Data Meets Durkheim

Ever since Durkheim’s Elementary Forms of the Religious Life (1912), most scholars have assumed that collective rituals function to arouse emotions and bond people together into more cohesive groups. In this post, I commented on this assumption and the remarkable fact that Durkheim’s big idea had never been tested or supported with anything other than qualitative observations. The post also noted that the idea was in fact being tested by Dmitri Xygalatas and that this seemed to be a nice start. In response to all this, Dominik Lukes left a comment which bears close consideration:

My first reaction was “proof, what proof.” [Durkheim’s idea] is so obvious, no additional proof is needed. And indeed, we could easily think of this in an axiomatic way. But on reflection, I think that it is far from obvious that “collective rituals function to maintain the cohesion of groups or society.”

This notion has such wide appeal because it resonates with so many situations all scholars can relate to both personally and professionally. It is hard to think of groups of any size that don’t develop some ritualized behaviors over time. But I could easily think of at least two other ways in which this could be explained.

1. Since all groups have to have some cohesion to be groups in the first place (groups with no cohesion are not groups but just random groupings), rituals could be an expression of that cohesion rather than a necessity for it. This would explain why those who express stronger attachments are more likely to take part in the rituals.

2. Rituals could be an expression of something else (e.g. human tendency toward repetitive behavior) and groups simply have to accommodate them. They may become carriers of cohesion once they’re in place but they don’t do anything to maintain it.

But I don’t think you can test for these hypotheses because you always have to find some proxies – such as expressing a feeling of belonging – and make some testable predictions about them. But because group cohesion is such a complex thing (frankly we don’t have an unproblematic definition of group or ritual – let alone cohesion) you can’t formulate any definitive set of testable prediction types. Or it will be so large you’re bound to meet some and fail others simply by chance. But I think we can take these scenarios as metaphors and play them out on any given piece of evidence and see what that can reveal about them.

Can Durkheim’s idea be tested? In this recent Aeon article, Dmitri Xygalatas explains his research and findings in much greater detail. His methods and results appear sound, though the intensity of the rituals he has studied may limit any larger inferences. It’s a fantastic article, but while reading it I kept thinking about Dominik’s comments (above) and the alternative ways in which Xygalatas’ results might be construed. Regardless, some testing is better than no testing, and it’s nice to see data meeting Durkheim.


Did you like this? Share it:

3 thoughts on “Data Meets Durkheim

  1. Chris Kavanagh

    I actually conduct research in this very area and to offer some attempt to address Dominik Lukes comments:

    1. I don’t think anyone argues that ritual performance is a necessity for ALL forms of group cohesion. Just that it works as a method to achieve this. One empirical way to test this would be to get individuals to perform rituals or ritualised actions together and compare them to people performing other actions designed to be similar in form but shorn of ritual aspects i.e. present the same actions as clearly instrumental. If limited forms of cohesion could be formed in such artificial conditions it would be a useful preliminary indicator that ritual performance is generating cohesion rather than solely a consequence of it. There are lots of other ways to come at this, including using correlational research at actual ritual field sites (i.e. taking the same measures pre and post performance).

    2. This is why it is useful to deconstruct the various aspects of ritual. Synchronous performance, for instance, or their repetitive nature as there are variations in these aspects between rituals (i.e. some are actually performed very rarely and are novel for most participants). These factors could/are be examined both experimentally, using correlational methods with actual ritual performances and through analysis of archaeological and ethnographic records (see Atkinson & Whitehouse 2011 for example).

    I also don’t really get why testing hypothesis and using proxies to address large collective concepts is a bad thing.. isn’t that how we find out if our theories are correct?

    Anyway, Dimitris article was great and provided a nice coherent narrative to his experimental work. I suspect/know the reality is slightly messier than the article suggests but I think it’s always good when scholars can provide a coherent account of their body of research.

  2. Dominik Lukes

    Thanks for the shout out Cris. I actually read that same Aeon article and didn’t even make the connection. The thrust was ‘people take part in rituals, stuff happens to their bodies and minds both directly and indirectly’ so I was not all that intrigued. It strikes me as incredibly useful piece of the puzzle and certainly provides important bits of evidence.

    But what was missing in this (and the whole entreprise of classical anthropology) is the notion of heterogeneity of both experience and interpretation (conceptualisation). I actually saw firewalking first hand many years ago in Bulgaria. It was put on for a group of Peace Corps trainers. I wager that for most of us the experience was similar to that of any ‘folklore’ exhibit. Mild bemusement and slight mental discomfort. It was no different from a magic show or tightrope walking. An exhibition of a great amount of skill and some risk but no sense of a deep cultural engagement. But there may have been others who would have been deeply touched. But that reminded me that what’s not a part of this narrative is the story of the guy who cleans up after the ritual the morning after. What’s missing is the ‘progressives’ who (through contact or internal analysis) think of the ritual as something symbolically inappropriate. The people who genuinely feel oppressed by it and are pressured into participation. This is not to advocate for an individualist point of view. The social, group perspective is incredibly valuable. But it exists alongside the others – not in competition to them.

    In response to Chris’ comment. This is my problem with what Xygalatas is doing. Not the data he’s collecting or how. Or not even the interpretation. It’s the spirit of trying to make anthropology somehow more ‘scientific’ because its insights are “a mile wide and an inch deep”. I’m all for postulating research objectives in ways that can be tested by further research. But not to shore up the thick description type of stuff but to provide another perspective (model, metaphor, …). It’s a mistake to think that measuring people’s heart rates or scanning their brains somehow gives us deeper knowledge about their nature. (Not that Xygalatas necessarily does that other than by implication.) It certainly does not allow us to make conclusions like:

    “One important takeaway is that when ideology and collective arousal merge, emotional reactions can spread like wildfire through a crowd.”

    There’s nothing in Xygalatas’ research (at least how described in the Aeon article) that actually describes this situation. It simply describes a conventional view of the crowd (going back to at least le Bon) adjacent to quantitative evidence based on proxied measures such as heart rate.

    Note: This is not an anti empiricist attitude. I’ve been equally critical of anthropological lyricism. This is an argument for a very strict empiricism.

    Note 2: Testing hypotheses and using proxies is definitely a good thing. But it is also absolutely not how we find out if our theories are correct. It is how we determine the limits of the utility and applicability of such theories.

  3. Chris Kavanagh

    Dominik> Thanks for the response.

    I find little to disagree with in your recommendations but I do think that there are empirical ways of addressing some of the issues you highlight. There will inevitably be differing reactions to ritual situations by observers and participants who have different levels of involvement with the relevant group or even just personality/demographic differences. I think in Dimitris research he touched on this with his HR research by showing that heart rate alignment was not evident between all observers/performers of firewalking rituals but rather only appeared in those who had a relational tie with at least one other firewalker. It’s also worth noting that this study found such an effect to be occurring BEFORE the firewalking had taken part, which strongly suggests it is not just the physical participation in such an event which matters. This is just one piece of data but it shows these matters are not entirely ignored. I also think it would be worthwhile to collect data from groups with differing levels of involvement and commitment and to try and parcel out the mediating effects of things like obligation and devotion (in fact thats what my present research involves). However, these are not issues that are impossible to address or to control for, to some extent, using empirical methods.

    Heterogeneity is real, but in modern social/cultural anthropology it is often emphasised to the point of absurdity, obscuring evident trends within groups/cultures due to a theoretical commitment. You don’t have to think of groups or cultures as being entirely bounded heterogenous entities to discuss dominant trends within them.

    Oh and in response to both Note 1 and 2. You are correct, it is good to hold people’s claims to the empirical fire, so to speak and, while I was speaking in layman’s terms, I agree that technically it is always inaccurate to suggest that, using the kinds of methods that cognitive anthropologists use, that a theory is ever ‘proved correct’.

Leave a Reply