The behavioral knowledge graph
Every edge is a typed claim. Reinforces, counteracts, depends on, creates risk of, explains, generalizes, instance of — seven verbs that turn a list of concepts into a system you can reason over.
Edge types
All edges
Click either node to jump into it.
Biases like loss aversion bottom out in dopaminergic and amygdala circuitry — neuroscience supplies the mechanism BE describes.
Reward prediction error and threat response explain why heuristics fire the way Kahneman & Tversky catalogued them.
Strategic equilibria become feedback loops once you let the game iterate; both predict structural lock-in.
Complexity is what systems thinking becomes once agents adapt to each other in real time.
Emergent behavior is hard to reason about without knowing the topology that carries the interactions.
Cialdini's six principles are applied biases — persuasion is BE deployed as influence.
Influence techniques that exploit System 1 can erode the consent moral philosophy requires.
Classical economics assumes rationality; BE was built precisely to falsify that assumption.
Org psych supplies the empirical base — psychological safety, climate — that leadership models then operationalize.
Leadership that protects experimentation directly increases the rate at which novelty survives contact with the org.
Alignment is a multi-agent game; AI safety reduces to mechanism design once capability is high enough.
Modern AI is information compression at scale — Shannon's bounds shape what alignment can do.
SDT's needs map onto ancestral fitness signals: autonomy = agency, relatedness = coalition, competence = status.
Hyperbolic discounting is the temporal frame of present bias; both predict the same myopic choice pattern.
Influence travels through ties; persuasion's reach is set by network structure, not just message quality.
Feedback loops are signal channels; noise and delay determine whether the loop stabilizes or oscillates.
Every leadership decision is a value trade-off — moral philosophy is the layer that names which trade was made.
Cultures that meet SDT needs reliably outperform on the engagement constructs org psych measures.
Maslow, Graves, and the chakras are pre-empirical maps of the same underlying needs SDT later formalized.
The 20 questions only produce insight if you can trace second- and third-order effects through the system.
Negotiation is the most-played repeated game in human life — Fisher & Ury is applied Schelling.
Real innovation requires variation, selection, and retention — the same engine that drives complex adaptive systems.
Loss aversion is the canonical Prospect Theory finding; BE is the layer that names and measures it.
Amygdala activation to potential losses is roughly twice the ventral striatum response to equivalent gains.
Present bias is a temporal-discounting phenomenon — the future self is treated like a stranger.
Principal–agent is a textbook asymmetric-information game; the solution space is mechanism design.
Goodhart is what happens when a feedback loop optimizes a proxy rather than the underlying stock.
In-group preference is an ancestral coalition-detection adaptation, not a modern pathology.
We over-value what we own because giving it up registers as a loss, not a foregone gain.
Sticking with the default avoids the loss of changing — two biases pointing at the same behavior.
Discounting the future self also discounts the future cost of overruns we're predicting today.
Measuring agents on a proxy guarantees they will optimize the proxy instead of the principal's real goal.
Individuals' confirmation bias compounds inside cohesive groups, producing groupthink.
Safety to dissent is the single most studied antidote to convergence-driven group failure.
The overjustification effect only happens to behaviors that were intrinsically motivated to begin with.
Sunk-cost reasoning is the individual-level engine behind organizational escalation spirals.