Denmark

**Author(s):** Bluet, A; Osiurak, F; Claidière, N; Reynaud, E. &lt;br&gt;&lt;br&gt;INTRODUCTION: Our technologies have never ceased to evolve, allowing our lineage to expand its habitat all over the Earth, and even to explore space. This phenomenon, called cumulative technological culture (CTC), has been studied extensively, notably using mathematical and computational models. One main factor that has received extensive attention is demography, and more precisely population size. However, the cognitive capacities needed for the emergence and maintenance of CTC remain largely unknown. In the literature, the focus is put on the distinctive ability of humans to imitate, with an emphasis on our unique social skills underlying it, namely theory of mind (ToM). A recent alternative view, called the technical-reasoning hypothesis, proposes that our unique ability to understand the physical world (i.e., technical reasoning; TR) might also play a critical role in CTC. Here, we propose a simple model that focuses on micro-societies experiments. Classically, we observe three forms of social learning in this literature: reverse-engineering (RE) where only the technology is present; observation (OBS) where a receiver witnesses a demonstration of the building of the technology by a sender; and communication (COM) with verbal interaction between a receiver and a sender. Thus, our model will implement these three social learning. Our model is composed of a simple environment with only one technology transmitted between generations of individuals. These individuals have two cognitive skills: ToM and TR, and can learn using the aforementioned different social-learning conditions. Their goal is to improve as much as possible the technology. Afterwards, we extend this model to include multiple chains of transmission in order to assess the effect of population size on CTC. &lt;br&gt;OBJECTIVES: To model the transmission of technical information between two individuals in a micro-society. Mainly, we intend to investigate the role of ToM and TR in the social-learning conditions found in the micro-society literature. Another goal of this model is to replicate and solidify the results of said literature. Finally, we intend to test our model on a population scale. This allows us to test the different hypotheses of the impact of population size on CTC. APPROACH: In our model, we assume an environment composed of a single technology. We further assume that this technology can be decomposed into components, called traits. For example, if the technology represents a vehicle, traits represent the wheels, the seating, the steering wheel, etc. In each generation, an individual appears in the environment. This individual has a set of knowledge (named cog) for each technology’s trait, and two cognitive skills, namely TR and ToM. First, the individual learns from previous generations using one of three social-learning forms (RE, OBS, COM) to improve its cogs. This learning phase depends on its cognitive skills (TR and ToM; depending on the social-learning form). Afterwards, the individual tries to improve the technology by refining its traits. Refining the technology’s trait equals increasing its quality. Subsequently, the individual might have an opportunity to innovate, which is equivalent to adding a new trait to the technology in our model. This opportunity comes with a fixed probability, and the individual needs to cease it either by having a good enough technology or by being himself a quite advanced individual compared to its environment. Finally, the individual disappears from the environment and is replaced by a new one. This process is repeated for N generations (N = 1000 if not stated otherwise), thus creating transmission chains. In the extension of our model, we run multiple transmission chains at the same time. Each chain has its technology and individual, the latter of which can learn from conspecifics in other transmission chains. This emulates a fully connected population of individuals. It is important to note that we assume that innovations are transmission through social learning, meaning that an individual will bring back through learning the innovation from another transmission chain.&lt;br&gt; RESULTS: In our model, all three social learning forms yield CTC in the form of exponential growth of technology’s quality. However, both RE and OBS produce lower technology quality than COM, the former both having very similar behaviours and results, which agrees with classical results of the literature on transmission chains, where RE has the lowest technology quality, closely followed by OBS and then well ahead COM. When comparing model of COM with and without ToM, we see little difference between the two, demonstrating a limited impact of ToM. Further analyses indicate that ToM has an impact only during the first generations, indicating that its effect may be to accelerate the acquisition of an optimal technology. When we add multiple transmission chains, we observe an effect of population size on CTC. This effect is weaker the greater the population size, hinting at a higher impact of adding individuals for a smaller population size. Further analysis showed that our model is driven by the frequency of innovation, thus explaining this result. Indeed, adding individuals is equal to adding potential innovators. However, we observe that once population size reaches a certain threshold, adding more individuals does not bring more innovation because there are already enough innovators. &lt;br&gt;DISCUSSION: Our model results are in accordance with the transmission-chain literature, where communication conditions often outperform non-communication ones. This both strengthen the result of said literature and validate our model. Furthermore, we observe exponential growth of technology’s quality, something that is also observed in our cultural evolutionary history (Lehman, 1947). We posit that this exponential growth is due to a snowball effect: an increase in technology’s quality is accompanied by an increase in the understanding of this technology, thus increasing in return the capacity of the individual to increase the technology’s quality and so on. This snowball effect occurs because TR supports both the transmission of information and the modification of the technology. This effect of both the understanding and the technology’s quality improving together is something that we have already shown in a recent study using transmission chains, where an increase in the speed at which a wheel rolls down an axis is correlated to the increase in the understanding of how mass repartition affects this movement. Our results show that ToM has little to no impact on CTC. This seems surprising because ToM can only impact learning positively in our model. However, further analysis showed that ToM has an impact early on in our model, acting as a catalyst for the first burst of increases in technology’s quality. In this regard, we consider the impact of ToM as a launching pad for the snowball effect described earlier, being neither necessary nor sufficient for the emergence of a CTC. When applied to the scale of a population, our model shows interesting and well-known results. Indeed, a larger population size has been extensively linked to an increase in CTC and this effect has also been linked to the frequency of innovation. These are encouraging results, and in the future, we intend to further extend our model to implement other aspects of demography that have been shown to have an impact on CTC, such as population structure and dissemination of knowledge.

Cultural Evolution Society 2022

Impact of technical reasoning, theory of mind and population size on cumulative technological culture: insights from a model of micro-societies

cognitive sciences / technological culture / technical reasoning / theory of mind / individual-based modelling / population size

AIAI 2020 (16th International Conference on Artificial Intelligence Applications and Innovations) should be held in Halkidiki, Greece. Due to current Covid-19 situation the conference will be held virtually. Underline will host a repository of video talks and preserve them for the future. Videos will be released on Underline one day after the conference.

AIAI 2020

**Author(s):** Bluet, A; Osiurak, F; Claidière, N; Reynaud, E. INTRODUCTION: Our technologies have never ceased to evolve, allowing our lineage to expand its habitat all over the Earth, and even to explore space. This phenomenon, called cumulative technological culture (CTC), has been studied extensively, notably using mathematical and computational models. One main factor that has received extensive attention is demography, and more precisely population size. However, the cognitive capacities needed for the emergence and maintenance of CTC remain largely unknown. In the literature, the focus is put on the distinctive ability of humans to imitate, with an emphasis on our unique social skills underlying it, namely theory of mind (ToM). A recent alternative view, called the technical-reasoning hypothesis, proposes that our unique ability to understand the physical world (i.e., technical reasoning; TR) might also play a critical role in CTC. Here, we propose a simple model that focuses on micro-societies experiments. Classically, we observe three forms of social learning in this literature: reverse-engineering (RE) where only the technology is present; observation (OBS) where a receiver witnesses a demonstration of the building of the technology by a sender; and communication (COM) with verbal interaction between a receiver and a sender. Thus, our model will implement these three social learning. Our model is composed of a simple environment with only one technology transmitted between generations of individuals. These individuals have two cognitive skills: ToM and TR, and can learn using the aforementioned different social-learning conditions. Their goal is to improve as much as possible the technology. Afterwards, we extend this model to include multiple chains of transmission in order to assess the effect of population size on CTC. OBJECTIVES: To model the transmission of technical information between two individuals in a micro-society. Mainly, we intend to investigate the role of ToM and TR in the social-learning conditions found in the micro-society literature. Another goal of this model is to replicate and solidify the results of said literature. Finally, we intend to test our model on a population scale. This allows us to test the different hypotheses of the impact of population size on CTC. APPROACH: In our model, we assume an environment composed of a single technology. We further assume that this technology can be decomposed into components, called traits. For example, if the technology represents a vehicle, traits represent the wheels, the seating, the steering wheel, etc. In each generation, an individual appears in the environment. This individual has a set of knowledge (named cog) for each technology’s trait, and two cognitive skills, namely TR and ToM. First, the individual learns from previous generations using one of three social-learning forms (RE, OBS, COM) to improve its cogs. This learning phase depends on its cognitive skills (TR and ToM; depending on the social-learning form). Afterwards, the individual tries to improve the technology by refining its traits. Refining the technology’s trait equals increasing its quality. Subsequently, the individual might have an opportunity to innovate, which is equivalent to adding a new trait to the technology in our model. This opportunity comes with a fixed probability, and the individual needs to cease it either by having a good enough technology or by being himself a quite advanced individual compared to its environment. Finally, the individual disappears from the environment and is replaced by a new one. This process is repeated for N generations (N = 1000 if not stated otherwise), thus creating transmission chains. In the extension of our model, we run multiple transmission chains at the same time. Each chain has its technology and individual, the latter of which can learn from conspecifics in other transmission chains. This emulates a fully connected population of individuals. It is important to note that we assume that innovations are transmission through social learning, meaning that an individual will bring back through learning the innovation from another transmission chain. RESULTS: In our model, all three social learning forms yield CTC in the form of exponential growth of technology’s quality. However, both RE and OBS produce lower technology quality than COM, the former both having very similar behaviours and results, which agrees with classical results of the literature on transmission chains, where RE has the lowest technology quality, closely followed by OBS and then well ahead COM. When comparing model of COM with and without ToM, we see little difference between the two, demonstrating a limited impact of ToM. Further analyses indicate that ToM has an impact only during the first generations, indicating that its effect may be to accelerate the acquisition of an optimal technology. When we add multiple transmission chains, we observe an effect of population size on CTC. This effect is weaker the greater the population size, hinting at a higher impact of adding individuals for a smaller population size. Further analysis showed that our model is driven by the frequency of innovation, thus explaining this result. Indeed, adding individuals is equal to adding potential innovators. However, we observe that once population size reaches a certain threshold, adding more individuals does not bring more innovation because there are already enough innovators. DISCUSSION: Our model results are in accordance with the transmission-chain literature, where communication conditions often outperform non-communication ones. This both strengthen the result of said literature and validate our model. Furthermore, we observe exponential growth of technology’s quality, something that is also observed in our cultural evolutionary history (Lehman, 1947). We posit that this exponential growth is due to a snowball effect: an increase in technology’s quality is accompanied by an increase in the understanding of this technology, thus increasing in return the capacity of the individual to increase the technology’s quality and so on. This snowball effect occurs because TR supports both the transmission of information and the modification of the technology. This effect of both the understanding and the technology’s quality improving together is something that we have already shown in a recent study using transmission chains, where an increase in the speed at which a wheel rolls down an axis is correlated to the increase in the understanding of how mass repartition affects this movement. Our results show that ToM has little to no impact on CTC. This seems surprising because ToM can only impact learning positively in our model. However, further analysis showed that ToM has an impact early on in our model, acting as a catalyst for the first burst of increases in technology’s quality. In this regard, we consider the impact of ToM as a launching pad for the snowball effect described earlier, being neither necessary nor sufficient for the emergence of a CTC. When applied to the scale of a population, our model shows interesting and well-known results. Indeed, a larger population size has been extensively linked to an increase in CTC and this effect has also been linked to the frequency of innovation. These are encouraging results, and in the future, we intend to further extend our model to implement other aspects of demography that have been shown to have an impact on CTC, such as population structure and dissemination of knowledge.

technical paper

### Welcome to the fourth Cultural Evolution Society Conference, CES2022! 

As last year, Underline will support the online component of the conference. The ongoing COVID-19 pandemic is causing a great deal of uncertainty for all of us. Therefore, we seek to offer a hybrid format that will hopefully allow some folks to travel to Aarhus, others to meet up locally or regionally in CES hub-lets, and other yet to participate online. The field of cultural evolution is growing in size and scope; its theories, methods and applications - both academic and applied - are getting ever more robust. Do join us in what promises to be an exciting exchange of ideas and a celebration of our field.

##### ***For SoMe interaction and live tweeting please tag @CulturalEvolSoc and use #CESconf2022***


#### The CES 2022 local organising committee
* Luseadra Joy McKerracher
* Jesper Sørensen 
* Niels Nørkjær Johannsen 
* Benjamin Grant Purzycki 
* Mette Løvschal 
* Felix Riede
* Iza Romanowska 
* Karsten Olsen

Please click on the button bellow if you want to watch the videos

The field of cultural evolution is growing in size and scope; its theories, methods and applications - both academic and applied - are getting ever more robust. Do join us in what promises to be an exciting exchange of ideas and a celebration of our field.

**Author(s):** Morin, O., Mueller, T.F., Morisseau, T., Winters, J. **Author(s):** No **Author(s):** No The interpretation of referential messages is biased by communicative salience: some referents are easier to communicate because they come more readily to mind. We explored two aspects of communicative salience, in a referential communication task involving colors. We predicted that lexically focal referents would be communicatively salient, and that the codes used to indicate them would be more stable. We tested these hypotheses with an online gaming app, the Color Game, in which players had to use ambiguous black and white symbols to communicate specific color hues to other players (> 200,000 trials). Focal colors, centrally associated with basic color terms, are easier to communicate than less focal ones. Code stability is contingent on the type of colors conveyed, with participants being more likely to keep using the same symbols when referring to focal colors. Communicative salience impacts the evolution of semantic conventions. These findings bridge the pragmatics of referential communication and the evolution of lexical semantics.

Communicative salience shapes the emergence and stability of semantic conventions

**Author(s):** Krajinovic, A; Rodrigues, X **Author(s):** No **Author(s):** No Linguistics has shown that the evolution of grammar follows a pattern known as grammaticalization, where structures evolve from more lexical and iconic meanings to more grammatical and abstract meanings. Lack of precise data on language transmission patterns presents an obstacle in understanding the factors that drive this grammaticalization process. In this talk, we argue that multimodal constructions such as internet memes follow a similar evolution pattern, and can thus provide a wealth of data on this phenomenon. We analyze twenty different memes, each consisting of an image macro with or without text. We use the number of browser searches, provided by Google Analytics, as a proxy for the time-dependent transmission rate of each meme. We then collect information on the date, format and semantic change over time of a sample of meme instances shared on Twitter and Reddit. We show that statistically, wider transmission leads to a higher rate of semantic change in the direction of higher abstraction. We also show that the memes go through stages parallel to grammaticalization, such as use in new contexts, semantic bleaching, erosion, and decategorialization, and that these stages follow the same temporal ordering across memes (see our example at https://bit.ly/3a4JFbk). Social networks can therefore provide a valuable lab for the study of communication patterns and language evolution.

From iconic pictures to “grammatical” templates: Modeling the evolution of internet memes

**Author(s):** Sebire, I; Camargo, C; Introduction: Popular media plays a significant role in shaping and reflecting culture and society. In this paper, we investigate patterns in the use of recurring storytelling devices, i.e. common tropes, in popular media. Examples of tropes include the "mad scientist", "villain wearing a dark palette", or even "melancholic song playing during sad moment". Using a database of media tropes, we compare clusters of tropes with actual genre classifications, and study the differences in media through geography and time. Methods: This study uses a dataset collected from TVTropes, a wiki-format crowdsourced website describing tropes in popular media. We apply approaches from cultural evolution along with data science methods such as topic modelling, sentence embedding, and stochastic block models, to simultaneously cluster tropes and the works they feature in. # Results & Discussion: We show the applicability of the nonparametric stochastic block model-based topic model topSBM to identify a hierarchy of genres and subgenres, as well a hierarchy of storytelling tropes present in the data over time. We end by discussing the strengths and limitations of machine learning in cultural evolution.

Genres, subgenres, and storytelling tropes: a data science approach

**Author(s):** Fedorova, N.; McElreath, R. Introduction: Shelter is a key human technology, and varies extensively throughout the ethnographic record. Like other technologies, human investment in shelter is strategic and has consequences for the evolution of settlement landscapes. We present findings from a household study of the drivers of building behaviour in the ger districts of Ulaanbaatar, Mongolia. Methods: We formulate a Stochastic Dynamic Programming (SDP) optimality model to generate non-null predictions about building behaviour. With the use of Approximate Bayesian Computation we compare simulated model outputs with empirical data and select parameter combinations that can generate the observed data. Results: The SDP model allows us to explicitly interrogate trade-offs between saving and building through time. Preliminary results suggest that the dynamics of saving influence the time horizon and extent of building, through changes in family composition. Conclusion: Human decisions are always balancing trade-offs, which are difficult to access with traditional analysis methods. We hope that through the robust comparison of empirical work to simulated non-null predictions we take productive steps to understanding human landscape investments.

Strategic investments in housing in Ulaanbaatar, Mongolia

**Author(s):** Miton, H; Morin, O. Cattle brands (ownership marks left on animals) may be subject to forces influencing other graphic codes: the copying of constituent parts, pressure for distinctiveness (e.g. to distinguish a mark from its nearest neighbors), and pressure for complexity (e.g. to be harder to fake). The historical record of cattle brands in some US states is complete due to legal registration, providing a unique opportunity to assess how sampling processes leading to time- and space-averaging influence our ability to make inferences from limited datasets in fields like archaeology. In this preregistered study, we used a dataset of ~81,000 Kansas cattle brands (1990-2016) to explore two questions: (1) the relative influence of copying, distinctiveness, and complexity, and (2) the effects of time- and space- averaging on statistical signals. By conducting generative inference with an agent-based model, we found that the patterns in our data are consistent with random copying and pressure for intermediate complexity. In addition, by comparing mixed and structured datasets, we found that these statistical signals of random copying are robust to, and possibly boosted by, time- and space-averaging.

Distinctiveness, copying, and complexity in the cultural evolution of cattle brands

**Author(s):** Deffner, D; Mezey, D; Wu, C; Schakowski, A; Kahl, B; Romanczuk, P; Kurvers, R Introduction: A fundamental challenge in collective problem solving is the dynamic integration of individual and social information. Previous theory has shown that social learning needs to be selective in order to be adaptive, yet it is still largely unknown how agents collectively negotiate this balance. Methods: We investigate these trade-offs in the context of collective foraging (collecting spatially distributed resources) and present novel agent-based simulations and human foraging experiments in a virtual world. This immersive environment creates a natural trade-off between individual search and social learning and provides spatial trajectories as well as visual field data that we analyze through Bayesian computational modeling. Results: Simulation results show how different amounts of social learning maximize collective returns depending on the structure of the environment and visual constraints. Experiments revealed that people copy selectively and can adapt social learning to the demands of the physical and social environment. Discussion: Our results shed new light on how groups collectively adapt to the demands of the environment through exploration and selective imitation.

Selective social learning in collective foraging

 Transient diversity in a cooperative group involves the maintenance of diverse solutions while the group attempts to reach a consensus. I argue that any mechanism that slows diffusion and delays consensus will increase the quality of solution at the cost of increased time to reach it. I further argue that this is a general principle analogous to the role of positive assortment in the evolution of cooperation. Several processes facilitate the evolution of cooperation—kin selection, direct reciprocity, indirect reciprocity, group structure, partner choice—but all these do so by directing the benefits of cooperation toward cooperative individuals. I review specific mechanisms that facilitate transient diversity, synthesizing evidence from both empirical studies and diverse formal models—multi-armed bandits, NK landscapes, cultural attractors, cumulative innovation models, and evolutionary transmission models—to support my argument. Apparent exceptions this principle occur only when problems are sufficiently simple that they can be solved by mere trial and error. This work has implications for the cultural evolution of teams, population structures, and the accumulation of innovations.

Maintaining transient diversity is a general principle for improving collective problem solving

Introduction: Santa Claus is a morally concerned, omniscient, interventionist supernatural agent. Belief in a ‘real’ Santa emerges in early childhood, is highly contingent upon adult endorsement, and typically ceases in the 7th year. Studying children may yield insights into the development and evolution of religious cognition. We expect that belief and exposure to cultural rituals will predict children's moral behavior during the Xmas period. Methods: We collected longitudinal data from parents (on Prolific) regarding their child’s belief in Santa, cultural practices and symbols, Santa’s moral concerns, and child prosocial behavior in the preceding week both before the Xmas period (T1: 25 November), and for the apex of the Xmas festival (T2: 26 December; N = 97). Results: Preliminary results suggest that disobedience/aggression did not change over the Xmas period, but that prompted prosociality increased, and unprompted prosociality decreased. The change was not predicted by belief in Santa, ritual salience, Xmas-relevant CREDs, or age. Discussion: Our real-world data among children have implications for the generalizability, learnability, and evolvability of the ‘Big Gods’ hypothesis.

We’ve made a list and we’re checking it twice: A two-wave longitudinal study of children’s moral behavior during Christmas

Introduction: Cultural evolution of norm psychology holds that groups first evolved norms regulating interpersonal behaviors. Second, individuals in groups with stable prosocial norms evolved norm psychology helping with internalization of such norms. I argue that perception of moral norms as absolute and universal is internal part of norm psychology and test three hypotheses derived from such reasoning. Methods: I utilized data from Pew, FBI crime records, WVS, and NSYR (National Study of Youth and Religion) and hierarchical regression framework. The Pew and FBI data served to test the association between % of moral absolutists in the US state and crime rates per 100K (H1 group benefits). WVS data were used to test whether moral absolutists justify norm transgressions less than relativists (H2 norm compliance). NSYR data allowed me to test relationship between moral relativism and likelihood of being jailed, arrested, and found guilty (H3 individual benefits). Results and conclusion: I found mixed support for the hypotheses which is in conflict with clear results from controlled experiments highlighting the importance of observational data in the study of norm psychology evolution.

Moral absolutism as a component of norm psychology: The relationship between moral perception and compliance with norms

Demonstrating how individual-level cultural transmission scales up to produce population-level social norms is one of cultural evolution’s evergreen challenges. Spoiler: This talk does not solve that challenge. Happily, even small steps towards understanding the interplay between micro and macro scales are useful. I focus my discussion on kinship terminologies, which are the patterns of words that languages have for denoting and referring to family members. Our systems of reckoning relatives vary cross-culturally, and the structural features of these norms persist and recur across the globe. In some languages, siblings are distinguished by age; in others by gender; by both aspects; or not at all. The women I call my ‘aunty’ in one society might be called ‘mother’ in another. I present collaborative work from a team project that explores terminologies as a product of cultural evolutionary processes shaped by cognitive and interactional constraints. Observations from Datooga-speaking children and their families demonstrates the contexts of individual transmission and learning of kinship terms, while phylogenetic analyses explore functional explanations for the persistence of patterned structural systems across languages. Reconciling these scales might be too grandiose an aim, but the effort helps us consider the relationship between language and culture, as well as the status of “shared ancestry” as an explanation.

Premium content

Impact of technical reasoning, theory of mind and population size on cumulative technological culture: insights from a model of micro-societies

Downloads

Next from Cultural Evolution Society 2022

Communicative salience shapes the emergence and stability of semantic conventions

Similar lecture

Real-time prediction of online shoppers’ purchasing intention using random forest

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES