user self-perception.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Workflow audit by trained raters who map each consequential action to its control point and assign the 1-5 authority-retention score; agreement across raters could be reported as ordinal ICC (two-way, agreement). Complementary objective indicator: proportion of consequential actions gated by a human confirmation step.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0145 — Responsibility Gradient
description: A proposed monotonic relationship in which the share of accountability assigned to the human increases with the extent of authority delegated to the AI, treated as a relationship between delegated scope and assigned responsibility rather than as a single event.
operationalDefinition: TREND/relationship. Operationalized as the monotonic association between (a) delegated-authority level, scored per task on a defined 1-5 delegation scale, and (b) the share of outcome accountability formally assigned to the human, expressed as a 0-1 proportion from a responsibility-assignment matrix. The construct is the direction and strength of the relationship (rank correlation / slope) between (a) and (b) across a task set, not a single ratio value and not a per-turn count.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Per-task coding of delegation level and of assigned human-accountability share, then estimation of the rank correlation (Spearman's rho) between the two across tasks; positive monotonic association supports the construct. Delegation-level coding agreement could be reported as weighted kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0859 — Agent Handshake
description: The initial exchange that opens an interaction between a human and an autonomous agent, in which the agent's stated capabilities, permissions, and limits are surfaced; distinguished from later turns by being the first establishment of the working terms.
operationalDefinition: EVENT. One occurrence is counted at the first turn of a human-agent session in which the agent discloses, or the human sets, the agent's scope (capabilities, permissions, or constraints) before task work begins. A session in which work starts with no such scope-setting exchange is coded as absent.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Presence/absence coding of the opening scope-setting exchange per session from transcripts, plus latency measured as number of turns until first task action. Two raters code an overlap sample; agreement on the present/absent judgement could be reported as Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0860 — Delegation Depth
description: The breadth of decision latitude granted to an autonomous agent, ranging from narrow execution of pre-specified steps to discretionary action within broad goals; a graded property of a delegation arrangement rather than a discrete occurrence.
operationalDefinition: STATE (graded). The granted latitude is rated per delegation arrangement on a 1-5 ordinal scale, anchored 1 = agent may only execute explicitly enumerated steps, 5 = agent may choose means and sub-goals within a high-level objective without per-step approval. Scoring is from the configured permission set, not from observed outcomes.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rater coding of the agent's permission configuration onto the 1-5 latitude scale; inter-rater reliability could be reported as ordinal ICC. Objective complement: mean number of distinct action types the agent may perform without seeking approval.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0914 — Physical Presence
description: A condition in which an AI system has a physical, embodied form present in the user's environment, treated as a graded contextual property of the interaction setting rather than as a discrete event or an effect claim.
operationalDefinition: STATE. Coded per interaction setting on an ordinal scale of physical co-presence: 0 = screen/text only, 1 = audio or projected presence, 2 = stationary embodied device, 3 = mobile/object-handling embodied robot sharing the user's space. Coding is based on the system's observable form, independent of any behavioural outcome.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Setting-level coding of the co-presence level (0-3) by raters from session metadata or video, with agreement could be reported as weighted kappa. Distinct from any outcome instrument: this schema indexes the embodiment condition only, not its consequences.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0888 — Human-in-the-Loop
description: An established interaction-design pattern in which a human acts as a supervisory checkpoint within an otherwise automated decision loop, retaining the ability to review or intervene; a superordinate design concept rather than a single countable occurrence.
operationalDefinition: Superordinate framework term; not operationalized as a per-interaction event. It names a class of control architectures (human review embedded in an automated loop); specific instances are studied through narrower constructs such as confirmation-gating or override behaviour rather than by counting the pattern itself.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Not directly measured; superordinate term.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0889 — Agent Ensemble
description: A configuration in which several AI agents operate jointly on a task under human supervision, characterised by coordination among the agents and a human orchestration role; a structural property of the setup rather than a discrete event.
operationalDefinition: STATE. Coded per task configuration by (a) the count of distinct agents acting on the shared task and (b) presence of a human orchestration point that assigns or arbitrates among them (present/absent). A single-agent setup, or a multi-agent setup with no human orchestration point, is coded accordingly.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Configuration coding from system logs: number of concurrently active agents per task and a binary human-orchestration-present flag; the flag's inter-rater agreement could be reported as Cohen's kappa. Agent count is a direct log readout, not a rating.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0868 — Rollback Option
also known as: The Rollback Option
description: A property of a control architecture in which AI-initiated actions can be reversed to a prior state, characterised by the availability and coverage of undo or restore mechanisms across the system's actions; an affordance rather than a discrete event.
operationalDefinition: STATE. Operationalized as reversibility coverage: the proportion of the system's consequential action types for which a documented undo/restore path exists, scored 0-1 from the system specification. Optionally banded (none / partial / full). It indexes designed affordance, not how often undo is used.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Specification audit enumerating consequential action types and checking each for a documented reversal path; reversibility coverage could be reported as the proportion reversible. Two auditors independently classify a sample; agreement on the per-action reversible/irreversible decision could be reported as Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0921 — Swarm Access
description: A condition in which a user can draw on the combined capacity of multiple networked agents under human supervision, treated as a graded property of the available capability rather than as a discrete event or a scaling claim.
operationalDefinition: STATE. Coded per configuration by the number of networked agents a user can concurrently marshal and whether their outputs are aggregated under a single human-supervised control point (present/absent). It describes accessible breadth of agent capacity, not the quality of any result.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Configuration readout from orchestration logs: count of concurrently reachable agents and a binary aggregated-under-human-supervision flag; flag agreement could be reported as Cohen's kappa. Distinct from Agent Ensemble by indexing reachable breadth rather than active joint operation on one task.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0915 — Embodiment Effect
description: An observed difference in interaction dynamics associated with an AI system having a physical, embodied form versus a text-only form; reported as a measurable contrast condition rather than asserted as a fixed causal law.
operationalDefinition: STATE/contrast. Operationalized as the difference in a pre-specified interaction measure (e.g., user compliance rate, turn latency, or trust rating) between matched embodied and text-only conditions of the same system. The construct is the between-condition difference on that measure, established by comparison, not a per-turn count.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Within- or between-subjects comparison of embodied vs text-only conditions on a registered outcome measure; effect could be reported as a standardized mean difference (Cohen's d) with confidence interval. The embodiment condition itself would be coded as in Physical Presence.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0944 — Autonomy Ladder
description: A staged classification of increasing agent autonomy, ordering arrangements from fully human-controlled through to agent-autonomous with a reserved human veto; a superordinate ordinal scheme used to locate cases rather than a countable occurrence.
operationalDefinition: Superordinate framework term; not operationalized as a per-interaction event. It provides an ordered set of autonomy levels onto which specific delegation arrangements are classified; measurement attaches to placing a case on the ladder (see Delegation Depth), not to counting the ladder itself.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Not directly measured; superordinate term.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0954 — Congruence Review
description: A periodic review in which an AI system's observed behaviour is checked for continued correspondence with its intended goals and constraints; distinguished by a documented review occurrence that compares behaviour against a stated specification.
operationalDefinition: EVENT. One occurrence would be coded when a documented review compares sampled system behaviour against its stated goals/constraints and records a conformant-or-deviation verdict. Routine use without such a documented behaviour-versus-specification comparison is not counted.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Audit-record coding: count of completed congruence reviews per period, each yielding a conform/deviation verdict, with deviation severity rated 1-5. Two coders rate a verdict sample; agreement on the conform/deviation call could be reported as Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0009 — Speed Limit
also known as: The Speed Limit
description: A working condition in which the pace of an AI-supported task is kept within the rate at which the human can review the outputs, described as the relationship between output throughput and human checking capacity rather than as advice.
operationalDefinition: STATE/ratio. Operationalized as a pacing index: the ratio of AI outputs reviewed by the human to AI outputs produced over a task window (0-1). Values near 1 indicate throughput stays within review capacity; low values indicate outputs accrue faster than they are checked. Computed from logs, not self-report.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Throughput/review counting from interaction logs: outputs-produced and outputs-reviewed per fixed window, expressed as the reviewed-to-produced ratio. No rater judgement is required; reliability rests on log completeness, which is could be reported as the proportion of outputs with a recorded review status.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0632 — Offline Moment
also known as: The Offline Moment
description: A discrete instance of a user deliberately ending or stepping away from AI interaction, marking a boundary in the work session; distinguished from incidental gaps by being an intentional disengagement.
operationalDefinition: EVENT. One occurrence would be coded when a session log shows an intentional disengagement marker, operationalized as a user-initiated session close or an idle gap exceeding a pre-set threshold (e.g., >= 30 minutes) that is not an automatic timeout. Brief within-task pauses below threshold do not count.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Session-log event extraction: count of qualifying disengagement events per day, with each event timestamped and its duration recorded. Threshold and user-initiated-close detection are rule-based; a manual audit sample verifies classification, could be reported as percent agreement with the rule.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0951 — Value Constraint Encoding
also known as: Value Lock
description: A condition in which a set of intended human values is encoded into an autonomous system's operating constraints such that the system's behaviour remains bounded by them; described as the presence and coverage of such encoded constraints rather than as a guarantee.
operationalDefinition: STATE. Operationalized as constraint coverage: the proportion of a predefined value-constraint checklist that is verifiably encoded and enforced in the system's behaviour under test (0-1). It indexes how comprehensively the stated values are bound into operation, not whether values are 'truly' fixed.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Constraint conformance testing: a fixed battery of value-relevant test cases is run and the share the system handles within the intended bounds is could be reported as coverage, with the residual logged as out-of-bound cases per 1000 test cases. Test-case pass/fail is rule-scored; an audited subsample reports percent agreement.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0962 — Pause Protocol
also known as: The Pause Protocol
description: A discrete deliberate interruption inserted before a critical AI-supported decision, creating a checkpoint at which the human reconsiders before proceeding; distinguished by an observable hold step preceding the decision.
operationalDefinition: EVENT. One occurrence would be coded when an observable hold step is recorded immediately before a designated critical decision, operationalized as an explicit confirmation prompt, an enforced wait, or a logged user-initiated stop preceding commitment. Proceeding to the critical decision with no such hold step is coded as absent.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Event coding from decision logs: count of decisions preceded by a qualifying hold step, expressed as a proportion of critical decisions, plus mean hold duration. Two coders judge the hold-present/absent marker on a sample; agreement could be reported as Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0120 — Range Framework
description: A defined usage boundary specifying what an AI system is permitted and not permitted to do within a given context, expressed as an explicit scope agreement; a superordinate governance construct rather than a per-interaction event.
operationalDefinition: Superordinate framework term; not operationalized as a per-interaction event. It denotes the documented permitted-use scope for a context; observable measurement attaches to adherence to that scope (e.g., out-of-scope action rate) rather than to counting the framework itself.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Not directly measured; superordinate term.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0958 — Accountability Chain
description: A defined chain of responsibility specifying, across roles and levels, who decides, who reviews, and who is answerable for AI-supported outcomes; a superordinate governance structure rather than a discrete countable event.
operationalDefinition: Superordinate framework term; not operationalized as a per-interaction event. It denotes the mapping of decision, review, and answerability roles for an AI-supported process; measurement attaches to properties of an instantiated chain (e.g., completeness of role assignment) rather than to counting the chain itself.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Not directly measured; superordinate term.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0955 — Transparency Layer
description: A condition in which an AI system's processes and outputs are rendered inspectable and explainable to relevant parties, described as the graded extent of available, accurate disclosure rather than as a single event.
operationalDefinition: STATE (graded). Rated per system on a defined transparency rubric covering disclosure of AI involvement, of provenance/sources, and of decision rationale, each scored 0-2 and summed to an ordinal index. Scoring is from inspectable system artefacts (notices, logs, explanations), not from user opinion.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Rubric-based rater scoring of the disclosure dimensions (AI-use notice / source provenance / rationale availability) to form the transparency index; inter-rater reliability could be reported as ordinal ICC. Items map to concrete artefacts so scoring is auditable.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

### AUG-0771 — Minor Protection
description: A protective standard governing minors' interaction with AI systems, characterised by the presence and coverage of age-appropriate safeguards such as age assurance and content or contact restrictions; a graded property of a system's safeguards rather than a discrete event.
operationalDefinition: STATE. Operationalized as safeguard coverage: the proportion of a predefined minor-protection control set (age assurance, restricted content categories, contact/data limits, escalation paths) that is present and active in the system (0-1). It indexes implemented protection breadth, not a single intervention.
measurementSchema: Proposed measurement protocol (not yet empirically validated): Compliance checklist audit against the defined minor-protection control set; coverage could be reported as the proportion of controls verifiably active, with each control independently checked. Two auditors classify a sample; per-control present/absent agreement could be reported as Cohen's kappa.
author: Andreas Ehstand (ORCID 0009-0006-3773-7796)

## END

Total terms: 204
License: CC BY-NC-ND 4.0 — Attribution
Citation: https://zenodo.org/communities/augmanitai
Contact: ehstand.schule@gmail.com