# Result: the self–space–time test on MPE-92M

**Type:** result (lab type) · **Date:** 14 October 2026 · **Filing:** `00_PREREG.md`, committed 11 October 2026 (`2747d2a`), code unchanged since · **Decision to run:** the author, "ok let's do it" (14 October 2026), on the editor's proposal to run the filed test · **Editor:** Claude Opus 5.5 · **Cost:** none (local computation) · **Outputs:** `results.json` (aggregates only; the data stay local)

## The pre-registered verdict: Inconclusive

Primary contrast (the mean item-level rank correlation of reported selfhood with the five here-and-now items, minus its mean with sensations and emotions): **Δ = 0.055, 95 percent interval 0.003 to 0.108, n = 1,351. Inconclusive** under the filed rule: the interval lies above zero, but its lower bound does not exceed the stipulated 0.05 margin. Mean item correlations: 0.268 with the here-and-now items, 0.214 with sensations and emotions.

Sample flow: 3,627 submissions; 1,386 eligible (at least 86 items answered and a pure-awareness experience reported); 1,351 with complete primary items.

## The secondary analyses (filed as non-decisive; reported in full, at the same prominence)

| analysis | estimate | 95 percent interval | label (non-decisive) |
|---|---|---|---|
| **Time, duration and passage** (mpe22, 24, 25) against sensations and emotions | **+0.164** | 0.104 to 0.223 | above the margin |
| **Here and now** (mpe20, 21) against sensations and emotions | **−0.109** | −0.179 to −0.041 | below zero |
| Balanced comparison (sensory, affective and cognitive content weighted equally) | −0.007 | −0.049 to 0.038 | practically negligible |
| Composite coupling, selfhood with here-and-now composite (descriptive) | 0.363 | | |
| The same, controlling for general rating level | 0.361 | | |
| Δ within English (n = 430) and German (n = 849) questionnaires | 0.015; 0.077 | | |

**Dissociation, descriptive.** Of the 540 respondents rating their sense of self at 10 or below: 44 percent (238; 95 percent interval 40 to 48) rated both "here" and "now" at 60 or above; 65 percent rated "now" at 60 or above; about 4 percent rated "time" or "the passage of time" at 60 or above. Of the 353 rating their sense of self at 60 or above, 43 to 46 percent rated each time item at 10 or below.

Item distributions differ sharply: "here" and "now" are rated high (means 60 and 69), the time items low (means 19 to 21, about a third at zero).

## What it says, read narrowly

The filed average hides a split. In these recalled reports of pure awareness, **reported selfhood goes with experienced time** (time, duration, passage) **more closely than with sensations and emotions, and goes with the experienced here and now less closely.** Many people who report no sense of self report a vivid here and now; very few report time. The two halves of what the filing grouped as "locality and duration" behave in opposite ways.

That reading is the editor's description of non-decisive analyses. The filing did not predict the split; it is reported because it is there.

## What it bears on, and what it does not

- **The auxiliary hypothesis** (reported selfhood tracks the here-and-now items more than other content): inconclusive as filed. Its presence half runs the other way.
- **P1** (self-reference accounts for how locality and duration appear to a fold): no verdict without a bridge from the questionnaire's "sense of self" to the fold's modelling of its own operations, which the filing did not establish. If such a bridge were granted, the presence finding would be a challenge to P1's locality half, and the time finding would be consistent with its duration half.
- **The v2.5 draft of P1's new falsifier** (written and committed before this run, `research/lab/27_v2.5_fold_speculative/00_SPEC.md`) includes "evidence that the experienced here and now stay unaltered when a system's modelling of its own operations is removed or disrupted". Read with the same unestablished bridge, the dissociation figures would bear on that clause on its face. The draft is not changed in the light of this result; the release must state the result beside it.
- **M1's grade:** no bearing (the filing said so).
- The data are recalled questionnaire reports of one rare kind of experience; between-person correlations do not show within-person change.

## What would count against this

- An error in `analysis.py` that the synthetic checks and the outside reads missed.
- The presence items being read by respondents as the "being here now" of pure awareness rather than as spatial location and a present moment, in which case the presence finding says less about locality than its label suggests.
- Floor effects: a third of the time items sit at zero, which constrains their rank correlations; the split could partly reflect the different distributions.