Computational behavior theory and cultural evolution

New paper: The logic of fashion cycles

Plos ONE has accepted our paper “The logic of fashion cycles,” where Alberto Acerbi, Magnus Enquist and myself present a new theoretical model to understand fashion cycles (see my previous post on dog breeds). You can download a preprint, and here is the abstract:

Many cultural traits exhibit volatile dynamics, commonly dubbed fashions or fads. Here we show that realistic fashion-like dynamics emerge spontaneously if individuals can copy others’ preferences for cultural traits as well as traits themselves. We demonstrate this dynamics in simple mathematical models of the diffusion, and subsequent abandonment, of a single cultural trait which individuals may or may not prefer. We then simulate the coevolution between many cultural traits and the associated preferences, reproducing power-law frequency distributions of cultural traits (most traits are adopted by few individuals for a short time, and very few by many for a long time), as well as correlations between the rate of increase and the rate of decrease of traits (traits that increase rapidly in popularity are also abandoned quickly and vice-versa). We also establish that alternative theories, that fashions result from individuals signaling their social status, or from individuals randomly copying each other, do not satisfactorily reproduce these empirical observations.

Triple Coffee Gelato

The most common problem with coffee gelato is its consistency: adding coffee means taking out either milk or cream, which often results in a icy rather than creamy texture (even at professional Italian gelaterie). My solution is to use coffee concentrate. I use a Moka pot, the Italian standard way of making coffee in the house, and I concentrate the coffee by re-cycling the water through the machine for multiple brewing cycles, each time replacing the used coffee grounds with fresh coffee. For this recipe I use what is marketed in Italy as a “3 serving” machine, yelding about 1/2 cup of coffee. I re-cycle 3 times, hence the name of this recipe. The base gelato recipe is from here.

  • 1 cup whole milk
  • 1/4 cup sucrose and 1/4 cup glucose (why glucose?)
  • a pinch of salt
  • 1 cup heavy cream
  • 7 large egg yolks
  • 1/4 teaspoon vanilla extract
  • 1/2 cup coffee concentrate (see below)

Heat up the milk, sugar and vanilla extract in a saucepan until the sugar dissolves. In a bowl, stir the egg yolks until homogeneous, then pour the warm milk over them stirring constantly. Scrape everything back into the saucepan and cook (without boiling) for a few minutes, stirring continuously with a spatula until the mixture coats the spatula. Let cool (or not), add the heavy cream and the coffee concentrate, and pour in the gelato maker.

Variations: I often complement this gelato with toffeed nuts. Melt 1/2 stick of butter with 1/4 cup sugar and let go on high heat until the mixture starts to brown. Browning happens fast, so be on your guard. To verify that the toffee is ready, drop a drop in cold water cup. It’s ready when it solidifies into a hard lump. Then you pour the toffee on parchment paper, add the nuts (I use pecans) and let cool in the freezer to make the whole thing brittle. Mix the nuts in when the gelato is fresh out of the machine. After discovering this variation, I am not making plain coffee gelato again.

Fashions in dog breeds

I have recently attended a one-day course on data visualization with Edward Tufte and I have tried to put his advice on virtual paper in this supergraphic on the popularity of dog breeds, using AKC data (courtesy of Hal Herzog). The graph shows the popularity of 100 breeds over time (most popular breeds first), indicating the maximum in popularity and other peaks (if any). I have produced this graph as an inspiration for my ongoing work on cultural dynamics (some features are idiosyncratic to the data analyses I am making). Here are a few things I see in the graph:

  • Many breeds have had a clear peak of popularity, after which their diffusion declined to low values. This applies especially, but not only, to breeds used purely as pets – such as the all-time favorite, the poodle.
  • The faster a breed rises in popularity, the faster it goes back to its pre-spike level (this is not only a visual impression, it can be put on strong statistical grounds). A similar phenomenon has been observed for first names.

What else can you see? And how to explain it? In am working, with Alberto Acerbi and Magnus Enquist on an explanation of fashion cycles based on the cultural dynamics of preferences, as foreshadowed in our previous work on how social learning influences openness to new information. The paper is now under review at Plos ONE.

Lemon gelato (or other citrus, or passionfruit…)

This lemon gelato, somewhat inspired by this one, is not a lemon sorbet, because it has milk and cream, and is not a custard gelato, because it does not have eggs. It ends up tasting like frozen lemon cheese-cake. You are welcome to crumble some graham crackers in it!

  • 1 1/2 cup milk
  • 1 cup heavy cream
  • zest of 1 lemon
  • 1/2 cup lemon juice (about 3-4 lemons)
  • 1/4 cup sucrose + 1/4 cup glucose
  • 2 pinches of salt (see the lemon sorbet for why more salt is better!)
Warm up the milk and sugar until the latter dissolves. Whisk in slowly the lemon juice, then add the zest. Pour in the ice-cream maker, add the heavy cream, and go!
UPDATE: You can substitue another citrus for lemon. You can also make passionfruit gelato by omitting the lemon juice, adding 2 cups pureed passionfruit and 2 tbsp sugar.

Ginger gelato


  • 1 cup milk
  • 1/4 cup glucose + 1/4 cup sucrose (why glucose?)
  • 2-3 pinches of salt
  • 1 largish knob of ginger
  • 8 medium egg yolks
  • 3/4 cup heavy cream

This recipe builds on the vanilla gelato recipe. Peel the ginger and slice thinly. Boil it for several minutes together with the sucrose in as little water as possible (a small pot helps). Leave in the fridge overnight or for a few hours for best results. Strain the ginger syrup but don’t throw out the pieces of ginger – it’s candied ginger and you may want to add it back to the ice cream later!

Warm up the ginger syrup with the milk and salt in a pot. Mix the egg yolks in a bowl and pour the warm milk over them while whisking. Pour back into the pot and warm on low heat mixing constantly with a spatula until – the books say – the mixture start to coat the spatula (you quickly learn to recognize this). Remove from heat, let cool for how long you have patience, add the heavy cream and the candied ginger pieces, mix, and put in the ice cream maker.

NOTE: You may be tempted to cook the ginger and sugar directly in the milk, or worse ad grated ginger to an ongoing custard. These procedures may curdle the milk or custard, though, and even if they don’t you won’t get the candied ginger, which adds considerably to the final result!

Video: Pete Richerson, How Possibly Language Evolved

Part of the Cultural Evolution Seminar Series at Brooklyn College

(Download video, 620 MB)

Abstract: Human language has no close parallels in other systems of animal communication. Yet it is an important part of the cultural adaptation that serves to make humans an exceedingly successful species. Evolutionary scholars have have converged on the idea that the cultural and innate aspects of language were tightly linked in a process of gene-culture coevolution. They differ widely about the details of the process, particularly over the division of labor between genes and culture in the coevolutionary process. Why is language restricted to humans given that communication seems to be so useful? A plausible answer is that language is part of human cooperation. Why did the coevolutionary process come to rest leaving impressive cultural diversity in human languages? A plausible answer is that language diversity functions to limit communication between people who cannot freely trust one another or where even truthful communications from others would result in maladaptive behavior on the part of listeners.

Pete Richerson‘s primary research for many years has focused on the phenomena of cultural transmission of information and the evolutionary phenomena that derive from cultural transmission. He has been especially interested in the trade-offs involved in using other people as a source of information. Even if we imagine that the cultural system of humans has been adaptively optimized by natural selection, maladaptive cultural variants can still evolve. People cannot take advantage of normally adaptive rules like “imitate the successful” without incurring the risk that the appearance of success is a sham behind which lurks a culturally transmissible pathology. Prof. Richerson has investigated such processes with mathematical models, laboratory experiments, and most recently field investigations.

Video: Alex Mesoudi, The Experimental Study of Human Cultural Evolution

Part of the Cultural Evolution Seminar Series at Brooklyn College

(Download video, 580 MB)

Abstract: A growing body of theory has begun to analyse human culture – the body of beliefs, skills, knowledge, customs, attitudes and norms that is transmitted from individual to individual via social learning – as a Darwinian evolutionary process. Just as the biological evolution of species can be characterised as a Darwinian process of variation, selection and inheritance, so too culture exhibits these basic Darwinian properties. I will present the results of a series of experiments that have simulated cultural evolution in the lab using methods from social psychology. One set of studies using the “transmission chain method” have identified a bias in cultural evolution for information concerning social interactions over non-social interactions, as predicted by the “social brain” theory of human intelligence. Another set of studies have simulated the cultural evolution of prehistoric arrowhead designs, testing hypotheses that different patterns of arrowhead variation are caused by different ways in which arrowhead designs were transmitted between prehistoric hunter-gatherers.

Alex Mesoudi is Reader of Psychology at Queen Mary College, University of London. He is a psychologist who studies cultural evolution as a Darwinian evolutionary process, similar in key respects to the evolution of biological species. Culture can consequently be studied using similar methods, concepts and tools that biologists use to study biological (gene-based) evolution. My own research uses a combination of laboratory experiments and theoretical models to simulate the processes of cultural evolution, with the aim of identifying the small-scale, individual-level cultural  processes that generate the patterns and trends that are observed in human culture.

Video: Alberto Bisin, The economics of cultural transmission and socialization

Part of the Cultural Evolution Seminar Series at Brooklyn College

(Download video, 520 MB)

Abstract: Parents often spend considerable energy in teaching children their own beliefs, a behavior that has both economical and cultural underpinnings. I will present an economic analysis that predicts that minority groups spend more effort in teaching their children, and I will apply it to empirical data on marriage patterns. The analysis predicts that individuals from smaller groups most often choose to marry within their own group, and spend more resources in transmitting their beliefs to offspring.

Alberto Bisin is Professor in the Department of Economics at NYU. He is an economist interested in the interaction between cultural and socio-economic forces, expanding economic theory that has traditionally explained human behavior as determined by cost-benefit analyses independent of one’s cultural background.

New evidence for early learning of sexual preferences

We have recently posted on our sex research blog the results of our latest survey on sexual preferences. We surveyed sexual attraction to such traits as tattoos, eyeglasses, eye color, skin tone, stature, heir lenght, hair color, body type, long nails, painted nails, eyeglasses, tattoos, body hair, facial hair, armpit hair, and smoking. We found evidence that heterosexual individuals often prefer traits they have seen during their childhood in the parent of the other sex. That is, heterosexual males appear to prefer traits found in their mother, while heterosexual females trait found in their father.

A summary of this study and of previous ones on early learning of sexual prefernces is provided in Hanna Aronsson’s Ph.D. thesis introduction.

Video: Stefano Adamo, The Social Diffusion of Specialist Knowledge

Part of the Cultural Evolution Seminar Series at Brooklyn College

(Download video, 660 MB)

Abstract: I argue that the social diffusion of specialist knowledge is contingent upon a combination of environmental and cognitive factors that make such ideas significant to the lay person and motivates their social transmission and retention. The same combination of factors, however, also engenders an incomplete comprehension of the ideas being spread. I propose a qualitative method to understand what makes specialist knowledge relevant and anticipate how lay peoplemay retain and spread such knowledge.

Stefano Adamo is Reader in Italian History and Culture at the University of Banja Luka, Bosnia Herzegovina, and fellowat the International Center of Economic Research, Turin, Italy. His research interests include the history of ideas and the cognitive theory of culture,especially the history of economic concepts (money, market, etc.) and their social diffusion.