Features, Politics, Social Science

Beware of Root Causes

What does it mean to claim that something is the ‘root cause’ of a problem in society?

It’s common, for instance, to hear the assertion that ‘the root cause of terrorism is Western foreign policy’. The implication being that the responsibility for terrorist attacks ultimately lies at the feet of the West since its interventionist foreign policy has destabilized the Middle East – irrespective of any other source of causality. The phrase ‘root cause’ invites us to become privy to society’s underlying pathologies that, if remedied, could improve the world beyond the scope of someone merely observing the surface.

Much like a bug in a software program causing a computer to shut down, or a leaky pipe causing subsidence under a house, the language of root causes implies that problems can be traced back through a cascading chain of events to an initial fault. Under this assumption, our goal should be to target that underlying problem rather than the ways in which the problem manifests itself in society.

A framework based around root causes is alluring for several reasons. It appears to simplify complex problems into pieces that can be more easily targeted, as a root cause may be easier to deal with than the broader problem that stems from it. There is also the possibility of killing two or more birds with one stone, as multiple problems may have the same root cause. Perhaps most inviting though is the appearance of profound insight into the nature of a problem, and a framework for how one should proceed thereon. Consider this passage from a paper advocating for ‘root cause analysis’ when attempting to understand and solve large-scale social problems:

Once a problem’s root causes are found, night becomes day. What was a murky cloud of befuddling complexity becomes a social problem structure so clear that its correct solutions are obvious.

If only it were so simple.

Tracing a lineage of causes back to their root may work reliably in non-complex, relatively closed systems. A well-trained mechanic can normally diagnose the root cause of why your car has broken down so it will continue to run well in the long term. But for complex social problems, the task of finding something that could be described as a root cause is far more difficult. Most social problems are multifaceted, with different causal layers and dimensions interacting in ways that are not reducible to a single phenomenon. They can co-evolve with other social problems in ways that are counterintuitive and unpredictable. More often though, the concept of a root cause is deployed in ways that serve particular discursive — and often political — ends.

Consider the reality of domestic violence (otherwise known as intimate partner violence), which has received significant attention from government and media in recent years. Australian state and federal governments have allocated substantial funding to programs that claim to address the ‘root cause of domestic violence’; allegedly, gender inequality. It’s thought that because the majority of perpetrators of domestic violence are men, then the underlying driver must be men’s views toward women, and the socio-cultural environment that supports and enables such views.

Under these assumptions, it is necessary to address the root of the problem if we want to eliminate gender-based violence permanently. It also happens that this view aligns with a feminist perspective that views gendered violence as fundamentally different from other forms of violence, and therefore a manifestation of a patriarchal social order. Addressing the root cause of DV would ostensibly kill two birds with one stone.

It’s easy to see why the prospect of uncovering and describing root causes is so attractive to researchers, activists, and policymakers. For researchers, seeking to identify the root cause of a social issue appears intellectually ambitious, revealing the base social structure of what we experience in everyday life. For activists, identifying and tackling the root cause of something is likely to be far more satisfying than merely addressing the symptoms. For policymakers in government, the prospect of funding programs and initiatives that address the root of a problem is also appealing, given that it may reduce their need for future expenditure in dealing with the manifestation of the problem in society.

In the case of domestic violence, the gender-inequality-as-root-cause hypothesis has come to dominate the policy and activism space, in contrast to rigorous academic research that acknowledges a range of contributing factors such as poverty, mental illness, alcohol abuse, personality pathology, and poor verbal ability – none of which can be described as ‘root cause’. It can also be falsified relatively easily, if only one has the inclination to do so. Same sex couples – where, by definition, gender inequality cannot be a contributing factor – experience similar rates of domestic violence as heterosexual couples. Developed countries that experience the highest levels of gender equality, primarily in Scandinavia, also experience above average rates of domestic violence than the rest of the EU. It’s difficult to see how something could be a root cause for a societal problem when the problem persists in contexts where that variable is removed or minimized.

Jordan Peterson also encountered the gender-inequality-as-root-cause hypothesis in his infamous interview with Cathy Newman about the gender pay gap. Newman claimed that the root cause of the gender pay gap was discrimination against women. Peterson objected to this, not because there is no discrimination against women, but because it’s merely one factor among many that contribute to an overall disparity in pay between men and women. Peterson understood that these sorts of problems are multivariate, with many different factors contributing and interacting in ways that are difficult to separate from each other. To parse these multiple contributing factors requires relatively complex statistical techniques whilst working with comprehensive data sources, something that too many journalists and activists show very little interest in doing. And crucially, it is impossible to legitimately characterize any of these variables as ‘root causes’, the elimination of which would rectify the problem as it manifests in society today.

Claims about knowledge of root causes should not go unquestioned. When they are encountered we should ask: does this social problem actually have a root cause, or does it have multiple factors that coalesce to contribute to it? What evidence is there that the purported root cause of a problem precedes and gives rise to other relevant phenomena? And perhaps most importantly: what political or discursive purpose does it serve to claim that something is a root cause?

Diagnosing a root cause for some societal ill is almost always a political move designed to draw attention toward certain explanations whilst simultaneously drawing it away from others. It attempts to present certain causal factors as having a primacy that is at best unfalsifiable, and at worst flatly contradicted by real-world examples. Speaking of root causes of complex social problems conveys an epistemic authority that is often not earned.


Andrew Glover is a sociologist based in Wollongong, Australia. He tweets at @theandrewglover


  1. Jonathan says

    I think this is a fine article, but readers should be careful to avoid conflating people who consider intimate partner violence a feminist issue and people who view it as a product of misogyny.

    • Gregory Bogosian says

      So what is the actual difference between those two views?

  2. ga gamba says

    I think this essay does well describing the root cause of the root-cause problem.

    A well-trained mechanic can normally diagnose the root cause of why your car has broken down so it will continue to run well in the long term.

    This is a closed-system, so a root cause can identified therein.

    • TarsTarkas says

      The root cause of any automotive mechanical breakdown is white patriarchy, because the overwhelming number of inventors of automobiles were white cisgender males/sarc.

  3. dirk says

    Here speaks the gardener: it’s a pity that the illustration above this piece shows a type of plants that does not need roots to propagate, cuttings of stems or even just leaves will do. The plants are the sedum type of stonecrops (among others) , well known to people without green fingers, because they care for themselves, even without soil and regular rain or watering. Sedum nussbaumerianum is one of them, I’m not joking, and metaphore intended.

    • Wilson Hill says

      “…does not need roots to propagate…”

      Good catch, but I’d say it actually seems rather appropriate.

      • dirk says

        Survive even on a roof in dry climates, so, no good example of plants needing deep rooting system, as neither is e.g. tumbleweed. I,d put a maize plant with its extensive root system.

  4. Robert says

    I wouldn’t say it’s impossible to legitimately identify statistically measured attributes as “root causes” unless that is to say there will always ever only be one “root cause”, which as you aptly pointed out is extremely unlikely. It’s just really… really hard and time consuming (potentially to the order of decades). Also, saying it’s almost always a political move to identify root causes is making the same error you’re cautioning against, after a fashion.

    Good article otherwise. It’s always proper to be extremely skeptical when someone is trying to sell silver bullets.

    • TarsTarkas says

      If they are selling silver bullets, either the Lone Ranger has retired or there are werewolves about!

  5. nicky says

    There is nothing wrong with root causes as such, only with wrongly identifying them. In case of partner violence, it must be deeply rooted in our biology. Things like sexual jealousy (and what are the root causes of sexual jealousy?).
    Personally I don’t think that partner-violence is caused by gender-inequality, but both may possibly share some root causes. 🙂

    • AC Harper says

      Just one of the causes of DV, mentioned by Jordan Peterson, is alcohol consumption.

      Of course if alcohol consumption was found to be a *significant* factor how straightforward would it be to fix that bit of the problem? Blaming gender-inequality is much easier and panders to peoples’ preconceptions.

  6. MCA says

    The whole “root cause” thing reminds me of a concept from biology, the “key innovation”, a trait which, once it evolved, is causally linked to the subsequent success of that lineage. The archetypal example is powered flight, which has led to rapid diversification, high speciation rates, and geographic expansion each of the four times it evolved.

    Like in sociology, there is a strong impulse to grab any interesting adaptation at the base of a lineage and just declare it as the “key innovation” by fiat (and, while not political, the proposed trait almost always lines up with the proposer’s research interests). Unlike sociology, though, there seems to be a strong culture of pushback on premature claims, probably a leftover from Gould’s critique of adaptationist storytelling, and emprical/statistical tools to test hypothesized key innovations

    • nicky says

      Talking about biology, there is a distinction there betweem ‘proximate’ and ‘ultimate’ causes.
      Is there a parallel between ‘root’ causes and ‘ultimate’ causes?

  7. Jacob says

    Also circular causation, arguably including all attractors since the system’s evolution depends on its state, meaning that designating a root is arbitrary (or trivial in the case of one dimension).

  8. In using the term ‘root cause’ it needs to be clearly understood that it is a specific term used within a specific context and that is the analysis of accidents. The concept of root cause was necessary to discipline investigators to track a chain of events back to a ‘root cause’ in order to eliminate superficial findings, such as ‘human error’ or ‘mechanical failure’ being used to explain accidents. However, an accident is a specific unique outcome with vastly different chains of events, although some events are more common across chains. In order for a root cause to be assigned to a social outcome suffered by a group, it would be necessary to examine the entire chain of events for every individual in the group in order to identify the root cause for each and every individual’s outcome. The most likely result from this very, very expensive approach would be the identification of many differing root causes. This is in fact what root evaluation finds across accidents, most famously within the aviation industry where vast sums of money, time and resources are spent on this. Even in aviation, with all the money and resources dedicated to the activity, many root cause investigations cannot identify a root cause. The term root cause has a particular function within accident investigation. Its only function within social analysis is as shorthand for the words ‘in my humble opinion’. Two words for four. It should not be read as anything else.

  9. Good points in this article. We should be critical of “teleological” and “origins” thinking. Mythological constructions like “patriarchy” ignore the actual historical conditions and imagine that the enormously diverse cultures throughout history have all been the same.

    • peanut gallery says

      They aren’t big fans of things like context or nuance. Humans have been around a long time. Guess who predominately hunters? The male/female thing is a cooperation. Some cultures have valued this cooperation in different ways of the centuries, but it’s not as if the two sexes didn’t need one another to survive.

      What’s odd about this time in history is that the sexes can technically survive without one another… They just paint everything as PATRIARCHY! Yeah, cause men have had it super easy through out history…. Context matters.

  10. dirk says

    In biology, ecology and pathology, looking for the root cause in a complicated matter is more rule than exception. What was the root cause of the pest? The air? The rats? the hygiëne? Only after the invention of the microscope and bacteriology, the root cause was found. Why do certain crops die all of a sudden? Is it a fungus, a virus, or maybe the lack of certain micro nutrients? Same with malaria. Why did certain bird eggs not hatch in the 1950?? Was it the scarcety of feed for the brooding birds, or the long trek? No, It was the insecticide DDT. Now, it is tempting to look for similar root causes in society and humanity. Alas, this looking seldome results in clearcut answer, like in biology. But sometimes it brings us further, nevertheless.

  11. DBruce says

    Henry George identified the root cause of the boom-bust cycle, of persistent poverty, of growing wealth inequality. 150 years later he has still not been refuted.
    But who talks about Henry George?

  12. Chris says

    Nice & short article and logical conclusion, but at the very least brief on the root cause explanation.
    Those who have been involved in actual root cause analysis of complex systems, know that there is a difference between a probable cause (not proven, but worth investment to rectify) and a true “root cause”. Contributing factors are the fun part in as far as the root cause is always there, but the problem doesn’t manifest itself unless you have a certain combination of contributing factors to create the perfect storm.
    The other part any investigation needs is a concise and complete problem statement. That is the easiest part to mess with; i.e. perfectly logic argumentation about the cause of a “fake” problem. That is where I disagree with some comments that limit the concept to accidents, because it is de facto applied to significantly more “problems” as as can be seen by the existence of this article alone.
    The problem description is the give-away! In the example of domestic violence as used by the author, it is obviously incomplete (as he shows when adding same sex couples) and therefore it would be surprising to find a “complete” solution if you only look at half the problem.
    There are plenty of tools (some simpler than others) to deal with multiple variables, but who believes their proper use will do much good here is delusional. In politics and many other social settings, the “mechanic” has no interest in fixing the car, but making sure he can work on it for as often as possible – you will always get a partial fix, because he will present only part of the problem (fixing each part of the actual problem individually likely creates new problems).
    No matter where you life, it’s a beautiful scam and it can be witnessed leading up to every election.

  13. Bryan Atkins says

    I see the author’s point … and think there is truth in his work.
    As with almost all phenomena, one can add complexity. Below, I do that, literally.

    As a counter, I submit a few emergent causes … and contend there’s no one to blame and plenty to fear in the oncoming, onrushing collapse that brings the promise of “premature and perverted death” on an unprecedented scale.

    Re Blame: Many of our Larger problems are a function of the dominant phenomenon of our era: exponentially accelerating complexity. Briefly, add ~6 billion people since 1900, give billions of people access to exponentially more powerful tech that is largely driven by exponentially accruing knowledge.

    These phenomenon are emergent. They tremendously expand human reach & therefore alter myriad relationships in-and-across geo eco bio cultural & tech networks. This ongoing dynamic has generated unprecedented environs, too complex for a locally-coded species to navigate, to process.

    Consider the world’s dominant app: humans deploying monetary code. It lacks the ability to accurately calibrate relationship-value information in & across the aforementioned networks.

    The information-processing efficacy of the app has been crushed by complexity.
    Exhibits A & B: Sky; Ocean.

    Processing complex information with sufficient reach speed accuracy & power is the essence of passing multilevel selection tests. Think of an immune system. If it can’t sufficiently process viral-invader information, the biological system collapses unto death.

    Add this fundamental & terrifying code selected for relationship interface: Fitness > Truth.
    “Fitness and truth are utterly different things.”
    “Organisms that see the truth go extinct when they compete against organisms that don’t see any of the truth at all, literally none of the truth at all, and are just tuned to the fitness function.”
    “Perception is not about seeing truth; it’s about having kids.” Donald Hoffman

    Subcodes of Fitness>Truth are Me>U; Us>Them; Short Term>Long Term
    A Horror of Fitness>Truth (there are more): Genocide is an app.

    “Under pressure, any group of us can be as brutal as any of those we deplore: genocide by tribal animals is as natural as breathing…” James Lovelock

    “It should not be thought that war, often accompanied by genocide, is a cultural artifact of a few societies. Nor has it been an aberration of history, a result of the growing pains of our species’ maturation. Wars and genocide have been universal and eternal, respecting no particular time or culture.” E. O. Wilson – The Social Conquest of Earth

    The Horror… will continue to emerge. Love your loved ones.

    Our situation? … distilled, alphabet-coded from more than 2,000 years ago …
    From the IChing:
    The wall falls back into the moat. Use no army now. Make your commands known within your own town.
    Perseverance brings humiliation. The change alluded to in the middle of the hexagram has begun to take place. The wall of the town sinks back into the moat from which it was dug. The hour of doom is at hand. When matters have come to this pass, we should submit to fate and not try to stave it off by violent resistance. The one recourse left us is to hold our own within our intimate circle. Should we persevere in trying to resist the evil in the usual way, our collapse would only be more complete, and humiliation would be the result.

  14. Andrew Roddy says

    Ironic that Jordan Peterson should be referenced here. It would not be any great stretch to characterise the Professor as a Root Cause Warrior/Worrior. In his own intrepid adventures, sniffing out root causes for perceived maladies in ‘Western Society,’ he never seems need to look any further than to something he calls post-modern neo-Marxism. But since the thrust of this piece seems clearly to be identifying root causes as a dangerous fallacy of the left this irony is left unexplored.
    Our tendency to seek out root causes the way pigs dig for truffles may be an interesting one to explore but that is hardly the intention here. This piece confines itself strictly to the Quillette lens.

Comments are closed.