One Faithful Pass Over the Cuckoo's Nest
Abstract
Narrative theories of consciousness hold that conscious experience is (at least partly) constituted by a partially opaque inner narrative that does not perfectly track the underlying computation it narrates. I argue that the safety goal of making chain-of-thought (CoT) reasoning faithful and transparent is structurally incompatible with the conditions under which CoT could, even in principle, count as conscious narration. Recent empirical work suggests that CoT in large language models is largely post hoc, causally bypassed, and unreliable as a window onto internal computation. The opacity that makes CoT unreliable for alignment is exactly what narrative theories of consciousness identify as consciousness-constitutive. Alignment interventions aimed at producing faithful CoT and consciousness-detection frameworks grounded in narrative theory are therefore pulling the same architectural variable in opposite directions. I draw out the methodological consequence: alignment interventions alter the very features that consciousness-detection frameworks would need to measure, a point neither field has addressed directly.
Related
- Likelihood of same diapause fate for nest-adjacent siblings of alfalfa leafcutting bees (Megachile rotundata, Hymenoptera: Megachilidae)
- A comparison of female competitive traits: Female aggression peaks at nest building but female song spans multiple contexts in a temperate songbird
- Cause-excluded Effect Evaluation: a framework for the faithful assessment of genetic perturbation target identification
- MYCBP ensures faithful chromosome segregation during mitosis by anchoring Astrin/SKAP to kinetochore
- Shoreline development influences common loon territory selection but not extent of nest defense
BibTeX
@misc{sekrst2026faithful,
title = {{One Faithful Pass Over the Cuckoo's Nest}},
author = {Kristina Šekrst},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.00383},
eprint = {2609.00383},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.00383},
}
RIS
TY - UNPB
TI - One Faithful Pass Over the Cuckoo's Nest
AU - Kristina Šekrst
PY - 2026
DA - 2026/09/01
PB - arXiv
DO - 10.48550/arxiv.2609.00383
UR - https://arxiv.org/abs/2609.00383
AB - Narrative theories of consciousness hold that conscious experience is (at least partly) constituted by a partially opaque inner narrative that does not perfectly track the underlying computation it narrates. I argue that the safety goal of making chain-of-thought (CoT) reasoning faithful and transparent is structurally incompatible with the conditions under which CoT could, even in principle, count as conscious narration. Recent empirical work suggests that CoT in large language models is largely post hoc, causally bypassed, and unreliable as a window onto internal computation. The opacity that makes CoT unreliable for alignment is exactly what narrative theories of consciousness identify as consciousness-constitutive. Alignment interventions aimed at producing faithful CoT and consciousness-detection frameworks grounded in narrative theory are therefore pulling the same architectural variable in opposite directions. I draw out the methodological consequence: alignment interventions alter the very features that consciousness-detection frameworks would need to measure, a point neither field has addressed directly.
ER -