Despite lots of talking about diversity, the situation is still grim in most companies in the US for people of color and women. That’s not to even mention the many dimensions of diversity that we talk about even less, like disability, or socioeconomic class, religion, or the intersectionality of all of the above (and more).

There’s barely any publicly available data in the first place, but from what little we do know, more than 70% of company leaders are white men, and at the CEO level, this number jumps to more than 90%. The numbers are especially stark in tech where many firms have no women of color in leadership at all, and the representation of black and Latinx talent has barely budged over 14 years.

Image: Fortune

So it’s a sweet irony that a concept that might help those dismal diversity numbers comes to us from computer science.

In 1976, the (white, cis, male) computer scientist John H. Holland published a chapter in a now—and probably even then—obscure book called Automata, Languages, Development. It was on something he called the “spontaneous emergence of self-replicating systems.” While the chapter furthered his ongoing work on genetic algorithms, this publication is important because it marks his first traceable use of the term “emergence,” for which Holland would later become widely known (in the admittedly narrow fields of complexity and computer-science).

But this esoteric concept can also help us understand something very tangible: how to actually increase diversity in organizations.

What is emergence?

Emergence says that the whole is not just greater than, but can even be different from the sum of its parts. It’s the idea that there are some issues we cannot fully understand by simply looking at the constituent parts and adding them up.

Consciousness is a good example. A perfect understanding of all the neurons and synapses in a brain is not enough to understand why humans are self-aware. Instead, we have a sense that neurons and synapses interact, which gives rise to something else: the ability to think, reflect, and, say, enjoy Season 8 of Game of Thrones.

Emergence can give rise to “emergent outcomes,” which are hard-to-predict big events that show up as a result of a whole bunch of decentralized, micro-level behaviors. One tangible example is a traffic jam. Every morning, lots of people decide to drive to work, and every morning this results in rush hour. Occasionally it is really bad; an extreme example is the 2010 traffic jam outside of Beijing that lasted for 12 days. The cause? A nasty mix of roadwork, heavy volumes of trucks transporting coal, vehicle breakdowns, and minor accidents. Eventually, people falling asleep at the wheel further worsened the congestion.

As much as honking your horn makes you feel better, emergence says it’s both not the other driver’s fault (so there’s no point in honking) and also all of the other drivers’ faults (but honking still won’t solve anything anyway). The key to emergence is that the individual actors do not need to deliberately set out to cause the things to which their actions give rise. Presumably, no one leaves their house in the morning with the goal of being 12 days late to work, yet the collective result of many of people independently choosing to do what’s best for themselves in a given moment (go to work) is a bunch of cars idling on the highway. Rush hour is fairly predictable, but the big jams are hard to see coming. (If they were, presumably, we could avoid them.)

Emergence is a hallmark of something called a complex system. This is a network of individuals who are interacting with each other while serving their own self-interest, not the common good; it’s basically a fancy way of saying we’re all just trying our best. A market of stock traders, a flock of birds, and a team of humans in a company can all be thought of as complex systems. The emergent outcomes—both good and bad—of these systems might be stock market crashes, a flock weaving itself into an aerodynamic V shape—or the creation of either a diverse or overwhelmingly homogenous team.

The connection between emergence and diversity

Similarly, a lack of diversity in companies can be caused—at least in part—by a whole bunch of people behaving in ways that make sense for them in the moment. They may not immediately cause obvious harm, but in aggregate, they produce stark homogeneity.

For example, a manager inviting a team member out to lunch to discuss career goals isn’t problematic on its own. The manager might even be responsible and take each team member out to lunch in an effort to not play favorites. But if during that lunch, the manager is a little more casual and a little more friendly with the employee who is more similar to him (and it’s usually a him), and this behavioral tendency repeats over years, emergence suggests that the aggregate outcome will be much greater than just a little more bias in favor of similar employees. It’s a game of math.

Just as individual neurons firing in your brain can give rise to consciousness, individual team members interacting over many years can give rise to toxic cultures. These environments are so unfavorable to some employees, often those from groups who are historically marginalized, that they are made miserable, if not driven out completely.

How to use emergence to increase diversity

This relatively esoteric concept of emergence is useful in understanding why well-intentioned people can’t seem to budge representation numbers in a company, even when they don’t feel like they are biased or discriminatory.

In fact, most companies are obsessed with outcomes, such as “How do I get more people of color into leadership?” That’s a fine goal, but a more useful question to ask to get there might be: “How do I change the inputs that gave rise to this homogeneity in the first place?”

We can use emergence to our advantage to help explain this.

We’re all (hopefully) very aware of things like unconscious biases and microaggressions, but we often don’t actually know how to translate that awareness into actions. Emergence can help those with both conscious and unconscious anti-diversity behaviors understand that while these behaviors look and feel tiny, they can have a big aggregate effect. There’s also often an asymmetry in perception: Those who are on the receiving end of these slights feel them acutely, and those who perpetrate them cannot (or will not) understand why each little offense matters so much.

One of the biggest barriers to improving diversity in companies is that most behaviors don’t feel particularly bad. Take microaggressions, for example. Microaggressions are an important concept with a terrible name: just because they are small doesn’t mean they aren’t mighty. Microaggressions refer to subtle, often automatic behaviors we all engage in that can cause harm to others. But the term “micro” is often wrongly misconstrued as “insignificant” and “aggression” immediately leads to defensiveness in the part of the deliverer of the harmful words or behavior. Some examples of microaggressions in workplaces include interrupting or talking down to someone in a meeting, asking where someone is “really” from, or showing surprise when someone demonstrates even a basic grasp of mathematics (even if you think you are being complimentary). A major challenge surrounding microaggressions is convincing their perpetrators that those small asides matter. In fact, emergence says it’s the little things that add up that matter the most.

Emergence helps us see why tiny slights matter. But these slights are largely invisible to those of us committing them.

To understand, people in management positions need to first become better listeners. Companies are rife with people articulating what’s problematic—the manterruptions, the credit-claiming for ideas from people of color, and on and on—but little is done about them. If you’re a white person, listen to what people of color are saying, and take them seriously. If you’re a man, listen when a woman says something is wrong with your behavior, and take them seriously.

And if you don’t hear anything: Ask, listen, and ask again.

The logic of emergence can be liberating because it helps us understand that outcomes sometimes aren’t really anyone’s fault in particular. But it’s also frustrating, because it then means that we need to work as a collective whole to solve it. It’s the opposite of reductionism, which is the optimistic hope that if only we could understand every single part, we could perfectly explain the whole.

Emergent outcomes—climate change, war, national prosperity, and meaningful diversity—are not easily engineered or changed. But recognizing these issues as emergent helps us identify one place to start: ourselves.