01
ObserverA person in the present
Four operations:
- Recognize emotions
- Identify arguments
- Explore perceptions
- Dismantle behaviors
The shapes of uncertainty
A multidisciplinary search for managing the complexity of subjective definitions of reality.
Uncertainty is not a gap in your data. It is the condition in which every strategic decision is made. This page explains how people approach it, how AI systems now handle it, and why both tend to close it too soon.
A path driven by one persistent problem
Each field answered part of the question. None answered all of it.
How to think and act when the future cannot be treated as a linear continuation of the present.
In nonlinear markets, causes and effects stop lining up, and a plan built for stable conditions keeps looking right until it isn't. The question first appeared in organizations, then moved to how people reason and to the stories that shape which futures a company can imagine. Since 2022 it appears in the language models those companies now use.
Uncertainty and complexity: the disciplines in an integrated model
02
Observer · LimitsUncertainty · limits
Uncertainty belongs to the observer and is marked by limits.
- Cognitive structureFrancisco Varela
- Emotions and feelingsAntonio Damasio
- LanguageLudwig Wittgenstein
- Bifurcation pointIlya Prigogine
- Systemic constructivismPaul Watzlawick
- NeurobiologyMichael Gazzaniga
- ScenariosLisa Feldman Barrett
03
System · LimitationsComplexity · limitations
Complexity belongs to the system and is marked by limitations.
- Dissipative structuresIlya Prigogine
- AttractorsEdward Lorenz
- Wave functionNiels Bohr
- Wave–particle dualityLouis de Broglie
- Butterfly effectEdward Lorenz
04
FuturesPossible states
From the present, trajectories branch and cross toward possible futures.
- Predictive horizonHubert Reeves
- AutopoiesisHumberto Maturana and Francisco Varela
- Operational closureNiklas Luhmann
- Wave-function collapseNiels Bohr
Uncertainty and Complexity: The Disciplines in an Integrated Model
DOI 10.13140/RG.2.2.34338.45763
Premature closure and epistemic monoculture
Better prompts improve the output, not its kind. What is needed is not a better answer, but a different class of answer.
The dominant design pattern in applied language-model systems involves a single inferential regime: one prompt, one model, one analytical pass.
When every vendor, every team, and every agent runs the same pattern, you get the same answers as your competitors, only faster. The risk is not a wrong answer. It is an industry that can only see what one way of reasoning makes visible.
The unexplored space
Under standard instruction, models close the space of possibilities before exploring it.
Inside that closure, reasoning improves, but the type of knowledge produced does not change.
Monoculture across systems
A single inferential regime can only produce what that regime makes visible.
The convergence is not mainly a limit of model capability. It is an architectural consequence of single-regime processing.