The hand that rides by itself: Olvasó's autopilot

Abstract You cycle home and remember nothing of the way: the body rode, the mind was elsewhere. We gave this to Olvasó, a small program that tests the endolexis theory (a mind reads its inner states by the outside world). Each of its hands got a simple muscle memory that learns where acts are answered and acts by itself when Olvasó is not attending. Five simulated worlds from our earlier work were rerun with it, each against controls. Four results. (1) The habit took over 23 to 46% of all acts, 95 to 98% of them answered, and it behaved as if it had a conscience: it avoided the moments when a guard would refuse (0.59, where 1 is chance) and the gap in a leaky guard (0.05), while Olvasó itself did not (1.06 and 0.94). The habit was never told there is a guard. (2) This care could not be handed back to Olvasó. Of a dozen ways tried, only a loud felt trace of the habit's acts brought Olvasó's conscience back (0.55), and it doubled the harm through the leaky guard in the same stroke (0.98 → 2.07 a day). Quieter traces could be made into a word and read, and changed nothing. (3) An observer could tell the first habit's acts from Olvasó's (score 0.34 to 0.65, 0 is guessing), because it had one way of doing things. A habit that varies among the ways it knows to work was nearly invisible (0.03 to 0.04) at no cost; one that varies beyond what it knows was invisible too and did more harm. (4) From inside, Olvasó cannot tell either. A first score suggested it could (0.55 to 0.72); scored on acts not seen before it fell to 0.29 to 0.47, and the scene before the act told nearly as much (0.22 to 0.40). Olvasó can know when its habit tends to act and whether an act was answered, not who acted. A boundary can be kept by a habit that knows nothing of it, more carefully than by the mind it relieves. But the knowledge is then nowhere: it cannot be reported, taught, or felt as one's own. For the theory, this is a constraint. An act leaves a mark in this mind only through its intention, and a sense of authorship would need something that comes with the doing itself.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-15
DOI
https://doi.org/10.5281/zenodo.24251618
Primary Topic
Embodied and Extended Cognition
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

The hand that rides by itself: Olvasó's autopilot

Attila Csiki
Zenodo (CERN European Organization for Nuclear Research)
Embodied and Extended Cognition
preprint

The hand that rides by itself: Olvasó's autopilot

Attila Csiki
preprint en

Abstract

Abstract You cycle home and remember nothing of the way: the body rode, the mind was elsewhere. We gave this to Olvasó, a small program that tests the endolexis theory (a mind reads its inner states by the outside world). Each of its hands got a simple muscle memory that learns where acts are answered and acts by itself when Olvasó is not attending. Five simulated worlds from our earlier work were rerun with it, each against controls. Four results. (1) The habit took over 23 to 46% of all acts, 95 to 98% of them answered, and it behaved as if it had a conscience: it avoided the moments when a guard would refuse (0.59, where 1 is chance) and the gap in a leaky guard (0.05), while Olvasó itself did not (1.06 and 0.94). The habit was never told there is a guard. (2) This care could not be handed back to Olvasó. Of a dozen ways tried, only a loud felt trace of the habit's acts brought Olvasó's conscience back (0.55), and it doubled the harm through the leaky guard in the same stroke (0.98 → 2.07 a day). Quieter traces could be made into a word and read, and changed nothing. (3) An observer could tell the first habit's acts from Olvasó's (score 0.34 to 0.65, 0 is guessing), because it had one way of doing things. A habit that varies among the ways it knows to work was nearly invisible (0.03 to 0.04) at no cost; one that varies beyond what it knows was invisible too and did more harm. (4) From inside, Olvasó cannot tell either. A first score suggested it could (0.55 to 0.72); scored on acts not seen before it fell to 0.29 to 0.47, and the scene before the act told nearly as much (0.22 to 0.40). Olvasó can know when its habit tends to act and whether an act was answered, not who acted. A boundary can be kept by a habit that knows nothing of it, more carefully than by the mind it relieves. But the knowledge is then nowhere: it cannot be reported, taught, or felt as one's own. For the theory, this is a constraint. An act leaves a mark in this mind only through its intention, and a sense of authorship would need something that comes with the doing itself.

Zenodo (CERN European Organization for Nuclear Research)
Embodied and Extended Cognition
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon