Cultural Evolution in the Digital Age

As I mentioned in a previous post, I am beginning to collect some thoughts and materials for a project around the topic “Cultural Evolution in the Digital Age”. The goal of the exercise would be to investigate how new digital technologies change the process of cultural transmission and evolution, using methodological tools and ideas from cultural evolution theory (intended in a quite broad sense).

I am of course aware I am not exactly the only person interested in this topic, but I am slowly realising that the cultural evolution perspective could in fact provide something different from what is already around. A great share of the scientific work on these topics is carried on by computer scientists and physicists, and while generally excellent from the “technical” point of view (quantity and analysis of data), it often lacks, for obvious reasons, theoretical background in social and human sciences. The explanations presented are often ad-hoc models that reproduce  – usually very well! – the data of the specific case studied, but can difficulty be generalised, or used to support pre-existent theories.

Good examples exist, but not, to my knowledge, coming from a cultural evolution perspective. Cultural evolution, with its explicit quantitative character, and with a distinctive taste for wide generalisation (it is intended as a plus here) seems especially well suited to be used as a theoretical background.

The goal of the project would be to have, in a short-term prospect, enough material to plan some experiments/models/data analysis specifically dedicated to the topic, and, in medium-term, possibly write a grant and/or a book on the subject.

The particular sub-topics would be selected as they fit with researches in cultural evolution. For example:

  • Cultural transmission biases in the digital age. Cultural transmission biases (e.g. “copy prestigious individuals”) have great importance in cultural evolution theory. Their existence is justified with their adaptivity in small-scale societies, where, for example, “prestigious” individuals tended, on average, to be also skilled, and, moreover, their skills were – again on average – relevant for the individuals who copied them. What when we have access to a virtually infinite amount of models? What when the skills of prestigious individuals are not relevant? A discussion on this topic, from this perspective, can be found here, but no cultural evolutionary work, to my knowledge, addressed the question. The same logic can be applied to other biases. Conformist bias, for example, indicates a disproportionate tendency to copy the majority. In a small-scale society, the number of people performing a certain behaviour, or adopting a certain cultural variant, should be estimated, possibly in indirect ways. What when we are provided explicitly with precise signs of popularity (e.g. the number of likes in Facebook, or re-tweet in Twitter, etc.)?
  • Preservative versus transformative cultural transmission. Is better to think to cultural transmission as a mainly preservative process (where variants are copied with relatively high fidelity and errors are random) or as a mainly transformative one (where individuals actively recreate variants each time they are transmitted, and stability is due to the existence of stable “cultural attractors”)? This question is central in a current debate in cultural evolution theory (see here). Digitally mediated interactions can be considered as supporting highly preservative transmission, in opposition, for example, to “traditional” oral transmission. How, for example, a story spreads in an oral context versus in a social media? Is there some content that is easier to transmit when it needs to be reconstructed each time, or vice versa?
  • Are we getting more homogeneous, or more different? Again, this is such a common question that I feel almost uncomfortable to put it here, but the point is: is the cultural evolution perspective telling us something new about it? Here I refer to a tradition of modelling that originated with Robert Axelrod’s well-known Dissemination of culture model, but it is partly forgotten in contemporary cultural evolution. A series of models have indeed explored the consequences of various methods of transmission (and, explicitly, media) on cultural homogenisation and differentiation. I am convinced this tradition could be revamped, especially adding features like transmission biases (or cultural attractors) to the basic models, and it could tell us something interesting about how digital media are changing our society.

These are just some examples, I believe other subjects can be treated explicitly from a cultural evolution perspective, such as the whole echo chambers/information bubble issue (I wrote some musings recently about it here) and others. I am very open to any suggestion, and, especially, if someone is aware of some cultural evolution work on related topics (or s/he is actually doing something) I’d be happy to discuss!

Update 28 September 2016:

I recently uploaded a preprint (see here) where I explore at some length the first two topics, and I discuss more generally how cultural evolution could contribute to the study of digital media.

Update 18 January 2017:

Now a paper published in Frontiers in Human Neuroscience.


[The picture, from the flickr account of Stockholm’s Tekniska museet, shows the Telephone Tower of Stockholm at the end of XIX Century. Every telephone required a direct physical line to a phone exchange centre – the tower – where the calls were manually connected by an operator]


15 thoughts on “Cultural Evolution in the Digital Age

  1. Dear Alberto,

    I see this as a very important new domain for integrative research. Perhaps the Cultural Evolution Society we are forming will be able to expand the dialogue and invite deep collaborations around topics like this one.

    As I read your blog post, I began thinking about ways to formulate hypotheses based on cultural transmission research in primate studies (human-chimp comparisons, for example), early childhood development (where social learning can be enhanced or altered by digital technologies); the diffusion of memes and structures of social discourse (which can be tracked with much greater fidelity using social media); and other key sub-fields of cultural evolutionary studies.

    I see this playing out in the digital humanities and through databases like Seshat and the newly launched D-SPACE data archives — how they can be augmented or complemented by the Twitter database or specific mobile apps designed to collect social data on human language and behaviors. This is a large domain of interest for campaigners, people in marketing, public health and prevention science practitioners, and others with an interest in dealing with social challenges.

    All of which is to say, let me know how I can help!



    1. Hi Joe,

      Thank you very much for your sympathetic words. In fact this looks like a “natural” domain for cultural evolution investigations, and it is surprising that not much is around. I look forward to discuss more, and I will try to post on the blog the next developments. Let’s see what we can do!

  2. I hope you will forgive a comment by an interested layman. I’m a computer scientist by trade with a great interest in cultural evolution.

    On the subject of Cultural Evolution in the digital age, have you looked at what is happening in the field of artificial neural networks, particularly with so-called “deep learning networks”? The reason I mention it is that they could serve as a model (even if just by analogy) of how ideas are learned. Let me explain using a recent, Internet famous case.

    An AI researcher trained a neural network to generate playing cards for a game called “Magic the Gathering” (see The process was quite simple: teach the network by exposing it to a large number of existing cards (which are complex) and then let the trained network generate new ones. The result was extraordinary, the network generated results that can only be called creative.

    The neural network learns only by copying, it has no hard-coded rules or algorithms. To me that is analogous to how humans learn. My evolutionary interpretation of this experiment is that the network is converting the phenotype (the card samples) into a kind of genotype (meme?) which is the structure of the neural network. From that genotype the network is able to create new phenotypes (cards). The ways that it makes mistakes are also very interesting, its creativity may very well be those mistakes.

    I’m sure I have gotten lots of this wrong but I think it is at least suggestive of reasonable particulate model of cultural evolution and perhaps worth exploring.


    1. Dear Aaron,

      On the contrary, thank you very much for your interesting comment! I have some basic knowledge of what is going on with “deep learning networks”. In fact few years ago I was mainly working with (very simple) neural networks and simulated robots to study social learning and cultural evolution (you can have a look at my publications page if interested).

      The results obtained with “deep learning networks” are impressive, I agree, but many researchers believe that these results are obtained at the expenses of a realistic simulation of how human learning actually works (see e.g. I admit this is quite off of my expertise, but I’d tend to believe that some, albeit simple, “hard-coded rules or algorithms” are an important part of -biological- learning.

      Said so, my interest in the “Cultural Evolution in the Digital Age” project is more on how human behaviour and learning interacts with digital technologies, so, for example: how humans would evaluate information provided by a “deep learning networks” versus information coming from a “traditional” (i.e. human) expert? This is something getting nowadays more and more common…

      Thank you again for your comment, and happy to exchange ideas!


      1. Hi Alberto,

        Thanks for the reply. I will check your publications section for your papers on this topic. I look forward to reading them.


  3. Hello Alberto,

    Another comment and some questions, this time more on topic to your post. I’m trying to understand the research questions you have in mind for “Cultural transmission biases in the digital age” but they aren’t entirely clear to me. When you say “cultural transmission” what exactly is being transmitted? What is the unit of heredity? “Culture” is an abstract term whose definition, I believe, is controversial even among anthropologists. It seems to me that the identification of this unit of heredity, whatever one may call it, is essential to this science. In fact, I would go further and say that the proper object of cultural evolution is this unit of heredity rather than human beings.

    My second question is, what is the standard model for the transmission of culture between people? What I mean is, what is the process by which this unit of heredity replicates from one person to another person? I don’t mean the pathways identified by Boyd and Richerson but the actual process of replication from brain to brain. How does it work?

    I ask these question because I think they are foundational questions that would inform the research questions you pose for the interesting topic of cultural evolution in the digital age. I am also interested in your views on them.

    My own, tentative, view is that Schema Theory from cognitive psychology offers a potential answer. Schemata can be viewed as the unit of heredity with a transmission model following Transactional-Sociopsycholinguistics. These theories seem to have good support among cognitive psychologist and they have the added benefit of being empirically verifiable. Nishida [1999] has a nice summary of this approach in “A Cognitive Approach To Intercultural Communication Based On Schema Theory”. What Nishida, and Schema Theory in general, do not provide is an evolutionary context. How, for example, do we assess the fitness of a given schema within an ecology of schemata residing in the human brain?

    One other thing I like about Schema Theory is that it fully takes into account the high degree of plasticity in schema (which are unlike physical genomes in this regard) and the high degree of lossy-ness (to use a computer science term) in transmission.


    1. Hi Aaron,

      Thank you again for your comment. Your questions are in fact quite central – and in the same time, problematic – in the study of cultural evolution. I admit I do not have a direct answer for them. Various scholars have different ideas about what can be exactly considered the unit of heredity in cultural transmission (other serious cultural evolutionists, e.g. Dan Sperber, even believe that there is not really inheritance in cultural evolution, but this do not prevent to consider culture as an evolutionary system in certain respects). The same can be said for how exactly the transmission is implemented.

      My opinion is that one can develop an evolutionary theory of culture without knowing some of the details, similarly as Darwin developed an evolutionary theory in biology without knowing – or even getting wrong…- some of the details. (Then of course Darwin was right, while cultural evolutionists like myself can be wrong…). I wrote something related to these topics in a relatively recent paper together with Alex Mesoudi, which is open access and can be found here:

      I do not know about schema theory – happy to have a look at the reference you mention!



      1. Hello Alberto,

        Thanks for your reply. I feel fortunate to have the opportunity to exchange ideas with an expert in the field.

        I read your paper and thought it was very good. I find your selection theory more convincing than Sperber’s “attractors”. If we consider the human mind as the environment in which cultural micro-evolution takes place then I think we can dispense with attractors altogether. I think it is safe to assume that within the mind there are schemas (cultural traits) and psychophysical modules. We could then say that the fitness of a given schema is a function of its compatibility with these pre-existing schemas and with the modules. Compatibility with a given module would grant a certain schema higher fitness. The fact that modules (as genetically determined entities) are common among humans would explain why certain schemas are more common than others. Hence we could explain the same phenomena Sperber identified without the need for attractors.

        For example, if we assume hyperactive agency detection is such a module then we should see schemas that are compatible with agency to have greater fitness (be more readily accepted by a person and be more widespread in a population) than those that don’t. And indeed this is exactly what we do see in human history with supernatural-based schemas. It has taken humanity a very, very long time for scientific schemas to out-compete this terrible module. And even now its force is strong.

        I noticed also that your paper does mention schema theory in passing, “In line with schema theories from cognitive psychology, it was found that low-level actions…” Later on you write, ‘A coarse-grained description of the cultural trait is “a story involving a young lady, first oppressed by her stepmother and stepsisters, and then succeeding in marrying a prince”.’ but don’t make the connection to schemas. What you describe as a cultural trait exactly matches the description of a schema. Schemas “represent knowledge about concepts: objects and the relationships they have with other objects, situations, events, sequences of events, actions, and sequences of actions.” Perhaps schemas could be a useful replacement for the term cultural traits?

        Kindest regards,

        1. Hey Aaron and Alberto,

          I’ve been following this thread as it unfolds and want to support what Aaron is saying. As you may know Alberto, I studied with George Lakoff earlier in my career and have a lot of experience working with human conceptualization and cognitive semantics. When I first started digging into the research literature on cultural evolution a few years ago, my first observation was that the field could quickly gain much greater coherence by (a) taking the philosophical stance of embodied cognition as foundational for cultural expression; and (b) merging with the immense literature on human conceptualization having to do with social scripts, frame semantics, conceptual metaphors, and the foundational role of image schemas for orchestrating human thought and behavior.

          It just so happens that this readily extends to non-humans as well because all minds are embodied!



          1. Hi Joe,

            Thank you for your reply. With respect to the literature on human conceptualization, are you aware of any papers (or perhaps have written one yourself) that attempt to describe the items you mention in an evolutionary context? The cognitive psychology literature is enormous but I haven’t been able to find much that attempts to describe the changes in human concepts over time as a Darwinian process.


            1. Hey Aaron,

              There are books like The Origin of Ideas by Mark Turner and Women, Fire, and Dangerous Things by George Lakoff that draw linkages between human conceptualization as it is today and its relationship with the evolution of language earlier in our history. But the emphasis in them is not to characterize their findings in Darwinian terms. So the bridges are implicated but not fully explored.

              This is an area where more work will be needed to make the case with academic rigor. Weaving across the cognitive psychology literature to the works of people like Harvey Whitehouse, Pascal Boyer, and Justin Barrett (in anthropology of religion) and with efforts like that of Dan McAdams and other narrative psychologists will help. But so far as I know there hasn’t been anyone whose primary scholarly goal was to explicate this relationship.

              My own work has been non-academic as I am working to apply interdisciplinary knowledge to pressing social problems humanity is confronting today. So I have not put in the time and effort needed to achieve this goal. But I have scanned a breadth of literature showing that such a task is doable if someone has a PhD length period of time to weave the strands.



              1. Hi Aaron and Joe,

                Thank you for this interesting thread. It is, to be fair, slightly out of my current expertise, but I’d like to suggest a book I found very interesting in this respect (consider I read it quite a few years ago): Ed Hutchins’s “Cognition in the Wild”. Do you know it? While there is not an evolutionary perspective, it is a great – if perhaps demanding – reading on culture and situated cognition.

                Hope this might be helpful, can’t add much more!


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s