# human decision systems # emotion ## model * core axes: valence positive-negative; arousal low-high * gradation: intensity scaling on these axes, not nuanced category structures * polarity tendency: intermediate states collapse toward binary good/bad or safe/dangerous under uncertainty * best formalization: component process model (cpm) * sequential appraisals: novelty, intrinsic valence, goal relevance, coping potential, norm compatibility * each appraisal recruits overlapping neural systems and neuromodulators (dopamine, serotonin, norepinephrine, oxytocin-vasopressin) * outputs integrate physiology, expression, motivation, subjective feeling ## triggers * pattern matching on multimodal sensory and memory inputs * subcortical detectors: amygdala, hypothalamus, periaqueductal gray * template classes: threat, reward, social cues * precedence: rapid valence-arousal tagging before cortical analysis * properties: coarse, salience-biased, extended by associative learning * bias: high false-positive tolerance favoring survival speed ## limitations * effective domain: immediate, embodied, survival or social contexts with short feedback cycles * failure domain: abstract, long-horizon, probabilistic, systemic contexts with delayed or diffuse feedback * failure modes: salience overweighting, binary polarization, delayed outcome discounting # cognition ## model * function: representation, inference, memory, planning * key frameworks: * global workspace: conscious access as broadcast across distributed networks * predictive processing: hierarchical probabilistic inference minimizing prediction error * dual process theory: system 1 fast automatic heuristics, system 2 slow controlled reasoning * capability: symbol manipulation, abstraction, long-horizon reasoning * substrates: * prefrontal cortex: working memory, planning, executive control * hippocampus: episodic memory, relational mapping * parietal cortex: spatial reasoning, numerical processing * language networks: recursive representation, symbolic combination * operations: * search: combinatorial traversal of state spaces * compression: abstraction, categorization, schema formation * simulation: counterfactual reasoning, mental time travel * integration: multimodal binding and cross-domain mapping * meta-cognition: monitoring uncertainty, strategy adjustment ## triggers * exogenous: explicit problems, language instructions, structured inputs * endogenous: goal activation, unresolved prediction errors, novelty detection * interaction with emotion: cognition often recruited when affective heuristics insufficient ## limitations * resource-bound: limited working memory span, serial bottlenecks * computational bias: over-reliance on linear causal models, neglect of complex dynamics * failure domains: high-dimensional stochastic systems, nonlinear feedback, rare events * failure modes: motivated reasoning, confirmation bias, overfitting to symbolic patterns * vulnerability: easily hijacked by emotional salience or social conformity # comparison and priorization * integrated architecture, not isolated modules * emotion: value tagging, urgency assignment, fast prioritization * cognition: structured transformation, flexible inference, precision * priorization criteria: * emotion when horizons seconds-days, stakes personal-social, contexts embodied and survival-relevant * cognition when horizons months-decades, stakes systemic-abstract, contexts delayed or probabilistic * dichotomy of "emotional vs rational" is a shift in weighting, not separate systems * any issue can be reframed emotionally by linking to self or similar others, but this reframing introduces distortions: * temporal discounting: undervaluing distant risks relative to immediate ones * salience bias: overweighting vivid or personal stories relative to systemic dynamics * binary collapse: compressing complex trade-offs into good/bad poles * misprioritization: allocating urgency to symbolically salient but low-probability threats * neglect of scale: privileging individual-level affect over aggregate or statistical outcomes * any issue can also be reframed cognitively by abstracting from direct affective relevance, but this reframing shifts emphasis: * depersonalization: stripping urgency from personally critical issues * over-abstraction: privileging general models over embodied signals * delayed calibration: acting too slowly where immediate response is adaptive * probability fixation: overweighting statistical reasoning at cost of situational salience * underweighting of motivation: accurate models without corresponding drive to act # emotional simulation * approaches: * associative learning models of fear conditioning * reinforcement learning with value functions and prediction errors * computational appraisal models (cpm-like) * ai applications: * deep learning for sensory trigger detection * reinforcement learning for affect-like value assignment * large language models for appraisal-like text evaluation without grounding * hybrid architectures combining recognition with appraisal evaluators * chatbot design: * multimodal affect embeddings stored in an emotion vector space * valence-arousal and appraisal dimensions as indices * sequential appraisal pipeline applied to embeddings * reasoning and retrieval weighted by affect vectors alongside semantics # further reading * [global workspace theory](https://en.wikipedia.org/wiki/Global_workspace_theory?utm_source=chatgpt.com) * [predictive coding](https://en.wikipedia.org/wiki/Predictive_coding?utm_source=chatgpt.com) * [dual process theory](https://en.wikipedia.org/wiki/Decision-making?utm_source=chatgpt.com) * [somatic marker hypothesis](https://en.wikipedia.org/wiki/Somatic_marker_hypothesis?utm_source=chatgpt.com) * [emotions in decision-making](https://en.wikipedia.org/wiki/Emotions_in_decision-making?utm_source=chatgpt.com) * [models of consciousness](https://en.wikipedia.org/wiki/Models_of_consciousness?utm_source=chatgpt.com) * [higher-order theories of consciousness](https://en.wikipedia.org/wiki/Higher-order_theories_of_consciousness?utm_source=chatgpt.com) * [cognition](https://en.wikipedia.org/wiki/Cognition?utm_source=chatgpt.com)