So what is ‘social complexity’?
The key question that appears to remain largely unresolved after several days of talks is how to define social complexity. From my perspective, there were two main factors limiting our ability to satisfactorily answer this question. The first is because most formal definitions of complexity are fairly vague. For example, in my talk I defined ‘complexity’ as ‘when the whole is not predictable by the individual parts’, by which I refer to parts (i.e. individuals) as having fairly simple rules that together form more complex patterns of behaviour which their rules do not encode directly for. This definition nicely satisfies the need for generality (it is applicable to many fields), but is perhaps too general for gaining new insights in evolutionary biology.
The second issue I could see is that there was a tendency to define social complexity in a way that elevated researcher’s study organism to the top of the complexity ladder. Here there were two main camps: researchers studying social insects (extending to cooperative breeders) versus researchers studying group-living animals that contain multiple lineages (e.g. matrilineal societies such as hyenas or baboons). Tim Clutton-Brock captured the underlying cause of this issue most elegantly by highlighting a very clear negative correlation between species that have high organisational complexity (e.g. eusocial species) and those with high relational complexity (e.g. multi-male multi-female societies).
Over the course of the three-day conference, I found myself gravitating towards a definition that better linked individuals to ultimate drivers. My now updated definition of a social complex species would be something like: when the fitness of the individual is not only determined by its own behaviour, but also by the (social) behaviour of others. However, I would avoid ranking species by the degree to which their fitness is influenced by others (i.e. placing eusocial species at the top). Instead, I would consider individuals from species with low reproductive skew, but high influence by others, as having more potential to satisfy the definition that individual rules or phenotypic traits themselves do not directly encode variance in reproductive success.
(the above was written by Damien Farine)
Views from other lab members:
In the frame of this conference I gave my first talk as a PhD student. Thus, I had a very good chance to cope with the preparations that such a talk implies but also with the stress that beginners may experience. The audience gave me the impression that they were there to make the speakers think more deeply about the concepts and science they presented but not to judge or outcompete others. I was also very lucky to participate in this conference together with all the other members of our lab, who created a very co-operative, supportive but also fun environment. Overall this positive personal and academic experience of this meeting made it a great place to get feedback for my first PhD project; as it was a relatively small meeting with many senior experts in the field of socio-biology.
Many talks focused on defining social complexity and many others presented different aspects of social behaviour (i.e. what influences interactions with others and how these interactions determine individual fitness). By the end of the conference we had concentrated many definitions on social complexity which are not contradictory to each other but still differed. I was thinking that maybe we were all trying to define a term that it’s subjective by its nature: some may perceive a process as simple, others as complex. It could also be that all social systems are complex. Can you think of a social system that you could describe as simple? An example of this could be an asocial species, but even then, they interact with conspecifics when for example they fight for territories or find a mate – and these interactions could still have complex outcomes. So, before we define social complexity, we might need to ask whether there is social acomplexity.
The GFT conference was the first time during my PhD that I’d presented my research, and despite it being nerve-wracking, it was a great experience for lots of reasons. It was an awesome opportunity to present my work to some prolific researchers in the field, and get feedback on my ideas. After the talk I had some fascinating conversations with people from the audience, which opened up the potential for future collaborations – as well as getting me thinking about some questions that I hadn’t addressed. The thing I enjoy most about these specialist conferences is seeing the different ways that people tackle the same problem. In our case, how do you define and quantify social complexity? Even on the last day of the conference, this remained an open question, but my feeling was that at the least, researchers need to have a mechanistically-underpinned definition that links to their own question. Also a pro tip for presenters (based on an unfortunate personal experience): check your microphone is working before you start talking!
This conference was also a first for me, presenting my first poster on the early results of my master’s thesis. Although my work – a model of individual decision making – doesn’t deal with social complexity directly, I still gained a lot from the various perspectives presented, and from feedback from researchers working with the very real animals whose behaviours I aim to understand. Personal highlights included presentations on hyena’s climbing the dominance hierarchy in nepotistic systems (Eli Strauss) and work linking a single life history trait (the number of offspring) to social structure across the entire mammalian lineage (Dieter Lukas).
I would give my personal understanding about social complexity as this:
A magic black box where you mathematically input simple 1+1 and in most cases you get outputs that are mathematically unequal to 2, and what happened in the box is usually very difficult to infer with such output information.
I was one of the very few of our lab who came along to the conference without presenting a talk or a poster myself. This was on the one hand very convenient because I did not have to spend time memorizing a talk and didn’t have second thoughts on socializing in the evenings because I didn’t have to worry about being well-rested standing in front of an audience the next day. On the other hand, I did have something to talk about: I am currently working on my Master’s thesis for which I intend to develop a “framework for sociality” (or at least this is my working title), so the theme of the conference was highly relevant for me and I had a list of people from different fields I wanted to chat to about this, to get some new views on my ideas so far. Not presenting and being only at the beginning of the academic career path made me more anxious about approaching and talking to people whose names were on so many papers I have read, even though I’m usually not too timid. The small size of the conference, the friendly atmosphere and a number of socializing events after the official parts (plus a shout-out from Damien who was one of the first invited speakers to talk) all really helped to make me feel more comfortable and after I dared to speak to the first person I looked up to I felt much more confident. In the end, the conference gave me the opportunity to speak to quite a few people from different fields, from whom I got really helpful feedback and new ideas. It also helped me to get more of an impression which direction I wanted (or didn’t want to) head towards to and to personally meet people from labs I was interested in.
As a master student I had the extraordinary opportunity to have a talk accepted at the GTF conference. However, despite the presence of many important researchers in the room, the atmosphere was absolutely friendly and before giving my talk I wasn't feeling too nervous. I just enjoyed the moment very much. Then, after the talk I had the possibility to chat with many students and researcher interested in what I was doing and to receive important feedbacks and comments on my work. I can't say I have a very long experience with conferences, but the GTF conference was probably the one I have enjoyed most so far, because of both a very high quality of talks and speakers and because of the low number of people at the conference, which was ideal to meet and chat with almost everybody.
I think that the topic of Social complexity and complexity more in general, despite being hardly univocally definable, is somehow extraordinarily intriguing for Biologist. Indeed, Ernst Mayr, one of the prominent contributors in the development of the modern synthesis of Evolutionary Biology during the past century, loved to stress the uniqueness of Biology and its difference from all other sciences.
Biology, by definition studies nested level of complexity and have to deal with emergent properties as well as with continuous changes and discrete entities not easy to define (e.g genes, species, individual). For me, spending four days at the conference was also a wonderful occasion to reflect on the core of our discipline and on the reasons that push me to study and to do research.
Conferences are always a big highlight of my year, and this conference on social complexity was one of my favourites. It was the perfect size. One could attend every talk and read every poster. I always think conferences are important because publications are always describing the science that happened at least a year ago. At a conference, you can get a better picture of what science is happening now. What I love the most is meeting people who I previously only knew through their writing. Talking in person provides a much deeper picture into how people think and feel about particular scientific ideas (we are all more careful and conservative about what we say in a peer-reviewed article compared to what we say in a pub).
Re: definitions of social complexity
I don't actually think researchers of social behavior will settle on a single definition of social complexity that everyone agrees on. But that's ok. Instead, I think we'll have useful distinctions between different kinds of social complexity (such as Clutton-Brock's relational vs organisational complexity). The more you think about a topic, the more that topic starts to look like actually several different related topics. I think we'll also have even more operational definitions that define precisely how to measure each definition of "complexity" as a single value.
It seems to me that this is what happened with the most important concepts in biology. Take for instance "genetic relatedness". There's no standard definition, even for a topic that so many people study and so many of us measure. Depending on their research question, some biologists use "kinship" and "relatedness" interchangeably and others don't. Some people find the concept of negative relatedness useful and other don't. In addition to various conceptual definitions like "geometric relatedness" there are still more than a half dozen ways of calculating the relatedness between two individuals from the same genotype data. Even though every biologist agrees on what relatedness is in a vague sense, some definitions work better than others for specific situations or questions.
One unfortunate side effect of trying to come up with one singular definition is that each researcher thinks their definition is the "correct" one and that that other researchers that are using the term differently are in fact wrong, whereas they might simply be using the term in a different way. In my talk at the conference, I briefly described how this happened with the idea of "reciprocity" or "reciprocal altruism" in behavioural ecology. There are many debates in the literature about whether reciprocity is rare or widespread and whether or not this term applies in specific cases. But most of the confusion stems from the fact that there are multiple contradicting definitions, and each author feels confident that their definition is correct while being less aware of the popular usage of other definitions within different subfields.
In 1988, there was a conference calling together all the world's experts on reciprocity with the goal of precisely defining "reciprocity" or "reciprocal altruism" in a way that works for everyone and agreeing on how to test it. Sadly, it's been 30 years since that conference and people still disagree. So I guess my take home message is that it's important to define your terms and to know how different people define their terms. We can all strive for that, but a debate about what "social complexity" should mean might go on forever!