The question
Across what timescales can human change occur, and why must the timescale of observation match the phenomenon being evaluated?
Definition
A timescale is the temporal span over which a specified process or pattern changes, unfolds or is organised for the question being asked; a phenomenon may have meaningful dynamics at several timescales simultaneously.
Human change has no single clock.
Some differences become visible within moments. Others emerge across repeated experiences. Some become interpretable only across sustained observation, while developmental change has to be located within much longer trajectories.
These are not fixed temporal classes. They are different ways of locating change in time.
A change can be real at one scale without establishing the same conclusion at another. A person may fluctuate substantially from hour to hour while remaining relatively stable across months. A behavioural pattern may shift across repeated occasions without yet providing evidence of long-term persistence. Developmental change can contain rapid transitions within it.
The key distinction is between three things: the timescale of the phenomenon, the way it is observed, and the temporal scope of the claim made from that evidence.
Timescale refers here to the temporal span over which a specified process or pattern changes, unfolds or is organised for the question being asked. A phenomenon may have meaningful dynamics at several scales at once. There is therefore rarely one universally correct timescale for something as broad as mood, behaviour or cognition. The relevant scale depends on what aspect of that phenomenon is being claimed to change.
Human Change Has No Single Clock
D1.5 established that change should be interpreted through time rather than reduced to one isolated moment. D1.6 adds a second requirement: the temporal scale must fit the specific change under examination.
Attention can shift rapidly. An emotional response may rise and fall across an episode. A skill may change across repeated practice. A behavioural pattern may require many observations before its direction becomes clear. Developmental and personality changes may need to be interpreted across much longer horizons.
There is no contradiction among these examples because they concern different temporal properties.
The question “How fast can people change?” therefore has no single answer. Some outcomes can change within very short intervals. Others can only be established through repeated evidence. A change in understanding may occur rapidly; a claim about enduring behavioural or developmental change requires a broader temporal frame.
“Short-term” and “long-term” should therefore remain relational terms. A month may be a substantial observation window for one process and inadequate for another.
The speed of a change also tells us little by itself about its importance, magnitude or durability. Rapid does not mean superficial, and slow does not mean deep.
Five Temporal Lenses
The Library uses five broad temporal lenses to locate claims about change. Their boundaries are intentionally porous. They organise questions about time rather than divide human change into five objectively bounded categories.
Momentary
The momentary lens examines change over a comparatively brief interval relative to the process and question under study.
It may capture changes in attention, affect, physiological responding, immediate behaviour or other fast-moving processes. There is no universal numerical cutoff. What counts as momentary depends on what is being observed.
A momentary difference is entirely real at that scale. It does not automatically establish persistence beyond it.
Episodic
The episodic lens examines how a specified aspect of functioning changes across a bounded interaction, task, session or episode.
A conversation, social encounter, emotional episode or practice session can have a temporal structure that is not visible from one isolated instant.
Here, episodic refers to bounded episodes of functioning, not to episodic memory.
Repeated-experience
The repeated-experience lens concerns patterns that become detectable across repeated exposure, practice, interaction or attempts.
Repetition itself does not guarantee change. Some patterns strengthen, some remain stable and some weaken. The point is that certain forms of change can only be evaluated across multiple occasions rather than within one episode.
Learning and skill development often require this temporal perspective.
Sustained-period
The sustained-period lens is a practical Library category for observations extending beyond isolated episodes or immediate conditions.
Its purpose is to examine whether a difference continues, recurs or changes across a more extended interval.
That interval has no fixed duration. What counts as sustained depends on the phenomenon being evaluated.
This lens can provide evidence of persistence, but persistence is not the same as the full canonical concept of lasting change.
Developmental
The developmental lens interprets change within longer trajectories of maturation, ageing, accumulated experience and changing developmental context.
Developmental does not simply mean “the longest” or “slowest” category. A developmental trajectory can include rapid processes, episodic changes and repeated experiences within it. Its defining feature is that the meaning of change is considered relative to a broader developmental course.
These five lenses are therefore arranged for conceptual clarity, not as a sequence through which change progresses.
A person may be stable across years while varying considerably from day to day. A developmental shift may change how someone responds within a single interaction. Repeated experiences may produce no sustained change at all.
The scales overlap.
Resolution and Observation Window Are Different
Timescale describes the temporal organisation of the phenomenon. Observation introduces a separate problem: how time is sampled.
Two concepts matter here.
Temporal resolution is the fineness with which an observation design can distinguish differences across time.
Observation window is the total span over which observations relevant to a claim are collected.
These are not the same.
Imagine one study that measures emotional experience every minute for two hours and another that measures it once per month for five years.
If the construct can be measured meaningfully at that frequency, the first design may reveal rapid fluctuations. Its temporal resolution is relatively fine, but its observation window is short.
The second may reveal broad longitudinal trends. Its observation window is much longer, but its sampling is too coarse to describe minute-to-minute variation.
Neither design is inherently superior. Each is suited to different questions.
Research using experience sampling and other intensive longitudinal methods shows how frequent repeated observations can reveal within-person variation that broader averages may obscure. Questions about developmental or longer-run change, by contrast, generally require evidence extending across the relevant longer interval.
The crucial point is that observation does not create the temporal structure of the phenomenon.
A person can fluctuate rapidly even if measured only once a month. Sparse sampling may simply fail to reveal that variability. Likewise, a short study may fail to establish persistence even when the observed change later proves durable.
The evidence available therefore depends partly on how time was observed, but the phenomenon itself is not reducible to the measurement schedule.
Fast and Slow Processes Interact
Processes observed at different timescales can coexist and interact.
Repeated faster events can sometimes contribute to slower change. Practice across many occasions, for example, can alter later performance. That later performance then changes what happens during subsequent practice.
The relation is not simple accumulation. Repetition does not guarantee a particular long-term outcome, and slower change cannot always be reduced to the sum of faster events.
Relations can also run in the other direction. Momentary responding may depend partly on learning, developmental organisation or recurring patterns established over longer intervals. What happens in a few seconds can therefore be influenced by processes whose history extends much further back.
Developmental and dynamic-systems research offers useful ways of conceptualising such multiscale interaction. D1.6 does not need a full mechanism theory. The relevant principle is narrower: processes operating over different temporal spans can constrain, contribute to and modify one another without collapsing into a single scale.
This is also why timescale should not be confused with level of explanation. Neural, behavioural, cognitive and interpersonal processes can each operate across more than one temporal scale.
Temporal scale and explanatory level are different dimensions.
The Main Cross-Timescale Errors
Many mistakes in interpreting change arise when evidence from one temporal scale is promoted into a claim about another.
Short-to-long inflation
A difference observed over a brief interval is treated as evidence of longer-term persistence.
A person's anxiety is lower during one session, and that difference is taken as evidence of a durable reduction in their usual anxiety. A new behaviour appears for several days, and lasting change is assumed.
The short-term evidence may be valid. It establishes a difference within the interval observed. Claims about longer persistence require evidence extending into the relevant longer window.
Persistence is not movement up a temporal ladder. It is evidence that a difference continues or recurs across an appropriately extended interval.
Long-to-short erasure
The opposite mistake occurs when a stable long-run summary is used to dismiss shorter-term variability.
A person's average mood may remain similar across several months while varying considerably from hour to hour or day to day. Two people can have similar averages while showing very different patterns of fluctuation across time.
A coarse summary can therefore be stable while finer temporal dynamics remain highly variable.
Human functioning can be stable at one timescale and changing at another without contradiction.
Resolution mismatch
Sometimes the observation design is poorly matched to the process being claimed.
Monthly measurements cannot directly describe minute-to-minute fluctuations. Dense observations across one day cannot establish a multi-year trajectory.
The problem is not simply low frequency. A design can be too sparse, too brief or otherwise poorly suited to the temporal structure of the claim.
More measurement is not automatically better. Appropriate temporal design is the goal.
Scale reduction
A difference observed at one timescale is sometimes treated as a complete account of change at another.
A rapid neural difference, for example, may establish that something changed neurally over that interval. It does not by itself establish durable behavioural change.
Likewise, a long-run behavioural trend does not eliminate the relevance of shorter-term processes through which the pattern is produced.
Empirical claims linking change across temporal scales require evidence appropriate to that relationship.
Match the Timescale to the Claim
D1.6 can therefore be reduced to one canonical epistemic rule: match the timescale of the evidence to the timescale of the claim.
Three questions keep that rule precise.
First, what is changing? A momentary affective response, a skill, an interpersonal pattern and a developmental characteristic are different temporal objects.
Second, over what timescale is the relevant process or pattern being claimed to change? The same broad construct may show meaningful dynamics at several scales, so the claim must specify which one matters.
Third, does the evidence resolve and observe that timescale adequately? Appropriately fine observation over a short interval can reveal rapid dynamics. Extended observation can reveal persistence and broader trajectories. Neither substitutes automatically for the other.
This also clarifies questions such as “When does change become stable?” There is no universal duration after which stability appears. Stability must be evaluated relative to the phenomenon and observation scale. Longer observation can provide stronger evidence of persistence, but elapsed time alone does not establish the stronger criteria for lasting change.
The five temporal lenses are therefore not stages, and no timescale is inherently more meaningful than another. Different scales reveal different aspects of the same changing system.
Human change has no single clock. To interpret it accurately, the temporal scale of the phenomenon, the observation and the claim must be kept aligned.