Attraction and Selection in Cultural Evolution – ASCE

The experimental system that we need to develop should have the properties that make experiments in biological evolution successful: (i) the possibility of obtaining a large number of generations (ii) the possibility of repeating the same experiment several times (iii) the ability of the system to evolve in many different directions and (iv) the ability to quantify changes across generations and measure the space in which they operate.
We develop a new protocol that fits these criteria by taking advantage of the unique opportunity offered by a recently developed research platform that comprises fully automatized operant conditioning stations. This world unique structure offers the ideal environment to study social learning and cultural evolution as we have now demonstrated. In our first study, we implemented a computer-based experiment for monkeys to recreate the kind of cultural transmission of knowledge and skills that happen in humans. In a first phase, 15 baboons (Papio papio) learned to play a computer-based visual memory task. The baboons were shown a 4 by 4 grid with 4 red squares and 12 white ones. After a very short time (400 milliseconds) the red squares became white and the baboons’ task was to touch the squares that had just turned white on the touchscreen. Once the baboons had learned this task, the second stage of the experiment took place. We created a baboon visual transmission chain: the pattern of squares one animal touched became the patterns the next animal in the chain were given to memorize and reproduce. Then a third animal would learn from the patterns of the second, and so on. In this way, the baboons were essentially playing a visual version of the childhood game of “telephone”.

Using this new method allowed us to create a transmission chain with a huge number of possible outcome (1820 possible patterns to be exact, compared to only a few in previous studies) and made it possible to redo the experiment several times (we did six repetitions of the same experiment). We showed that when the patterns are transmitted between baboons they change in such a way that the baboons become better at the copying them. The initially random patterns generated by the computer gradually transformed into mathematically rare configurations known as tetrominos (the shapes used in the computer game Tetris). At the end of the transmission chain, baboons were more accurate in the memory game when the pattern was a tetromino compared to when it was not. In the same way that humans can perfect technologies, baboons were able to make small gradual improvements to the patterns provided by others, and these small changes accumulated to allow them to perform better at the task.
Following this initial study, we were able to introduce and control the action of selection during the transmission of the patterns between baboons. Selection acts to reduce the diversity of the patterns by eliminating certain grids according to a specified criterion. Using modelling and experiments, we showed that cultural selection has predictable consequences on cultural variation; it pulls the distribution in one direction. However, the consequences for cultural evolution are very different from those of biological evolution. In biological evolution, the variation is mostly random whereas it is directed in cultural evolution. This difference has an enormous impact on the relationship between the two processes. In biological evolution the gradual appearance of complexity is driven by selection whereas in cultural evolution this cannot be the case because variation is constrained by directed changes.

Work in progress.

Scientific articles :
- Claidière, N., Smith, K., Kirby, S., & Fagot, J. (2014). Cultural evolution of systematically structured behaviour in a non-human primate. Proceedings of the Royal Society B: Biological Sciences, 281(1797). doi: 10.1098/rspb.2014.1541
- Miton, H., Claidière, N., & Mercier, H. (2015). Universal cognitive mechanisms explain the cultural success of bloodletting. Evolution and Human Behavior, 36(4), 303-312. doi: 10.1016/j.evolhumbehav.2015.01.003
- Fagot, J., Marzouki, Y., Huguet, P., Gullstrand, J., & Claidiere, N. (2015). Assessment of Social Cognition in Non-human Primates Using a Network of Computerized Automated Learning Device (ALDM) Test Systems. JoVE(99), e52798. doi: doi:10.3791/52798
- Claidière, N., Scott-Phillips, T. C., & Sperber, D. (2014). How Darwinian is cultural evolution? Philosophical Transactions of the Royal Society B: Biological Sciences, 369(1642). doi: 10.1098/rstb.2013.0368
Our work in the media:
- Video summary of our first experiment (in French):
videotheque.cnrs.fr/index.php
- CNRS press release (in French):
www2.cnrs.fr/presse/communique/3797.htm

One of the major scientific challenges in modern evolutionary biology concerns the existence of non-genetic heritability and its importance in adaptive explanations. In particular, questions regarding the relationship between biological evolution and cultural evolution have become fundamental in biology as well as in social sciences. The purpose of this project of fundamental research is to shed light on these questions by establishing a new methodology for the study of cultural evolution.
By culture, we mean the set of objects, ideas and behaviours that are transmitted from individual to individual by non-genetic means. In recent years, important theoretical advances have clarified the relationship between biological and cultural evolution and raised the question of the balance between selection and attraction (i.e. directed transformations occurring during transmission). Ideally, transmission chains experiments, in which the results of one individual’s action are passed on to another individual, should easily provide an accurate estimation of cultural selection and attraction. However, transmission chains conducted so far have remained severely limited in this respect and there is to our knowledge no convincing experimental paradigm that can quantify some basic properties of cultural evolution such as the degree of fidelity of transmission, the power of cultural selection or the strength of attraction.
The experimental system that we need to develop should overcome these limitations and have the properties that make experiments in biological evolution successful. To succeed, we propose to take advantage of the unique opportunity offered by a recently developed research platform of the Laboratoire de Psychologie Cognitive (UMR 7290), that comprises fully automatized operant conditioning stations. This world unique structure offers the ideal environment to study social learning and cultural evolution and in order to study the feasibility of our proposal, we carried out a first pilot experiment with baboons that addressed the main technical challenges that we aim to overcome (large number of trials, repeatability of the experiment and the use of a non-verbal task). This pilot study, which was designed to demonstrate the potential of the approach taken, has been extremely successful.
Accordingly, we will estimate the balance between attraction and selection by applying our new methodology to 4 challenging questions in the field. Firstly, we will extend our pilot work by developing, carrying out and analysing transmission chains with non-human primates. Secondly, we will study the impact of social influence on social learning and cultural evolution, which, surprisingly, has never been studied in detail. Thirdly, building on previous studies showing that precursors of fundamental aspects of language can be found in non-human primates, we will address the question of the evolution of compositionality in non-human primates. Finally, we will also aim to demonstrate the strength of our approach to study cognitive capacities that are specific to humans by studying the cultural evolution of arguments.
The new experimental system that we have described opens the door to many questions that were, until now, impossible to address experimentally and we therefore expect to produce important methodological, analytical and theoretical advances in the field. The study of cultural evolution and social learning remains under-represented in France compared to other European countries. We aim to bring researchers from various institutions and disciplines together to create a core network of collaborations that will strengthen the domain, promote its development and facilitate the future emergence of research teams dedicated to this topic in France. This will open the way to a broad range of research opportunities that will facilitate the development of long term projects (5-10 years) and the prospect of employment for the coordinator.