Why Does Consciousness Ignite? A Bold Guess Based on Dynamic Resource Scheduling
Zenodo (CERN European Organization for Nuclear Research) July 28, 2026 preprint DOI: 10.5281/zenodo.21639567 (opens in new tab)
Study at a glance
AI-extracted from the abstract| Characteristics | Theoretical or philosophical paper |
|---|---|
| Key points | Argues that consciousness ignition may be a dynamic computational resource scheduling process in a limited-resource brain, rather than a simple binary information-entry event. Proposes that prediction error, value, urgency, risk, outcome modifiability, and an internal value model determine how much computational resource is invested, yielding conscious states of different intensities, and hypothesizes that dopamine and other neuromodulatory signals may submit importance alerts. The author states these mechanisms have not been proven. |
Abstract
Consciousness ignition is usually described as a whole-brain integration process after information enters the global workspace. But this paper offers a bold guess from the perspective of a limited-resource system: consciousness ignition may not just be an information-entry process, but a dynamic computational resource scheduling process. This paper argues that although the brain possesses powerful computing ability, it is not an infinite computing system. Therefore, the brain may need to continuously judge: which information is worth investing computational resources in, and how much should be invested. Based on this guess, this paper proposes several variables that affect conscious resource investment, including prediction error, value, urgency, risk, outcome modifiability, and individual internal value model. This paper further guesses that consciousness ignition may not be a simple binary process of “happens or doesn’t happen,” but more likely a continuous process of resource allocation. The system dynamically adjusts the level of computational resource investment according to the importance of different information, thereby producing conscious states of different intensities. On this basis, this paper puts forward an even bolder guess: dopamine and other neuromodulatory signals may not only participate in prediction-error-related processes, but may also participate in the discovery of “surprises,” and submit importance alerts to the system. Subsequently, the resource scheduling system evaluates whether to increase computational resource investment based on value, risk, urgency, and whether the outcome can still be changed, forming different degrees of consciousness ignition. This paper does not claim that the above mechanisms have been proven. It simply tries to propose a computational architecture perspective: consciousness may not be an automatic response to all information, but a dynamic choice made by a limited-resource system regarding “what is worth continuing to compute” and “how much computing power it is worth.”