Our first step in thinking about how to teach Durkheim to undergrads and graduate students involves discussing and evaluating his descriptive model (see Part 1 for background context on this essay). Admittedly, there will be times where I have to “peek” at the next essay, Part 3, because his descriptive model is difficult to untangle entirely from his explanatory model.

I would be willing to bet that the average sociology reader knows there are four types of suicide, according to Durkheim – egoism, anomic, altruistic, and fatalistic. I’d also be willing to bet that he actually only had three types of suicide (the first three in the list), but a seemingly last-minute decision on his part and the crystallization of a contemporary interpretation obscured this fact. Wait, what?!? Go back, re-read Suiicdei and you’ll see that these three types are the only three types he mentions save for a single paragraph in a single footnote at the end of the last substantive chapter (more on this soon). In what is called “Book III,” where Durkheim shifts to his big theoretical argument – something, strangely and unfortunately, overlooked by most sociological theory texts – fatalism is nowhere to be found. Rather, he conceptualized societies as either balanced in a ‘tug of war’ between three social currents: egoism (or liberalism), altruism (or intensive mechanical solidarity), and anomie (or acute/chronic dysregulation). The dominance of the first meant the cult of the individual was winning out and people were ‘bowling alone’. The second meant that society demanded people take their life under certain conditions (e.g., martyrdom). And the third meant moral anchorage was temporarily or indefinitely impossible due to some disruption in political, economic, and/or domestic life. These social currents (or social facts) are more like public opinions or dominant beliefs that spread through a population and were the etiological roots of higher-than-average suicide rates (more on this soon, I promise). Hence, three types were the foundation of his descriptive classification system and the explanatory model.

******

[Quick Aside for the Uninitiated Reader or for a Refresher: Durkheim’s theory is rather parsimonious. In essence, it hinges on two principles. First, the structure of suicide rates is a positive function of the structure of social relationships within a collective, broadly defined (groups, classes of people, geographic communities). Second, the structure of social relationships varies in terms of how integrative and regulative they are. The former is usually interpreted as social connections and solidarity, while the latter is about the group’s moral order and its ability to guide behavior consistently. When groups are not integrative or not regulative enough, or too much, then members become vulnerable to suicide. Why they die is not important, as subjective, specific individual decisions for Durkheim have no scientific merit. It is structural, macro, and beyond anyone’s individual choices.]

******

So, how did we get to fatalism? I’ve never dug deep on this, but if Durkheim was like you or me, as he was preparing his manuscript for publication, I am sure a friend, colleague (maybe his nephew Mauss?), or even his own sharp self noticed a potential “reviewer critique”: on the one hand, he had symmetry in egoism and altruism representing two opposing types of suicide based on integration. In that sense, he had met the criteria for a good descriptive mode (exhaustive and exclusive types with, potentially, clear explanatory power). But, anomie was its own thing. What was its logical opposite? In a hastily written – and that is the only conclusion one can draw from a crappy footnote in an otherwise exhaustively researched set of substantive chapters – Durkheim made a careless mistake and added a random type (fatalism) for, what I can only surmise, the sake of symmetry (it’s on p. 276 in the 1951 translation for the interested reader). I joke in talks that I dream that a footnote I write will become a full-fledged totem despite its undeserving nature. In elevating this fourth type, modern sociology (ironically) tossed aside Book III and the explanation Durkheim laid out, in favor of a four-fold model that looks nice and clean (see the figure below from Bearman 1991) but was not actually what Durkheim appears to have intended at all!

The lesson, here, is that exegesis is only as good as the exegetes and the degree of freedom contemporary exegetes have – structurally and symbolically – for dealing with old texts. Thus, smarter (and currently dead) famous sociologists treated Durkheim’s work as having four types which subsequently led to this model’s enshrinement in texts books purporting to teach classical theory which ultimately made someone like myself’s protestations fall on deaf ears (“Who are you to question Durkheim’s greatness!?”). The excavation of a four-fold model was not like paleontologists filling in an incomplete skeleton with drawings or fake bones, but rather an archaeologist recovering a fragment of a text and adducing an entire story from a sliver of writing. We invented a whole type and now must live with this situation and, because the model and the theory are a “settled” matter, we cannot really develop the four types beyond the superficial reading that most theory textbooks have ingrained in their students’ heads.

It is catastrophic in the sense of scientifically approaching suicide on two grounds. First, as my old advisor was fond of saying: once you’ve wrapped your favorite text in velvet, placed it on the altar, lit the incense, and prostrated yourself, there is no self-correction, no modification, no advancement. The ideas are set and we must contend with the taken-for-granted wisdom of the common (mis)interpretation of Durkheim’s Suicide. Second, because we are frozen, we are left with a purportedly general classification model that appears exhaustive and exclusive, but which raises doubts about both criteria. To put it differently, sociology has struggled to contribute to the study of suicide in a meaningful way, but this is a microcosm of many of the things we do. Teaching this to grad and undergrad students simply reproduces this problem, when the real lesson this text offers is how to evaluate theory, how to advance theory, and how to abandon the things not worth keeping anymore.

In that spirit, let’s look a little more closely at some of the issues presented by this four-fold model.

The Curious Case of Fatalism

Adding fatalism makes the model elegant looking, visually. It feels complete. But, what exactly is fatalism? Is it a worthy type? Durkheim resoundingly answers in the negative. In that footnote/paragraph, Durkheim notes there is another type of suicide but quickly tells the reader it is insignificant, sociologically, and only perhaps of interest to historians. He then anecdotally – which, again, is strange given the effort he puts into empirically validating his other types – lists three historical cases: slaves, middle-aged childless wives, and newly married men. The first case is always listed as the exemplar. Makes sense. What class of people are more overregulated than slaves? But the other two cases are headscratchers as he has no explanation for why he chose them or why they are fatalistic. We can fill in the blanks, I suppose. The first case, which may strike the reader as erring on the side of misogyny, probably comes from his own circle or extended network. Women, in those circles, would have largely been expected to be dutiful hood ornaments to their upwardly aspiring husbands and their essential role was having children – preferably boys, but children to transform the couple into a respectable family. Any middle-class family worth their salt needed kids, and any woman entering her 30s would need kids to maintain her social relationships and family’s status. A childless woman would be a social pariah. Her friends might shun her, while her husband may but distance between them. Consequently, we can presume many of these women became locked in a loveless, unhappy marriage or, worse, threatened with divorce and spinsterhood. All of this may be what Durkheim was thinking; but, we will never know. It is a fair question to ask how analytically and empirically similar this case is to slavery. They do not really seem to share etiological foundations and, even more vexing, they differ morphologically in very serious ways. These differences are beyond my interest, for now, and thus this discussion will await some future essay.

The third case, bachelor-turned-married men is also likely anecdotally drawn from French newspapers, gossip, his own circle, or just ‘common sense’. Like a slave, newly married men are suddenly “choked” with obligations, though for a relatively short period of time. In this sense, they go from a state of relative freedom to one of obligation, but is it akin to slaverly? Marriage often carries positive status enhancement as well as stability and support (after all, men seem to benefit from marriage as protective against suicidality). I won’t beat the proverbial horse, though, and let the reader decide, however, what to make of this and whether or not fatalism makes any more sense given his three sorts of parallel, mostly not cases. Beyond the empirical weakness of the type, there are theoretical dilemmas too.

The real trouble for making sense of fatalism as an authentic type emerges when we consider it within the context of his chapter on anomie – the only type he appears to have intended to link regulation and suicidality. The common interpretation, thanks to Parsons (1951), is that deregulation means normlessness or directionlessness (the word anomie, in Greek, does mean “a” – lack of “nom” law or order); and normlessness causes negative outcomes. The question, though, is what is the causal force of anomic suicide? Is it too little regulation? A curious puzzle arises when we consider that Durkheim believed economic booms and busts were two disruptions capable of deregulating society or sectors in society. Most people have studied busts, but would busts not contribute to the narrowing of choices and delimiting of aspirations? The choking of futures? Sounds a bit like anomie may include having too many moral dilemmas and the sudden loss of moral flexibility. What this suggests is Durkheim made a distinction about deregulation that was different from the egoism-altruism continua. There was no need for symmetry because the etiology of suicides caused by regulation had less to do with the structure of social relationships and everything to do with their disruption and disintegration and the sudden moral ambiguity that created. Fatalism radically re-orients the theory, causing the Parsonian version of normlessness to become the explanation, which may account for the surprising lack of empirical verification of Durkheim’s most famous type of suicide – especially considering how universally supported egoism is across disciplines.

This issue, however, points to two key takeaways. First, it is a great teaching moment about conceptual clarity. Anomie and egoism, regulation and integration, were never defined clearly. You can’t fault mid-century sociologists for trying to operationalize them while working with vague source material. It also reminds us that common interpretations are inextricably tied to a community of scholars and are not always correct. The classics suffer from being Rorschach tests. They would likey never be published today, because they fail at specifying the theoretical logic in precise, parsimonious fashion. But, for graduate education, we should be training them to see these flaws and work out a better, tighter theory while undergrads should be taught about theory instead of classical lore.

The second takeaway is that regardless of our own interpretation today, the addition of fatalism violates the exclusivity criterion because Durkheim probably never intended to have fatalism as a type. This is all fine and well and can be easily remedied with some logical exposition, but it suggests we have propped up a theory whose foundations are flawed. The conceptual ambiguity, however, surrounding Durkheim’s intended ideas about anomie and regulation, and the fact he saw it necessary to even add fatalism, also raise questions about the veracity of the three-fold model. Of course, it remains difficult to imagine sociology changing course and adopting the three types instead of the four types, or whether that would make a difference. In fact, we must applaud those who have massaged these inconsistencies to produce legible and compelling accounts, like Bearman’s (1991) and Pescosolido’s (1990; Pescosolido and Georgianna 1989) network approaches. But, too often we gloss over the inconsistencies without taking in the lessons they teach graduate students and the harm it has on the science of suicide. My colleague and I once wrote a paper attempting to totally re-theorize fatalism (2018 – also stored elsewhere on my website) and something peculiar happened. The editor of the journal loved the paper and noted, ironically, that it was about time someone addressed fatalism, while one of the reviewers pushed us to omit most of the reference to fatalism in the service of building a more comprehensive theory of structural and cultural regulation. Such a fascinating interplay between two very bright scholars. In the end, my sense is there is room for keeping fatalism, but perhaps the reviewer was right: the labels themselves are doing more harm than good. But, for now, I will offer one more example revealing the exclusivity problem.

Is it the Dogma or the Moral Community?

Durkheim’s most famous thesis was that Protestants died by suicide more than Catholics. He surmised that the former invited individualism, free thinking, fewer obligations to fewer people, and weaker communities. So important a thesis, Merton (1967) referred to it as sociology’s one law – nevermind extensive research, some using Durkheim’s own data (Halbwachs 1978), pointing to a lot of caveats with this so-called law and also showing the inverse in some cases (Travers 1990; Pescosolido and Georgiana 1989). Why it isn’t likely a law speaks to some interesting problems with his typology and, specifically, with egoism as a type.

The argument, of course, is that religion offers a moral community in which people find strong support and protection against pathology. It is important to note, that sometimes in purposive or unintentional defiance of the common interpretation of Durkheim, researchers presume protection comes in the way of the prohibitions some religions offer against suicide. A mortal sin, in Catholicism, backed by stigma, threats of excommunication of one’s everlasting soul, and, in medieval times, the actual public degradation of the whole family and the reduction in its social standing (Barbagli 2015). Indeed, this operational confusion points to the theoretical confusion surrounding religion – as well as the degree to which modern sociologists are willing to accept orthodoxy despite obvious alternative theses (which is a good thing!). This confusion, good for science as it may be, does undermine the exclusivity of the model. This is particularly problematic because Durkheim cannot settle on the explanatory logic of religion and suicide (see 1951, pp. 209ff.). He initially argues it’s because of the moral community, but then he argues the moral community provides regulative capacities (duh!), thereby making integration the direct and indirect cause (through regulation) of suicide rates. So, is it integration or regulation, because if it were the latter or some mixture of the two, then we cannot retain the notion that there are discrete types and, consequently, a true classification system.

Well, we could – Weber’s famous typology of action and of domination present ideal types that do not actually exist in empirical reality, but serve to study variation in real cases and the consequence of this variation. But, for Durkheim, these are ontologically real phenomena; the are social facts.

The point, ultimately, is that one can (and some have [Johnson 1965]) make the case that there is only one reason for suicide: disintegration, or a lack of integration (see also, Halbwachs 1978). Either social organization and networks are in a state of connective scarcity or a process of declining slowly or suddenly. But, that is not how we interpret Durkheim or how he is taught, or how most reviewers tasked with evaluating research tasked with testing these theses understand the conceptual or operational definitions. So…yeah, it’s not good.

Perhaps, though, you disagree with my first and second contentions. Fair. But, I have one more trick up my sleeve with regards to the exhaustive nature of the typology.

Suicide Suggestion, Contagion, Clustering

In the 1970s, David Phillips (1974) “discovered” the study of suicide suggestion, or what he called the Werther effect. In the 18th century, Göethe published The Sorrows of Young Werther in which the eponymous protagonist dies by suicide because of unrequited love. Following the publication of the book, several copycat suicides first within Göethe’s social network, then in the town and eventually the surrounding area happened in relatively rapid succession. The authorities banned the book. Borrowing the name, Phillips identified a strong association between being exposed to newspaper articles about suicide and subsequent spikes in suicide rates among the audience. The longer the exposure and greater the celebrity, the bigger the effect (Stack 1987). For decades now, using increasingly conservative and sophisticated statistical methods, this relationship has held firm (Stack 2004). And, seems to be even stronger when the exposure is relational, such as being exposed to a friend or family member’s suicidality (Abrutyn and Mueller 2014; Mueller and Abrutyn 2015).

Phillips astutely noted that Durkheim had rejected the idea of contagion as a mechanism or cause of suicide (see 1951:123-142). Phillips turned, instead, to Durkheim’s rival who he famously buried in public debates – Gabriel Tarde (Abrutyn and Mueller 2014). Tarde had written a book on the “laws” of imitation, which Durkheim ridiculed. (Admittedly, Durkheim’s argument is far more nuanced as he discerns between three definitions of imitation and accepting one as a real possible explanation of suicide, but not sociologically important. One more aside: after vanquishing Tarde, I would argue he stole his ideas when he began writing about the contagion of sacred beliefs and emotions in The Elementary Forms of Religious Life). With some theoretical ingenuity and synthesis, making Durkheim’s work fit a model of contagion is possible (Abrutyn and Mueller 2016), but it is neither implicit or explicit in his descriptive or, as we will see explanatory, theories. It is totally absent.

Not surprisingly, it wasn’t until the 1970s that someone even noticed the correlation between exposure and suicidality, and not for several more decades that someone bothered to demonstrate the relationship causally (Abrutyn and Mueller 2014; Mueller and Abrutyn 2015; Mueller et al. 2015). It is still not something taught in sociology courses unless they teach the small boutique literature that includes my own work with Anna Mueller and a few random papers by others (Bearman and Moody 2004; Baller and Richardson 2005, 2009). There is Durkheim and then nothing else; which may explain why sociology has barely contributed to suicidology (the interdisciplinary field) or the practical policy interventions designed to increase safety and reduce suicide.

Long before Phillips, anecdotal evidence of suicide epidemics was known. Durkheim was aware of small villages, sentry towers, regiments, and medieval monasteries as sites of tragedy. But, social science had also cataloged these sorts of events, unsystematically of course, a few decades before Phillips. The tendency for suicides to cluster in temporal and geographic space, ironically, is the most sociologically interesting case of suicide given some places are vulnerable and not others (Haw et al. X). That is, some Indigenous communities are sites of suicide clusters, but not all (X); some high schools have clusters, but not all; some prisons have suicide clusters, but not all. The variation in social environment seems eminently relevant to a Durkheimian or just an average sociologist interested in the social distribution of pathological behavior!

Durkheim’s theory, however, fails to even make sense of these. One cannot study what isn’t in the purview of the theory motivating research questions and puzzles. And, thus, the exercise of both identifying the descriptive theory that Durkheim lays out and evaluating it provides major lessons for graduate education and also questions whether classical theory is worth doing any more or at least in the conventional sense. In my last essay, I will return to the question of what do we keep and what do we toss. The third part of this series, however, will explore the explanatory model as the most important criteria of any descriptive theory is whether it produces good or bad explanations. Durkheim’s, to foreshadow this assessment, is mixed, but does do better than his descriptive model.