33

Complex Cognition (Fach) / Task 8 (Lektion)

Diese Lektion wurde von maxmenke erstellt.

Lektion lernen

zurück  |  weiter 1 / 1

• Game Theory ð  analysis of interaction between rational agents that share common knowledge of rules §  Assumes common knowledge of rationality – assume rationality of other players §  However, people rarely act 100% rational
• Behavioral Game Theory study of how people actually behave in interactive situations
• Games = interactions between 2 or more agents, each of whom have 2 or more strategies available (which are associated with payoffs that depend on the action of the other person)
• Prisoner’s Dilemma P  Joint defection = nash equilibrium – combination of strategies that cannot be improved upon, as long as the other player sticks to the strategy §  As there is only one equilibrium point (in other cases there can be more), mutual defection is seen as the dominant strategy §  Not ideal but the best they can get if they are both rational actors pursuing their self-interest (according to game theory!) In reality people cooperate more often than game theory predicts
• Public Goods Game P  Best collectively rational strategy is to contribute Best individual strategy is not to contribute
• Dictator Game P  Two anonymous players are allotted a sum of money (= stake) in a one-shot interaction P  Player 1 decides how to divide the money between the two P  Player 2 receives allocation & game ends Purer measure of intrinsic motivation for equal offers
• Ultimatum Game P  Two anonymous players are allotted a sum of money (= stake) in a one-shot interaction P  Player 1 decides how to divide the money between the two P  Player 2 accepts/ rejects possible offers before hearing the actual offer If he/she doesn’t accept, no one gets anything
• Third-Party Punishment Game P  Two anonymous players are allotted a sum of money (= stake) in a one-shot interaction P  Player 1 decides how to divide the money between the two P  Player 2 has no choice P  Player 3 (gets additional 50% of stake), before hearing the actual offer, decides which offers to punish with what amount of money P  If punished, player 1 loses triple the amount payed by player 3
• Evolution of Cooperation P  Capacity for cooperation can evolve within a population P  Axelrod’s computer tournament: tested strategies on prisoner’s dilemma in an evolutionary sense (successful strategy = offspring) §  Nice rules begin by cooperating – were preyed upon by nasty strategies §  Nasty rules begin by defecting – died out when there was no prey left §  Most successful strategy was tit-for-tat (defection punished, cooperation rewarded) P  Crucial ingredient in the rise of more complex societies was the development of new social norms & informal institutions that are capable of domesticating our innate psychology for life in ever-expanding populations §  More complex societies prosper and spread to the degree that their social norms & institutions (= market exchange, world religions) effectively sustain successful interaction beyond local networks of kin & long-term relationships §  Internalised those norms as proximate motivations which then recalibrated our innate psychology to support large-scale cooperation P  Norm stabilising mechanisms: diffuse costly punishment & reputational mechanisms facilitate trust, fairness & cooperation §  With increasing community size less reputation more punishment to maintain fairness P  Support: fairness to anonymous others positively correlates with market integration & adherence to world religions
• Factors influencing Extent of Cooperation Consideration of others Empathy  Fear Differences in Values  Trust  Generosity  Communication
• Anzeige
• Consideration of others g  Less cooperation when it was known that the other cooperated (went from 37% to 16%) g  Perspective taking reduces cooperation g  Thinking harder about a task reduces cooperation (more likely to behave fairly when less capacity for thinking)
• Empathy g  Critical for social functioning & promotes prosocial behaviour g  Empathic failures – cases in which individuals could feel empathy but don’t due to salient social & psychological factors §  No accurate mentalisation about experiences of others/ may be correct in mentalising but fail to share state or feel no concern for those states g  Expression of empathic processes is influenced by psychological factors that tune one toward/away from empathy (norms, emotions prior to encounters) §  Interventions must diagnose specific nature & precursors of empathic failures and tailor interventions accordingly g  Empathy alone may be insufficient to produce pro-social behaviour (esp. when parties differ in status/power) Interventions should promote equitable goals & norms in addition to empathy
• Fear g  Either of being taken for a sucker or of being punished for acting in self-interest g  Fear of defection leads to defection, fear of free-riders leads to free-riding g  The latter results in more generosity §  Dictator game: people offer less when their offers can’t be rejected but still more than they would if they were behaving out of pure self-interest g  Once fear is removed people behave in a greedy fashion §  Greed has stronger effect on behaviour than fear g  Revenge effects (punishment motivated by anger)
• Differences in Values g  People can be categorised according to their social value orientation §  Prosocial – maximise joint gain & equality in outcomes (most upward, fewest downward counterfactuals g more cooperation) §  Individualist – maximise own gains (same number of each type) §  Competitors – maximise relative gain (= difference between one’s own and the other’s outcome) (reversed pattern as compared to prosocial) Values can be learned (e.g. economics students less cooperative due to training, gets less with school years)
• Trust g  Trust leads to increased cooperation g  More likely to trust people with good reputation g  Prisoner’s dilemma: more cooperation when play against the same person g  Emotions affect trust ratings of others (except for family members) Anger = more trust, gratitude/happiness = less trust
• Generosity g  Noise – erroneous response g when one cannot show trust to someone else due to circumstances they had no control over (e.g. car accident on the way to work) §  Can ruin successful TFT and end in a vicious circle of retaliation Can be solved by adding generosity or forgiveness
• Communication g  If one can’t be generous, communication helps g  Can eliminate detrimental effects of noise (esp. in people with low-trust dispositions) g  Increases cooperation because it promotes group solidarity & allows to express commitment to mutually cooperate
• Moralistic Punishment – fosters group cooperation as punishment of non-cooperators motivates them to cooperate
• Antisocial punishment P  – punishment of high cooperators destabilises group cooperation & reduces the beneficial effects of moralistic punishment (most often lowest target highest cooperators) §  Do-gooder derogation – people who help others get criticised & ridiculed for their efforts §  Prevents escalation of generosity §  Risky tactic as it can make oneself look bad if done poorly
• Normative theory : punishment of all deviators no matter which direction
• Biological markets theory (Barclay) P  – we choose partners for their cooperative interactions hence we try to outbid each other in order to be chosen (= competitive helping) §  Anti-social punishment do-gooder derogation function as a way to prevent one’s competitors from gaining relative reputation which could make oneself look worse by comparison §  Applies to any social dilemma or cooperative sitch §  May explain why generous people sometimes hide large donations §  Best strategy for social competition: do good but not too good P  Support: antisocial punishment much higher in market condition (observer saw overall average contributions of each player) than in the control condition (no observer) §  Also contributions didn’t increase in market condition §  Moralistic punishment stayed the same across conditions thus can’t be explained by retaliation
• Appraisal theory P  emotions are adaptive responses elicited based on how an agent evaluates its situation §  Appraisals occur in response to external stimuli and thoughts and are directly related to motivational goals of the agent §  Appraisal (= input of emotional experience) g action tendency (delineate behavioural output) §  E.g. guilt: failed to live up to expectations g reparative tendencies §  E.g. anger: progress toward goal is blocked g punishment/ revenge tendencies
• Psychological game theory P  appraisals associated with anger/guilt may be captured as changes in an agent’s expected payoff following a new event/outcome
• Guilt P  Feel guilty to the extent that a co-player receives less than he expected P  Guilt aversion – belief-dependent guilt enables cooperation §  Support: in trust game one returns the amount of money proportional to what they believe the co-player expected §  Neural activity: Increase in network thought to be involved in negative affect, cognitive control, theory of mind (yellow) P  When maximising financial self-interest: activity in regions involved in reward processing & mentalising (blue) Shows competing motivations to maximise material payoffs and to minimise aversive psychological payoffs
• Anger P  Arises when one experiences a worse outcome than initially expected §  Drives costly punishment P  Magnitude of deviation from expectations correlated with activity in network associated with error-monitoring & emotion Frustrated anger model – consequent feelings of anger motivate decisions to punish an offender
• Self-interest motivation widely presumed to drive most economic behaviour which aims to increase personal well-being
• Greed P  self-interest taken to such an extreme that, based on prevailing social norms, it may be perceived as unacceptable/immoral §  Central issue: what is the limit of acceptable self-interest? §  “no stopping rule” when excessive behaviour has no impact on others (e.g. binge eating wouldn’t be called greed) Studying economics lead people to hold more positive views of greed + ignore effects of their own behaviour on others thus more comfortably engage in greedy behaviour
• Mechanisms influencing greed related attitudes P  Dominant assumption of self-interested maximisation: Homo Economicus isn’t concerned with others’ welfare §  Deemphasises social consequences of one’s actions g reduced fairness concerns P  Game theory’s emphasis on a cool analytic approach to interpersonal behaviour §  Interpersonal emotions (empathy, guilt…) don’t fit within these frameworks §  Crowds out salient effects of social emotions which usually limit self-interested behaviour through strong internal feelings P  Creates belief that others are also motivated by self-interest §  Creates false consensus that leads them to justify their own self-interested behaviour g self-fulfilling prophecy Support for this g  Keep much more money to themselves g  Open-ended questions suggest greater self-focus & less adherence to norms of fairness g  Definitions of greed were less cognisant of the impact of greedy actions on others g more positive perceptions of greed g  Reading a positive text about self-interest changes peoples (non-economics) attitudes & opinions about greed g easy to imagine that studying such texts for years would have a strong impact §  Controls for self-selection
• Brain ð  Behaviours like altruism, trust & fairness are associated with activation in reward-related brain areas and the PFC responsible for cognitive control, processing of emotions & integration of costs & benefits to resolve the conflict between self-interest & other-regarding motives ð  Social preferences are similar to preferences of one’s own rewards in terms of neural activation (= domain general = no distinct social reward pathway)
• Cognitive Control P  vmPFC – crucial role in decision making involving social preferences §  Integration of costs & benefits §  Lesioned: reject offers in UG more frequently g costs of rejecting positive offers seem to weigh less §  More activation when unfair offers accepted g downregulates resentment associated with unfair offers P  ACC – conflict monitoring §  Activity consistent with trade-off between self-interest & prosocial motives P  Anterior insula – activity associated with degree of emotional resentment of unfair offers §  Stronger activity = more likely to reject an offer §  Negative emotional state in response to unfair treatment by social entity P  DLPFC – affects fairness related behaviours not judgement §  Controls emotional impulse to reject unfair offers Right: when out of order it increases acceptance rate of unfair offers
• Anzeige
• Reward-Related Areas P  Striatum – activated by anticipated reward, reinforcement learning & social reward §  More costly donations = more activation P  Nucleus accumbens – activated when there is reward value in observing the punishment of unfair partners P  VTA – activated when receiving money & non-costly donations g shows that giving has its own reward value Oxytocin – reduces fear of betrayal (= become more trusting) by dampening amygdala activation
• Cultural Differences P  Payoff to cooperation: cooperation stronger in societies where there is more cooperation with non-immediate kin P  Market integration: more cooperation in societies where there is greater trading via markets P  People in collectivistic might be more inclined to see others as outgroup members thus have more anti-social punishment
• Social Class Differences g  Upper class shows more unethical behaviour g  Relative independence & increased privacy in professions g fewer structural constraints & decreased perception of risk associated with committing unethical behaviour g  Availability of resources to deal with downstream costs increase likelihood of such acts g  Independent self-construal g feelings of entitlement, inattention to consequences of one’s actions g  Reduced concern for other’s evaluations g  More likely to become leaders g more likely to receive economics education g promotes values that justify & moralise positive beliefs about greed

zurück  |  weiter 1 / 1