# Hypothesis

## Primary hypothesis

`HYPOTHESIS`

Electronic devices impose avoidable cognitive context switching when they require people to translate creative intent into filenames, folder structures, version numbers, storage decisions, and recovery procedures.

This burden can interfere with creative flow and collaboration. It may have a disproportionate effect on people with ADHD and other cognitive or learning disabilities, but PIXIE does not assume diagnosis, impairment, or a universal experience.

## Design proposition

`PROPOSED`

A device steward can reduce that burden when it:

1. asks what the person is trying to accomplish in ordinary language;
2. derives technical details from available context;
3. shows a plain-language preview;
4. performs only approved, reversible actions;
5. records what changed and how to undo it; and
6. improves suggestions from explicit corrections without silently changing the person's rules.

## Originating observation

`OBSERVED / INTERPRETIVE`

The initial insight emerged during the *50 Ways to Leave Another* production process, when a collaborator-supplied video showed practical difficulty with creative-file naming. Dominique Devereaux interpreted the interruption as a systems-design failure: the software was asking the creator to perform administrative translation that the software could handle.

The source person's experience is evidence of an originating case—not proof of the general hypothesis, endorsement of PIXIE, consent to participate, or evidence of a medical condition.

## Research questions

1. Which file and device-management steps most often interrupt creative work?
2. Which information can software derive reliably without asking the person?
3. Which actions must always remain explicit choices?
4. Does natural-language retrieval reduce time-to-find and duplicate work?
5. Does a reversible preview increase trust without increasing cognitive load?
6. How do needs differ across ADHD, dyslexia, autism, brain injury, fatigue, aging, and people without identified disabilities?
7. What observable interaction signals indicate interface friction without claiming to reveal emotion or diagnosis?

## Non-claims

PIXIE does not currently claim that it:

- improves mental-health outcomes;
- detects depression, anxiety, mania, crisis, or cognitive decline;
- knows what a person feels from device behavior;
- can safely make autonomous deletion or clinical decisions;
- represents all neurodivergent people; or
- has been endorsed by any named collaborator or platform provider.

## Success criteria

The hypothesis gains support only if consented testing shows measurable improvements such as:

- fewer questions required to save or recover work;
- reduced time to locate a known file;
- fewer accidental duplicates;
- successful return after interruption;
- accurate recovery from incorrect suggestions;
- increased user-reported control and confidence; and
- no material increase in privacy, loss, or unwanted-action incidents.