The Encryption Theory of the Evolution of Humor: Honest Signaling for Homophilic Assortment Thomas Flamson, Ph.D. UC Davis ~ Anthropology IBNeC / HBES Gramado, RS 2 September 2015
Variation & Assortment
Variation & Assortment
Ethnic Markers Barth 1969; McElreath, Boyd, & Richerson 2003
Between-Group Assortment
Between-Group Assortment
Between-Group Assortment
Within-Group Assortment
Homophily
Complementarity
Large-Scale
Avoiding Burning Bridges
The Encryption Theory of Humor Humor is an adaptation for honestly signaling compatibility in local cultural variation among potential long-term interaction partners Two Levels of Analysis: Cognitive: Humor is understood by means of an encryption-decryption process Social: Successful humor entails an inference of similarity, encouraging affiliation with compatible local peers
The Encryption Theory of Humor Humor is an adaptation for honestly signaling compatibility in local cultural variation among potential long-term interaction partners Two Levels of Analysis: Cognitive: Humor is understood by means of an encryption-decryption process Social: Successful humor entails an inference of similarity, encouraging affiliation with compatible local peers
Sample Cartoon
Humor As Encryption Flamson & Barrett 2008; Flamson & Bryant 2013
Humor As Encryption Ha ha ha! Flamson & Barrett 2008; Flamson & Bryant 2013
Model of Within-Group Signaling Smaldino, Flamson, & McElreath In Prep.
Model of Within-Group Signaling Smaldino, Flamson, & McElreath In Prep.
Communication Phase Smaldino, Flamson, & McElreath In Prep.
Overt Communication Smaldino, Flamson, & McElreath In Prep.
Covert Communication Smaldino, Flamson, & McElreath In Prep.
Assortment Phase??? Smaldino, Flamson, & McElreath In Prep.
Choosy Assortment Smaldino, Flamson, & McElreath In Prep.
Choosy Assortment 1+α+β 10 Smaldino, Flamson, & McElreath In Prep.
Dire Assortment 1+α+β 10 Smaldino, Flamson, & McElreath In Prep.
Frequency of Dire Situations (δ) Covert Outperforms Overt Under Most Conditions Group Size (m) Fitness Covert > Fitness Overt Proportion of Signals Received (r) (Other Parameters: Similarity (s) = 50% ; Similarity Benefit (α) = 1 ; Liking Benefit (β) = 1)
Frequency of Dire Situations (δ) Even With Noisy Signals Probability of Understanding Covert Signal (γ) Fitness Covert > Fitness Overt Proportion of Signals Received (r) (Group Size (m) = 10; Similarity (s) = 50%; Similarity Benefit (α) = 1; Liking Benefit (β) = 1)
Frequency of Dire Situations (δ) And With Extreme Benefits In Overt s Favor Probability of Understanding Covert Signal (γ) Fitness Covert > Fitness Overt Size of Similarity Benefit (α) (Reception (r) =.7; Group Size (m) = 10; Similarity (s) = 50%; Liking Benefit (β) = 1)
Conclusions Covert signaling is an effective strategy for within-group assortment Benefits of homophilic assortment + Continued interaction with remaining group members Particularly as Group size increases Information transfer becomes more reliable Covert signal noise decreases Benefits of homophily are moderate Smaldino, Flamson, & McElreath In Prep.
Humor and Personality Extraversion: preference for sexual & incongruity-resolution humor, production of affiliative humor Agreeableness and openness: production of affiliative humor Sensation-seeking: preference for nonsense humor Neuroticism: rejection of nonsense humor Aggressiveness/Type A: preference for aggressive humor Flamson In Prep.
Humor and Attitudes Conservatism: preference for incongruityresolution humor, rejection of nonsense & sexual humor, production of aggressive humor Religiosity: less humor production, rejection of sick humor, production of aggressive & self-defeating humor Prejudices: preference for derogatory or sexist jokes correlate with implicit preferences Flamson In Prep.
Humor and Compatibility Similarity in humor preferences: increases evaluations of strangers & altruistic behavior, correlated with social closeness & romantic relationships Humor priming: creates feelings of closeness with strangers, enhances group cohesion & performance Humor use: more used with closer friends, correlated with romantic relationship satisfaction, more frequent in groups with greater cognitive similarity Flamson In Prep.
Conclusions Humor preferences and production correlate widely with a variety of personality traits and attitudes Similarity in sense of humor enhances interpersonal attraction Humor positively tied to strength of interpersonal relationships and group performance Humor is strongly linked with within-group variation and homophilic compatibility Flamson In Prep.
Brazil Map Ceará Gramado
Assentamento Map
Johnson 1971 Pictures
MST Pictures
Antônio Family Picture
Chapel Picture
Board Meeting Pictures
Coletivo Pictures
Friendly Favors Pictures
Humor Preferences Study Assessments of prior knowledge Ratings of jokes randomly preceded by Irrelevant (High Encryption) or Relevant (Low Encryption) information Predictions: 1) Prior knowledge Higher joke ratings 2) Prior knowledge High > Low No prior knowledge Low > High Flamson & Barrett 2013
Humor Preferences Study Flamson & Barrett 2013
Online Humor Preferences Studies Study 1: Popular jokes Study 2: Obscure jokes Flamson & Barrett 2008
Social Assortment & Humor Preferences Social networks measured by ranking all other assentados by closeness Ratings task participants coded by average closeness per dyad Computed difference in joke ratings per dyad ( S1 S2 ) Prediction: Greater social closeness Lower difference in joke ratings Flamson & Barrett 2013
Social Closeness Predicts Similarity Knowledge Level Encryption Level Effect on Difference in Ratings Flamson & Barrett 2013 Average Social Closeness Ranking -0.026* Both Had Prior Knowledge -0.019 Both Did Not Have Prior Knowledge 0.018 Both Had High Encryption Jokes 0.026 Both Had Low Encryption Jokes -0.023 Transformed Age Difference 0.018 Practice The Same Religion -0.089* Have Different Nativeness Statuses -0.001 β Related -0.059 Live In The Same Vila -0.037 Mixed-effects restricted maximum-likelihood regression, * = p < 0.05
Social Closeness Predicts Similarity Knowledge Level Encryption Level Effect on Difference in Ratings Flamson & Barrett 2013 Average Social Closeness Ranking -0.026* Both Had Prior Knowledge -0.019 Both Did Not Have Prior Knowledge 0.018 Both Had High Encryption Jokes 0.026 Both Had Low Encryption Jokes -0.023 Transformed Age Difference 0.018 Practice The Same Religion -0.089* Have Different Nativeness Statuses -0.001 β Related -0.059 Live In The Same Vila -0.037 Mixed-effects restricted maximum-likelihood regression, * = p < 0.05
Social Closeness Predicts Similarity Knowledge Level Encryption Level Effect on Difference in Ratings Flamson & Barrett 2013 Average Social Closeness Ranking -0.026* Both Had Prior Knowledge -0.019 Both Did Not Have Prior Knowledge 0.018 Both Had High Encryption Jokes 0.026 Both Had Low Encryption Jokes -0.023 Transformed Age Difference 0.018 Practice The Same Religion -0.089* Have Different Nativeness Statuses -0.001 β Related -0.059 Live In The Same Vila -0.037 Mixed-effects restricted maximum-likelihood regression, * = p < 0.05
Flamson & Barrett 2013
Flamson & Barrett 2013
Conclusions Prior knowledge interacts with encryption level in evaluations of humor Replicates Flamson & Barrett 2008 Greater social closeness associated with similar evaluations of jokes Similarity in humor preferences reflects real-world social assortment Flamson & Barrett 2013
Social Assortment & Personal Evaluations Study Rated 10 targets on personal features: Funny: Tells good jokes Globally-relevant: Reputation (for telling good jokes), Respected, Lazy, Victim (of jokes) Dyadically-relevant: Friendly, Reliable, Frequent (spend time with) Predictions: 1) Higher funniness Greater social closeness 2) Effect of funniness on social closeness > Effect of globally-relevant traits 3) Higher funniness Higher dyadically-relevant traits Flamson & Barrett In Revision
Funniness Predicts Assortment Effect on Social Closeness Ranking Funniness Rating 0.230*** Same Location 0.997*** Mixed-effects restricted maximum-likelihood regression, *** = p < 0.001 Flamson & Barrett In Revision β Related 0.845*** Age Difference -0.029 Same Nativeness -0.025 Same Religion -0.018
Funniness Predicts Assortment Effect on Social Closeness Ranking Funniness Rating 0.230*** Same Location 0.997*** Mixed-effects restricted maximum-likelihood regression, *** = p < 0.001 Flamson & Barrett In Revision β Related 0.845*** Age Difference -0.029 Same Nativeness -0.025 Same Religion -0.018
Funniness Predicts Assortment More Than Globally-Relevant Traits Do Effect on Social Closeness Ranking β Funny 0.213*** Respected 0.122** Reputation 0.111* Lazy¹ 0.028 Victim¹ -0.027 Mixed-effects restricted maximum-likelihood regression, ¹ = Reverse-scored, * = p< 0.05, ** = p < 0.01, *** = p < 0.001 Flamson & Barrett In Revision
Funniness Predicts Dyadically-Relevant Traits Associations between funniness and various dyadic traits (β): Friendly: 0.221*** Reliable: 0.212*** Frequent: 0.245*** Mixed-effects restricted maximum-likelihood regression, *** = p < 0.001 Flamson & Barrett In Revision
Conclusions Closeness in social network is predicted by perceived funniness of target Perceived funniness is more predictive of social closeness than globally-relevant traits Perceived funniness is strongly predictive of other dyadically-relevant traits Interpersonal evaluations of humor ability reflect real-world social assortment Flamson & Barrett In Revision
Summary of Conclusions Covert signaling is an effective strategy for within-group assortment Humor is strongly linked with within-group variation and homophilic compatibility Similarity in humor preferences reflects realworld social assortment Interpersonal evaluations of humor ability reflect real-world social assortment Humor is an adaptation for honestly signaling compatibility in local cultural variation among potential long-term interaction partners
Additional Questions What are the linguistic features of encrypted humor?
Additional Questions What are the linguistic features of encrypted humor? What other features are assessed in assortment?
Additional Questions What are the linguistic features of encrypted humor? What other features are assessed in assortment? How does assortment change over time?
Additional Questions What are the linguistic features of encrypted humor? What other features are assessed in assortment? How does assortment change over time? What are the broader implications of the encryption model?
People reveal their character by nothing more clearly than what they find laughable. - Goethe, Elective Affinities Contact: tflamson@ucdavis.edu Thanks to the people of the Assentamento da Boa Ventura, Richard McElreath, Paul Smaldino, Clark Barrett, Dan Fessler, Rob Boyd, Martie Haselton, Greg Bryant, Allen Johnson, Emilia Yamamoto, Wall Hattori, and Carlos Caroso Soares. This research was made possible by funding from the International Society for Human Ethology, the UCLA Department of Anthropology, the UCLA Center for Latin American Studies, and the National Science Foundation.