In the jungle of cultural complexity

The jungle—an area overgrown with dense forest and tangled vegetation—is not unlike the concept of complexity within cultural evolutionary research. The notion of complexity has been used across many branches of cultural evolutionary research to describe and explain key concepts, such as cumulative cultural evolution (CCE) and “complex” cognition. However, whilst interest from diverse research areas has enriched the literature, it has also led to a dense network of inconsistencies, which need to be addressed. What are researchers from different areas trying to describe and capture when they refer to “complexity”? The Culture Conference 2018 aimed to provide a platform in which researchers from different disciplines could further integrate their ideas and begin to untangle their views on the relevance and use of this concept. The University of Stirling, Scotland, played host to the Culture Conference 2018: “Complexity in Culture,” following a move from its previous home at the University of Birmingham where it was founded by Elisa Bandini and Dr. Eva Reindl. During the two-day conference (12th & 13th July) the 63 attendees (Figure 1) were treated to presentations of 18 posters, 11 talks (including four by invited keynote speakers), and an enlightening panel discussion hosted by Dr. Anne Kandler. A wide range of disciplines was represented across attendees and speakers. These included comparative, developmental, and experimental psychology; animal behavior; archeology; cognitive science; modeling; linguistics, and management. Cultural complexity was, therefore, considered from diverse perspectives (Figure 2).

panel with three main topics for discussion (see sections in italic), which encompassed themes highlighted throughout the aforementioned talks and posters.
When asked to provide an explanation as to how complexity is used in the examination of culture within their discipline, panel members came to no consensus. Rather, some researchers stated that they focus on the complexity of the product, and others the complexity of the process whereby the cultural product is created. For example, in archeology the complexity of a product is often represented by the number of technounits it contains or by an inferred minimum number of intermediary steps, but as Tennie stated, archeologists do not have the luxury of being able to directly examine the processes involved in product production. Conversely, Emery, from the field of comparative cognition, specified that when he considers complexity he is generally thinking about "complex" cognition. That is, higher-representational levels and how individual units build upon/interact with each other, rather than perceptual responses to stimuli. Whiten described how in his research, better representing the complexities of the real world has been a major goal. For example, in his recent research exposing children and chimpanzees to microworlds affording cumulative cultural change, complexity is multifaceted because tools, actions, and even the decision process needed for task success, exhibit rising levels of complexity. 1 Even researchers within the same field (animal behavior) differed in whether they considered complexity in terms of the process or the product, highlighting the divisive nature of this widely used, yet poorly defined, term. Caldwell (previously inspired by audience member, Dr. Mark Atkinson, University of Stirling) suggested that the term CCE could be reserved for describing the process which produces accumulated products, and the products themselves referred to as cumulative culture (CC).
Bandini and Reindl had previously highlighted the issue of complexity in process versus product in their keynote talk, questioning the necessity of increasing complexity as a criterion for CCE. Showing pictures of a lighter and a match, they asked the audience which came first, before revealing that the lighter came first. This simple yet effective example was used to highlight that the evolution of some cultural traits involve decreases in complexity, suggesting that increases in complexity of the product should not be regarded as a defining feature of CCE. However, it was noted that an increase in efficiency of product use may require an increasingly complex production process, which confuses matters somewhat.
How do these different researchers measure the aspects of complexity most relevant to them? And, when considering CCE, are we more interested in relative measures, rather than absolute magnitude? If "complexity" within cultural evolutionary research is considered differently by different disciplines, is it a concept which can be consistently operationalized?
It seems there is no simple answer to these questions surrounding the measurement of complexity, especially when comparing across disciplines, described by Kandler as the "gold star challenge" in this field. Emery pointed out that the measurement of complexity appears to be domain specific, and that attempting to "simplify complexity" into a measure that can be used across disciplines would surely be detrimental to the field. In the same vein, audience member Professor-Christian Rutz (University of St Andrews) described comparisons across taxa as the "real challenge" and highlighted the difficulty presented by natural constraints, which differ between species. He therefore, suggested that researchers in different disciplines choose the metric which best allows them to measure directional change in their chosen system before comparisons across systems are made. Whiten showed, however, that all is not lost and useful comparisons can be made across different species when he referenced work by Claidière, who is using algorithmic measures of complexity similar to those used in linguistics research. Caldwell also brought linguistic research into the discussion in reference to the clearly defined metrics used. These have nonetheless failed to resolve debates between linguists as to "what is a complex thing." Caldwell suggested that as a field it would be helpful if we could agree on a broad standard, which, as a minimum, allows us to state that one thing is more complex than another is, allowing for greater communication between disciplines even if the way complexity is measured differs between them. How easy this will be, is yet to be determined…. As Watson commented, there are many different aspects of complexity to consider. In addition to the complexity of the process and the product, we must also consider the cognitive demands imposed by different tasks.
Prior to the discussion, the panel members and audience alike voted on the question below using an interactive system.
A defining feature of CCE is often cited as an increase in the complexity of traits/behaviors across generations, but does the term "complex" fully encapsulate the changes we want to describe? Some authors have identified additional means whereby a trait/behavior can change over generations to become more useful, therefore, should we be describing the changes to traits/behaviors as an increase in: • Complexity only • Complexity and/or efficiency • Functionality-as an umbrella term incorporating complexity, efficiency and others, for example, security, convenience • Robustness • None of the above The overwhelming majority of delegates voted for "functionality," with about 10% selecting "complexity and/or efficiency," and a few brave individuals choosing "none of the above." This vote highlighted that many academics are dissatisfied with using complexity to describe modified traits/the process during which traits are modified in the context of CCE, and therefore the need for further discussion.
Issues with using the term complexity were emphasized earlier in the conference by Bandini and Reindl in their keynote talk-"Complexity: an inefficient concept for the study of culture." They presented three problems: inconsistencies in assessing the degree of complexity for the same trait across different domains (for example, artifacts, knowledge, behavior, cognitive processes), the subjectivity of judgments made about levels even within these domains, and lastly, that increasing com- The question immediately sparked comments as to the criteria we use when defining CCE. Whiten started the dialogue, voicing his opinion that complexity need not be thought of as one of the essential criteria when examining potential instances of CCE; rather, very basic or core criteria like those of Mesoudi & Thornton 2 -repeated build-up of socially learned traits-can provide a definition of CCE, then how other features, such as complexity, may or may not emerge, can be further studied. In a supporting point (and in line with thoughts presented by Bandini and Reindl), Rutz reminded us that when talking about the accumulated products of CCE we are often referring to something which has undergone a decrease in complexity together with a paralleled increase in efficiency. Kandler related this point to the idea of robustness, stating that a trait may need to decrease in complexity in order to remain robust, that is to remain within a population and be transmitted between individuals. Once the trait has been adopted, it could then increase in complexity once again, expressing an emergent balance between robustness and complexity.
Noting that the panel members were relatively satisfied with labeling a cultural product/process as CCE without necessarily observing increases in complexity, Caldwell posed the questions: "Can we envisage a situation where complexity might increase but we might say that's not cumulative culture? Is that definitely cumulative culture if complexity is increased, or are there circumstances in which that could happen and we could say-no it's not?"