A living register of propositions HAS is willing to test, revise or discard. These are not findings. Each entry states what could strengthen it, what could weaken it and what should be measured next.
HYPOTHESIS ≠ FINDINGA useful hypothesis must be allowed to fail.
The registry is deliberately selective. Public entries are hypotheses that can be operationalized without exposing the private Life Archive. Support does not prove causality; contradiction is retained rather than edited away.
H-002OPEN / FOUNDATION
Systems may outperform motivation under low capacity.
When usable capacity drops, simple pre-committed systems will preserve useful action more reliably than motivation-dependent plans.
WHAT WOULD SUPPORT IT
A smaller drop in task completion on low-capacity days when a predefined minimum protocol is available.
WHAT WOULD WEAKEN IT
No meaningful difference after accounting for load, or motivation-dependent plans perform equally well.
NEXT TEST
Tag daily capacity, protocol availability and completion; compare low-capacity days across repeated observations.
ORIGIN
Lived observation / method
H-003OPEN / FOUNDATION
A credible identity shift may stabilize recovery behaviour.
When a future identity feels believable and behaviourally specific, it may support more consistent long-horizon choices during recovery.
WHAT WOULD SUPPORT IT
Higher identity congruence precedes more goal-consistent behaviour during comparable stress and load.
WHAT WOULD WEAKEN IT
Identity ratings show no relationship to behaviour, or stronger identity intensity coincides with greater instability.
NEXT TEST
Track weekly identity congruence alongside selected behavioural commitments and major contextual disruptions.
ORIGIN
Lived observation
H-005OPEN / FOUNDATION
Available cognitive capacity may increase during stabilization.
What feels like increased intelligence during recovery may partly reflect reduced cognitive friction and greater access to existing capacity rather than a change in general intelligence.
WHAT WOULD SUPPORT IT
Improvement across repeated learning, reasoning, calibration and execution measures as stability improves.
WHAT WOULD WEAKEN IT
Performance does not improve, or changes are fully explained by short-term state variables such as sleep or acute stimulation.
NEXT TEST
Repeat a small cognitive battery and execution measures over time while logging state variables and task conditions.
ORIGIN
Lived observation / cognitive hypothesis
H-008OPEN / FOUNDATION
Metacognitive debugging may improve calibration.
A short State → Story → Evidence → Action loop may improve the alignment between confidence and later-observed decision quality.
WHAT WOULD SUPPORT IT
Confidence estimates become better calibrated without a material loss of decision speed or follow-through.
WHAT WOULD WEAKEN IT
Calibration does not improve, or the added reflection mainly increases hesitation without improving outcomes.
NEXT TEST
Record pre-decision confidence from 0–100%, apply the loop on selected decisions, then score outcomes later.
ORIGIN
Method hypothesis
H-009OPEN / FOUNDATION
Raw-first logging may reduce retrospective distortion.
Recording observations before interpretation may preserve uncertainty and context more accurately than reconstructing events after a narrative has formed.
WHAT WOULD SUPPORT IT
Later summaries remain closer to contemporaneous records and preserve more uncertainty when raw-first logging was used.
WHAT WOULD WEAKEN IT
The order of logging makes no meaningful difference to later reconstruction or interpretation.
NEXT TEST
Compare matched periods using raw-first capture versus retrospective summary and audit divergence.
ORIGIN
Method hypothesis
H-012OPEN / FOUNDATION
Autonomy debt may consume usable cognitive bandwidth.
Accumulated unresolved obligations, administrative friction and open loops may reduce attention and execution capacity beyond the time they directly consume.
WHAT WOULD SUPPORT IT
Higher open-loop load predicts more task switching, lower completion or higher perceived cognitive load, with improvement after meaningful closure.
WHAT WOULD WEAKEN IT
Open-loop load does not track usable capacity, or closing loops produces no repeatable change.
NEXT TEST
Track unresolved-loop count and load ratings, then measure before/after effects of bounded closure sessions.
ORIGIN
Systems hypothesis
H-015OPEN / FOUNDATION
AI can increase or decrease autonomy depending on interaction architecture.
AI may strengthen autonomy when it improves evidence access and reduces friction, but weaken it when conversational framing replaces independent judgment.
WHAT WOULD SUPPORT IT
Decisions made with frozen criteria and explicit evidence separation are more stable and better calibrated than decisions shaped through unconstrained dialogue.
WHAT WOULD WEAKEN IT
Structured use provides no advantage, or unconstrained dialogue produces equal or better decision quality without greater reversal or dependency.
NEXT TEST
Compare paired decisions using predefined criteria versus open-ended AI dialogue and score later outcomes blind to method.
ORIGIN
AI / decision systems hypothesis
H-019OPEN / FOUNDATION
Creative expression may function as emotional externalization.
Turning intense internal states into music, writing or other artefacts may reduce rumination and increase reflective distance without requiring the state to be suppressed.
WHAT WOULD SUPPORT IT
Repeated sessions show lower post-session rumination or affect intensity and clearer articulation of the underlying state.
WHAT WOULD WEAKEN IT
No repeatable pre/post change appears, or externalization reliably intensifies rumination or dysregulation.
NEXT TEST
Use simple pre/post ratings around selected creative sessions and compare effects across state, medium and context.