Time-Lagged Activation Probe¶
Measurement probe — instantiates Activation Decay Measurement
Re-measures the same primed target at increasing delays after the cue, so you learn whether the effect lasts minutes, days, or only until the context shifts.
Everything downstream in this archetype needs numbers, and Time-Lagged Activation Probe is where the numbers come from. It goes back and re-reads the same primed target after real delays — not once, but on a schedule of widening intervals — to see how much of the effect is still there at each point. Its whole job is instrumentation: it fixes what "active" means, captures a fresh baseline immediately after the cue, picks a proxy that genuinely tracks the target, and then re-samples that proxy at, say, one day, one week, one month. The defining move — the one thing true of it and false of every sibling — is that it produces raw persistence observations across time and nothing else: no fitted curve, no threshold, no verdict. It answers only "how much is left at time t?" and hands the readings on.
Example¶
A company runs a thirty-minute anti-phishing training for staff. To find out whether the wariness actually lasts, security treats the training as a prime and instruments its decay. Immediately after training, a simulated phishing email goes out and 8% of trained staff click — that is the fresh baseline. Then, at deliberately spaced delays — three days, three weeks, three months — security sends fresh, equivalent simulated lures (a different template each time so the probe does not teach the answer) and records the click rate on each. The proxy is click-through on a matched simulated lure, chosen deliberately over a post-training quiz score because clicking behaviour tracks the real target — susceptibility — where a quiz tracks only stated knowledge.
The output is a small table: roughly 8% at day zero, 11% at three weeks, 19% at three months. That trail is the deliverable. The probe does not draw a curve through those points or set a retraining date; it hands the persistence observations to Decay Curve Fitting and stops there.
How it works¶
- Pin the target first. Name the specific primed state — here, susceptibility to a live lure, not "awareness" — so every later probe measures the same thing.
- Capture the baseline at t=0. Read the proxy immediately, while the prime is maximally fresh, to anchor "100% of the effect."
- Choose a proxy that co-moves with the target, and keep it lightweight enough to repeat cheaply.
- Schedule delays that resemble the real use window, widening the spacing so early fast decay and later slow decay both get sampled.
- Rotate parallel probe items so the act of probing does not itself re-teach the material.
Tuning parameters¶
- Interval spacing — logarithmic (dense early) vs. linear. Log spacing catches fast initial decay; linear is simpler but can miss the early cliff.
- Number of probe points — more points resolve the curve better but raise cost and the risk that probing itself alters the state.
- Proxy invasiveness — a covert behavioural read vs. an explicit test. Covert is truer to the target; explicit is cheaper but more reactive.
- Item equivalence — how tightly the rotating probe forms are matched. Loose forms add noise; identical forms leak answers.
- Longitudinal vs. cross-sectional — re-probing the same people (cheap, but each probe re-primes them) vs. a fresh sample per interval (clean, but costlier).
When it helps, and when it misleads¶
Its strength is that it converts "they were trained" or "we reminded them" into evidence a decision can stand on — it is the only sibling that touches ground truth about persistence. Its central failure mode is reactivity: a probe is itself a retrieval event, and retrieval strengthens memory, so probing the same people repeatedly quietly re-primes them and flattens the very decay you are trying to measure — the testing effect turned into a confound.[n1] The classic misuse is re-quizzing one cohort at each interval and concluding the effect is durable, when in fact the quizzes were unplanned refreshers. The guarding discipline is parallel forms, cross-sectional sampling where stakes justify it, and the lightest proxy that still tracks the target.
How it implements the components¶
activation_target_definition— names the precise primed state to be measured, so the proxy and every later probe stay aimed at one thing.baseline_activation_capture— the immediate post-cue read that fixes the reference level decay is measured against.observable_activation_proxy— selects and validates the observable signal that stands in for the latent primed state.delayed_probe_schedule— the widening timetable of re-measurements at realistic delays.
It produces observations, not a model: it does NOT implement decay_curve_model or activation_half_life_estimator — fitting a shape and a half-life to these points is Decay Curve Fitting's job, and it is this probe's nearest twin.
Related¶
- Instantiates: Activation Decay Measurement — this probe is the data-collection front end the whole archetype runs on.
- Sibling mechanisms: Decay Curve Fitting · Activation Window Thresholding · Refresh Cadence Adaptation · Contextual Reactivation · Decay Segment Comparison · Staleness Boundary Enforcement
Editorial Notes¶
Form Classification¶
Form family: Experiment, Test & Rehearsal
Rationale: Time-Lagged Activation Probe operates as an active test, trial, simulation, drill, or rehearsal that generates evidence through a deliberate attempt or perturbation because it re-measures the same primed target at increasing delays after the cue, so you learn whether the effect lasts minutes, days, or only until the context shifts.
Independent corroboration: The frozen evidence defines Time-Lagged Activation Probe as 'Re-measures the same primed target at increasing delays after the cue, so you learn whether the effect lasts minutes, days, or only until the context shifts', so its operative form is Experiment, Test & Rehearsal.
Nearest alternative: Monitoring, Sensing & Alerting — Time-Lagged Activation Probe includes features of ongoing observation, sensing, or alerting that detects and surfaces state without itself executing the response, but its defining operation is an active test, trial, simulation, drill, or rehearsal that generates evidence through a deliberate attempt or perturbation.
Review outcome: Independent reviewer agreement; medium confidence.
Origin Attribution¶
Primary origin: Psychology
Origin pattern: Single lineage
Present-day reach: Multi-domain
Rationale: The defining operation is: Re-measures the same primed target at increasing delays after the cue, so you learn whether the effect lasts minutes, days, or only until the context shifts. In the psychology lineage, that operation is specifically evidenced by authoritative or primary work that establishes time-lagged semantic priming effects by measuring target response after a preceding cue. This makes psychology the best historical origin, while the retained alternates document contributing methods and later applications rather than being mistaken for coequal origins.
Related originating lineages:
- Cognitive Science — Cognitive-science research on representation, learning, and recall supplies a parallel or contributing lineage for the mechanism's defining operation: re-measures the same primed target at increasing delays after the cue, so you learn whether the effect lasts minutes, days, or only until the context shifts.
- Data Science & Analytics — Data science's telemetry, modeling, profiling, and monitoring tradition provides a formative adjacent lineage for the same time lagged activation probe operation.
- Neuroscience — Neuroscience and memory-consolidation research supplies a parallel or contributing lineage for the mechanism's defining operation: re-measures the same primed target at increasing delays after the cue, so you learn whether the effect lasts minutes, days, or only until the context shifts.
- Statistics & Experimental Design — Statistics, experimental design, and measurement theory supplies a parallel or contributing lineage for the mechanism's defining operation: re-measures the same primed target at increasing delays after the cue, so you learn whether the effect lasts minutes, days, or only until the context shifts.
Review resolution: The blind reviewers disagree on primary lineage (statistics_experimental_design versus psychology), so I adjudicated the mechanism rather than inheriting either label. The defining operation is: Re-measures the same primed target at increasing delays after the cue, so you learn whether the effect lasts minutes, days, or only until the context shifts. In the psychology lineage, that operation is specifically evidenced by authoritative or primary work that establishes time-lagged semantic priming effects by measuring target response after a preceding cue. This makes psychology the best historical origin, while the retained alternates document contributing methods and later applications rather than being mistaken for coequal origins. The cited Meyer & Schvaneveldt, Facilitation in recognizing pairs of words directly supports the mechanism-specific operation and its disciplinary lineage. I retain all independently explained historical alternates without a numeric cap. origin_mode=single_lineage records how the mechanism arose; domain_reach=multi_domain separately records how broadly it can now be applied.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Researched adjudication after independent review; high confidence.
Sources consulted:
Notes¶
The probe is deliberately kept "dumb" — it fits nothing and decides nothing — precisely because it is the one place reactivity bites. Keeping it cheap and model-free is what lets it be run often enough, and with fresh items each time, to stay honest.
[n1] The testing effect (retrieval practice) is the well-established finding that the act of retrieving a memory strengthens it. It is a feature when you want durability, but a confound when you are trying to measure natural decay, because your measurement becomes an intervention. ↩